University of Southern California Dissertations and Theses

Prostate cancer: genetic susceptibility and lifestyle risk factors

(USC Thesis Other)

Prostate cancer: genetic susceptibility and lifestyle risk factors

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content PROSTATE CANCER: GENETIC SUSCEPTIBILITY AND LIFESTYLE RISK FACTORS By Anqi Wang A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (EPIDEMIOLOGY) May 2023 Copyright 2023 Anqi Wang ii Acknowledgements As I sit down to write these acknowledgements, I am filled with a sense of awe and gratitude. The journey that has led me to this moment has been one of the most challenging, exhilarating, and rewarding experiences of my life. I could not have done it without the support and guidance of countless ones along the way. First and foremost, I would like to express my deepest gratitude to my advisor, Dr. Chris Haiman, for his unwavering support and guidance throughout my Ph.D. journey. His encouragement, insights, and constructive criticism have been invaluable in shaping my research and developing my skills as a scholar. I would also like to thank Dr. Dave Conti, who, with his humor, patience, and vast knowledge in both biostatistics and epidemiology, has been an inexhaustible source of guidance and inspiration for me. I am deeply grateful and honored to have their mentorship. I am also incredibly grateful to the members of my dissertation committee, Dr. Anna Wu, Dr. Nick Mancuso, and Dr. Amir Goldkorn, whose knowledge, critiques, and support have been instrumental in refining my ideas, strengthening my arguments, and polishing my dissertation. Their expertise and dedication have helped me to grow as a scholar, and I am humbled by their contributions to my work. They are not only respected scholars, but also my role models in academic pursuit. I am particularly grateful for the supportive and collaborative environment of my lab. I would like to thank my lab members, Grace, Peggy, Susan, Victor, Loreall, Kimberley, Toni, and Amanda, for their invaluable contributions to my research and for making the lab feel like a second home. I iii want also to express my gratitude to Burcu and Fei, who are now two rising star professors in the field, for their invaluable suggestions and guidance to my research. I am also fortunate to have my peers in the lab, Alisha, Zhaohui, Raymond and David, for their invaluable contributions to my research and for their support and friendship throughout this journey. I am grateful to have had the support and companionship of Senkei, Keren, and Kaili, my fellow graduate students in the cohort. Together, we have shared in the joys and struggles of graduate school, and their passion for learning and dedication to work have always been a constant source of motivation to me. Lastly, I would like to extend my heartfelt appreciation to my family, who have consistently supported and encouraged me on this academic journey. I would also like to thank my significant other, Yili, who has been my rock and source of love and encouragement during the past five years. This dissertation wouldn't have been possible without his ongoing support, patience, and belief in me. Completing a Ph.D. is not the end of my journey, but rather the beginning of a new chapter. I am humbled and grateful for the support and encouragement of those who have contributed to my success along the way. iv Table of Contents Acknowledgements ........................................................................................................................................................ ii List of Tables................................................................................................................................................................. vi List of Figures .............................................................................................................................................................. vii Abstract ....................................................................................................................................................................... viii Chapter One: Introduction.............................................................................................................................................. 1 1.1 Overview of prostate cancer ................................................................................................................................ 1 1.2 Chronic inflammation and prostate cancer .......................................................................................................... 3 1.2.1 Intraprostatic inflammation on the mechanism of prostate carcinogenesis ................................................. 3 1.2.2 Atopic allergic conditions (AACs) .............................................................................................................. 5 1.2.3 Smoking ....................................................................................................................................................... 7 1.2.4 Air-pollution ................................................................................................................................................ 8 1.3 Genetic determinants of prostate cancer ........................................................................................................... 10 1.3.1 Evidence of genetic contribution to prostate cancer risk ........................................................................... 10 1.3.2 Genetic susceptibility loci ......................................................................................................................... 12 1.3.3 Polygenic risk score ................................................................................................................................... 14 1.4 Post-diagnostic factors and prostate cancer survival ......................................................................................... 15 1.4.1 Common medications ................................................................................................................................ 15 1.4.2 Dietary factors ........................................................................................................................................... 16 1.4.3 Obesity ....................................................................................................................................................... 18 1.4.4 Physical activity......................................................................................................................................... 19 1.4.5 Smoking ..................................................................................................................................................... 20 1.4.6 Summary.................................................................................................................................................... 21 1.5 References ......................................................................................................................................................... 22 Chapter Two: Specific Aims ........................................................................................................................................ 35 v Chapter Three: Characterizing prostate cancer risk through multi-ancestry genome-wide discovery of 187 novel risk variants ......................................................................................................................................................................... 36 3.1 Background ....................................................................................................................................................... 36 3.2 Methods ............................................................................................................................................................. 37 3.3 Results ............................................................................................................................................................... 49 3.4 Discussion ......................................................................................................................................................... 57 3.5 References ......................................................................................................................................................... 61 Chapter Four: Clonal hematopoiesis and risk of prostate cancer in large samples of European ancestry men ........... 66 4.1 Introduction ....................................................................................................................................................... 66 4.2 Methods ............................................................................................................................................................. 66 4.3 Results ............................................................................................................................................................... 71 4.4 Discussion ......................................................................................................................................................... 74 4.5 References ......................................................................................................................................................... 76 Chapter Five: Atopic allergic conditions and prostate cancer risk and survival in the Multiethnic Cohort Study ...... 79 5.1 Background ....................................................................................................................................................... 79 5.2 Methods ............................................................................................................................................................. 80 5.3 Results ............................................................................................................................................................... 85 5.4 Discussion ......................................................................................................................................................... 88 5.5 References ......................................................................................................................................................... 92 Chapter Six: Summary and Future Directions ............................................................................................................. 96 Appendix: Supplementary Materials.......................................................................................................................... 117 Chapter Three ........................................................................................................................................................ 117 Chapter Four .......................................................................................................................................................... 151 Chapter Five .......................................................................................................................................................... 157 vi List of Tables Chapter Four Table 1. Associations of four common CHIP genes and the aggregate of all identified CHIP genes with overall and aggressive prostate cancer risk………………………………………....101 Table 2. Associations of four common CHIP genes and the aggregate of all identified CHIP genes with age at blood draw and prostate cancer diagnosis in the UK Biobank………………102 Chapter Five Table 1. Baseline characteristics by status of AACs in the 74,714 men in the Multiethnic Cohort study…………………………………………………………………………………….103 Table 2. Hazard ratios of prostate cancer outcomes associated with AACs in the Multiethnic Cohort, 1993-2017 (N=74,714)……………………………………………………104 Table 3. Hazard ratios of prostate cancer associated with AACs status among subgroups in the Multiethnic Cohort, 1993-2017…………………………………………………………..105 vii List of Figures Chapter Three Figure 1. Manhattan plot of results from the multi-ancestry PCa meta-analysis........................ 106 Figure 2. Comparisons of genome-wide significant variants across populations. ...................... 107 Figure 3. Percentage of cases in the top (quintile 1) and bottom (quintile 5) GRS risk quintile based on GRS100, GRS181, GRS269, and GRS451 in the multi-ancestry sample. .......................... 108 Figure 4. Associations between standardized GRS and prostate cancer by populations. ........... 109 Figure 5. Cumulative absolute risk by age.................................................................................. 111 Extended Data Figure 1. Venn diagram of variants common (MAF>1%) among European, African, Asian, and Hispanic populations. ................................................................................. 112 Extended Data Figure 2. Odds ratios and 95% confidence intervals for one SD increase in GRS451 and total prostate cancer risk in the GWAS discovery and replication sub-studies and meta-analysis by ancestry. .......................................................................................................... 113 Chapter Four Figure 1. Manhattan plots of associations between CHIP genes and age at blood draw in the UK Biobank. CHIP: Clonal hematopoiesis of indeterminate potential. ..................................... 114 Chapter Five Figure 1. Association between antihistamine use on a) prostate cancer incidence and mortality, and b) prostate cancer survival in the multiethnic cohort. .......................................................... 115 Figure 2. Association between the duration of antihistamine use on a) prostate cancer incidence and mortality, and b) prostate cancer survival in the multiethnic cohort. ................................... 116 viii Abstract Prostate cancer is a significant public health issue and one of the most common cancers among men worldwide. The exact causes of prostate cancer are not yet fully understood, but several risk factors have been identified, including age, family history, genetic risk variants, and environmental factors. This dissertation aims to investigate the genetic and environmental risk factors for prostate cancer and their impact on the development and progression of the disease. Specifically, Chapter 3 focuses on identifying novel genetic risk variants for prostate cancer in multi-ancestry populations through large-scale genome-wide association studies (GWAS). Chapter 3 involves conducting ancestry-specific and multi-ancestry GWAS meta-analyses to identify novel genetic risk variants and develop a genetic risk score to characterize prostate cancer risk in multi-ancestry populations. Chapter 4 investigates the association between clonal hematopoiesis, an aging-related expansion of blood cells with somatic mutations, and prostate cancer. Chapter 5 examines the effect of atopic allergic conditions, a hypertensive immune state, on the risk of prostate cancer, disease severity, and survival in multi-ethnic populations. The findings from this thesis studies have the potential to contribute to a better understanding of prostate cancer etiology and the development of new strategies for prevention, early detection, and treatment. Ultimately, this research may help improve the quality of life and survival rates of those affected by prostate cancer. 1 Chapter One: Introduction 1.1 Overview of prostate cancer Globally, prostate cancer is one of the most common types of cancer incidence and causes of cancer death. 1 The Global Burden of Disease Cancer Collaboration estimated that in 2017, there were 1.3 million (95% uncertainty interval (UI): 1.2-1.7 million) newly diagnosed prostate cancer cases and 416,000 (95% UI: 357,000-490,000) prostate cancer deaths. 1 It is shown that men in countries with higher levels of socioeconomic development tend to be more likely to develop prostate cancer. 1 In the United States, prostate cancer is the most common non-skin type of cancer among men with an estimated incidence of approximately 248,530 per year in 2021. 2 Prostate cancer is also the second leading cause of cancer death in men in the United States, right after lung and bronchus cancer, with an estimated 34,130 deaths in 2021. 2 Prostate cancer exhibits the greatest variation across different race/ethnicity groups of all cancer types in the US, with African Americans in particular having significantly higher incidence and mortality rates than other racial/ethnic groups. For example, in 2013-2017, the age-adjusted incidence rate of prostate cancer among African American was 171.6 per 100,000 men, which is nearly twice as high as Non-Hispanic White (97.7 per 100,000) and Hispanic men (85.6 per 100,000), and more than three times as high as Asian/Pacific Islanders (53.8 per 100,000). 2 Similarly, during 2014-2018, the age-adjusted mortality rate among African American men was 38.3 per 100,000 men, approximately twice as high as Non-Hispanic White (17.9 per 100,000) and Hispanic men (15.6 per 100,000), and more than four times as high as Asian/Pacific Islanders (8.8 per 100,000). 2 To date, extensive efforts have focused on identifying risk factors for racial 2 disparities, as well as reducing racial disparities along the prostate cancer continuum, from screening and diagnosis to treatment and survival. 3 Current research has suggested that social- economic factors, lifestyle, environment, and germline genetic variation interact in a complex way to contribute to racial disparities. 3,4 The clinical feature of prostate cancer is highly heterogeneous, and while most men developed a low-risk indolent disease, a small proportion of cases develop a highly aggressive and potential lethal disease. 5 It is suggested that indolent and lethal prostate cancers may represent diverse etiologies. 5 Thus, they present differently in terms of clinical symptoms, treatment, and survival. Men with indolent disease generally have low staged tumor and exhibit fewer or no symptoms, which can be detected by a prostate-specific antigen (PSA) screening test. Due to the slow progression of disease for men with indolent prostate cancer, treatment is not usually required, but rather active monitoring is recommended. Indolent prostate cancer generally has favorable clinical consequences. Between 2010-2016, the 5-year survival rates for localized and regional stage prostate cancers were over 99%. 6 Prostate cancers that are lethal progress more rapidly and are usually detected at a more advanced stage. When the cancer is in a late-stage, the tumor may metastasize to nearby organs or tissues and cause pain. In contrast to the high survival for men with low-risk prostate cancer, metastatic cancer has only a 30% 5-year survival rate. 6 While it is important to detect the disease at an early stage, PSA screening may also lead to over-diagnosis and unnecessary treatment given that a large proportion of prostate cancers are indolent. 3 1.2 Chronic inflammation and prostate cancer 1.2.1 Intraprostatic inflammation on the mechanism of prostate carcinogenesis Prostate cancer is caused by a combination of environmental and genetic factors. Older age, a family history of prostate cancer, genetic susceptibility risk loci, and African ancestry are the most well-established risk factors for prostate cancer. 7 Over the last two decades, it has been proposed that chronic inflammation of the prostate is linked to prostate tumorigenesis, which can be induced by environmental factors such as bacteria or virus infection, gut microbiome, diet, obesity, hormone alternation, and physical injuries of the prostate. 8-10 Furthermore, the immune response triggered by inflammation can in turn influence tumor progression and treatment response, thereby altering prognosis and survival. 8-10 A complex series of molecular mechanisms underlie the progression of chronic inflammation to prostate cancer. Histologically, inflammation is commonly observed in prostate tumorigenesis. Inflammatory infiltrates are often seen in proliferative inflammatory atrophy (PIA) regions in prostate, which is characterized by proliferation of epithelial cells in the atrophic zone. 11,12 Certain external stimuli (e.g., infections, trauma, chemical exposure, etc. ) may trigger an inflammatory micro-environment, where immune cells release reactive oxygen species (ROS) and nitrogen species (RNS) as well as pro-inflammatory cytokines to induce genetic and epigenetic mutations in normal prostate cells, ultimately leading to tissue repair response and result in abnormal epithelial cell proliferation (PIA). 8,9 Furthermore, the mutated cells may in turn trigger inflammation and immune response resulting in oxidative stress, and PIA may further transit into prostatic intraepithelial neoplasia (PIN), 11,13 which is a precancerous lesion of the prostate. 14 During this process, it has been found that T-cells and macrophages are immune cells that are more 4 frequently involved in prostate carcinogensis. 15-17 Macrophages include tumor-associated macrophages (TAMs), tissue-resident macrophages, and myeloid derived suppressor cells (MDSCs). 18 TAMs and MDSCs are recognized as having a variety of tumor-promoting properties, including angiogenesis, invasion, and metastasis. 19 T-cells has been found to have both anti-tumor or pro-tumor effects. For example, CD8 + cytotoxic T (CTL) and CD4 + helper T type 1 (Th1) cells are indicators of an improved survival among multiple types of cancer patients, 20-22 while Th2, Th17 and T regulatory (Treg) cells infiltration often correlate with an unfavorable prognosis through a pro-tumor function or suppression of anti-tumorigenic response. 20,23-25 In this way, immunosurveillance against tumor cells and tumor-promoting responses are in a dynamic balance until one side prevails. Evidence from epidemiological studies is shown to support the role of inflammation in prostate cancer development. Early case-control studies before 2000 suggested a positive association between prostatitis and prostate cancer. 26 In case-control study participants from the Prostate Cancer Prevention Trial (PCPT) and its follow-up study PCPT-SELECT, inflammation of prostate tissues was prospectively collected through biopsy, and it was found that the extent of benign inflammation in prostate tissues was associated with a higher risk of prostate cancer. 27,28 Another case-case study among men with low-grade prostate cancer at diagnosis reported that men with chronic inflammation in prostate tissue tended to have a higher risk of lethal prostate cancer. 29 Nonetheless, most studies showed a negative association between needle biopsy-detected prostate inflammation and prostate cancer. 30 Such inverse associations might be due to a collider/stratification bias, given the study populations were selected conditional on elevated PSA level (men who were referred for biopsy might either have intraprostatic inflammation or prostate 5 malignancy that leaded to increased PSA level or other clinical indications). 31 However, it cannot be ruled out that these studies discovered a lower incidence of prostate cancer among men with prostatic inflammation due to enhanced immunosurveillance. A range of external environmental factors, including urinary bacteria or virus infection (i.e., sexually transmitted infections), gastrointestinal microbiome, diet and obesity, hormonal alternation, and corpora amylacea in the prostate gland have been linked to the progression of prostate cancer through chronic inflammation. 8,10 These factors can cause a local inflammatory response in the prostate or its surrounding tissues or organs. In what follows, we will focus on the biological potential and epidemiological evidence of other environmental factors to contribute to prostate cancer carcinogenesis through chronic inflammation that may not directly cause inflammation of prostate tissue, but rather through a systemic chronic state of inflammation within the body. 1.2.2 Atopic allergic conditions (AACs) As stated above, the persistence of an inflammatory environment could trigger prostate cancer development and progression through a complex immune response. 32,33 Particularly, atopic allergic conditions (AACs) such as asthma, hay fever, and food allergies have been considered to imply a heightened immune response status, thus may be associated with elevated risk of prostate cancer. Atopic allergic conditions usually refer to immunoglobulin E (IgE)-mediated hypersensitivity reactions to common allergens in the environment. When exposed to an allergen, people who suffers allergies are likely to release cytokines produced by Th2 cells at the affected site, whereas normal people prone to produce interferon-γ by Th1 cells in response to the 6 allergen. 34 Serum Th2 cells are more prominently expressed in prostate cancer patients than in normal individuals. 35 Meanwhile, in a study using the National Health and Nutrition Examination Survey 2005-2006, IgE level is found to be inversely associated with PSA level, suggesting a potential anti-tumor role of AACs. 36 Current epidemiological evidence on the effect of AACs on prostate cancer has been mixed. A meta-analysis of 14 studies (7 case-control, 7 cohort study) through 2015, reported an overall null association between asthma and prostate cancer (OR=0.99, 95% CI: 0.84-1.18) 37 . A prospective analysis using the Health Professionals Follow-up Study reported a significant protective effect of asthma on lethal (RR=0.71, 95% CI: 0.51-1.00) and fatal (RR=0.64, 95% CI: 0.42-0.96) prostate cancer 38 . By contrast, three large population-based cohort studies, the Busselton Health Survey HR= 1.89, 95%CI: 1.00-3.60) 39 , Melbourne Collaborative Cohort Study (HR=1.25, 95% CI: 1.05- 1.49) 40 and Taiwan National Health Insurance Research Database (HR=2.36, 95% CI: 1.22-4.57) 41 , reported deleterious effects of asthma on total prostate cancer. In terms of allergy, a meta-analysis of 6 studies by 2015 reported a slightly increased but null association for atopy (RR=1.25, 95%CI: 0.74–2.10) and for hay fever (RR=1.04, 95%CI: 0.99–1.09), whereas a slightly decreased but non- significant risk (RR=0.96, 95%CI: 0.86–1.06) for any allergy. 42 The Health Professionals Follow- up Study reported a significant increased risk of total prostate cancer among men with hay fever (RR=1.07. 95% CI: 1.01=1.13). 38 The Busselton Health Survey also found a large increased risk of prostate cancer with any atopy (HR=2.49, 95%CI: 1.04-5.93). 39 In addition, few above- mentioned studies investigated the influence on aggressiveness of prostate cancer 40,43 , and most studies were based on White populations. Therefore, there is a need to explore how AACs 7 contributes to the development of prostate cancer, in particular aggressive and fatal prostate cancer in a diverse population (Section 2.1). 1.2.3 Smoking Smoking has been found to both promote systematic inflammation and have an anti-inflammatory effect. 44,45 Moreover, there is evidence to suggest that smoking is associated with inflammation of the prostate. In the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) study, men with slightly elevated PSA (2.5-10ng/ml) underwent biopsy to detect inflammation. Current smokers were significantly associated with increased risks of both acute (OR=1.30, 95%CI: 1.08– 1.56) and chronic inflammation of the prostate (OR=1.24, 95%CI: 1.05–1.47), whereas former smokers were not (acute OR=1.01, chronic OR= 1.00, both P’s>0.05). 46 Furthermore, smoking may raise blood levels of interleukin-18 (IL-18) and IL-12, pro-inflammatory cytokines commonly detected in areas of chronic inflammation, and may further contribute to increased inflammation and poor survival in men with prostate cancer. 47,48 Previous results on smoking and total prostate cancer incidence remains inconclusive. A meta- analysis of 51 prospective cohort studies showed that current smoking tended to be associated with lower risk of prostate cancer incidence (RR: 0.90, 95% CI: 0.85-0.96), but higher risk of prostate cancer mortality (RR: 1.24, 95% CI:1.18-1.31) 49 . Another systematic review suggested that current smoking is likely to be associated with about 2 to 3 fold-greater risk of aggressive prostate cancer risk 50 . Previous analysis in the Multiethnic Cohort (MEC) Study also found that smoking has a protective effect on total prostate cancer (HR=0.72, 95%CI: 0.63–0.83) 51 ; however, the association with aggressive prostate cancer has not been studied. On the other hand, findings regarding 8 association between smoking and prostate cancer survival or recurrence are more consistent. In population-based cohort studies, current smokers at the diagnosis of prostate cancer also tended to show a poorer survivorship and higher risk of recurrence compared with non-smokers, with HRs for prostate cancer specific mortality ranging from 1.14 to 2.0 52-57 . Generally, smokers are less likely to undergo PSA screening. 58 Also, smokers tend to have lower total PSA level compared to non-smokers. 59 Whether the inverse association between smoking and prostate cancer is an artifact of lower screening in smokers isn’t clear. Given this issue, I will also assess the association between smoking and aggressive prostate cancer, as well as the effect of smoking on prostate cancer-specific mortality in the MEC, taking into account the effect of PSA screening. 1.2.4 Air-pollution Pollutants in the air that come from traffic and industry can trigger an inflammatory reaction and hypersensitive state, particularly in the respiratory tract. 60-62 In addition to causing worsening asthma symptoms, 63,64 air pollution has also been shown to be associated with the onset of new asthma conditions in childhood. 65-67 Epigenetic studies have shown that traffic-related air pollution (TRAP), such as polycyclic aromatic hydrocarbon (PAH) and diesel exhaust may lead to Treg cells suppression, Th1 cells to Th2 cells conversion, and Th2/Th17 differentiation. 68-71 Epidemiological evidence on air pollution and prostate cancer risk is limited, although a positive link between air pollution and prostate cancer incidence has been suggested. A majority of studies have focused on the effect of traffic-related air pollution (TRAP), specifically nitrogen dioxide 9 and nitrogen oxides, on prostate cancer risk and reported a statistically significant positive association. 72-77 A cohort study in German reported a 1.23 (95%CI:1.08–1.39) fold increased risk of prostate cancer incidence per 10 μg/m 3 increase in PM10. A retrospective population-based study in Shanghai also reported a statistically significant positive correlation between industrial waste gas emissions and prostate cancer incidence. 78 It's worth noting that none of these research studies were carried out in the United States, and most focused on White and Asian populations. Therefore, there are large research gaps regarding 1) the effects of other types of air-pollutants 2) the association of air pollution with risk of different prostate cancer outcomes 3) and impact across diverse race/ethnicity groups, all of which need to be further addressed. 10 1.3 Genetic determinants of prostate cancer 1.3.1 Evidence of genetic contribution to prostate cancer risk Prostate cancer is a highly heritable type of malignancy. As stated above, family history is a well- recognized risk factor for prostate cancer. Familial aggregation of prostate cancer was first reported by Morganti et al. in 1956. 79,80 Although some early segregation studies suggested that prostate cancer is likely to be inherited in an autosomal dominant pattern, more recent data tended to support a polygenic role in prostate cancer inheritance. 81-85 In later epidemiological studies, family history was consistently associated with an increased risk of prostate cancer, with the risk being strongest if first-degree relatives are affected. 86 A family history of female breast cancer was also shown to be associated with an increased prostate cancer risk. 87 Moreover, having a family risk of prostate cancer is also associated with earlier onset of prostate cancer, higher prostate cancer mortality, higher recurrence rate after radical prostatectomy, and higher incidence of secondary primary cancers, but there were no clear clues as to the impact on prostate cancer survival. 88-90 Although studies based on familial relationships provide strong evidence of heritability, it is difficult to distinguish between shared environment and genetics between family members. While in twins, their inherited genomes are identical (for monozygotic twins) or nearly half identical in average (for dizygotic twins). Assuming the environment was shared to the same extent between monozygotic and dizygotic twins, the heritability of prostate cancer can be estimated by comparing their concordance. In 2000, a study with 44,788 pairs of twins from the Swedish, Danish, and Finnish twin registries estimated the heritability of prostate cancer to be 42% (95%CI: 29%-50%) of, which is the highest among all the common cancer types examined in the study. 91 In 2016, an extended study twins from the Nordic countries (Denmark, Finland, Norway, and Sweden) with 11 203,691 twin pairs reported a 57% heritability (95% CI: 51%-63%) and an excess familial risk for dizygotic twins (which can be generalized to siblings) of 22% (95%CI: 18.8-25.7). 92 Additionally, racial disparities in prostate cancer incidence and mortality may also indicate a heritable component in prostate cancer carcinogenesis. A whole-genome admixture scan in African Americans revealed a region on chromosome 8q24, enriched with prostate cancer susceptibility risk loci with significantly different frequencies between West Africans and European Americans. 93 This genetic association is stronger in younger men, consistent with the overall trend of earlier prostate cancer diagnoses in African Americans compared to other populations. A subsequent genotyping in the region, with five major race/ethnicity populations in the US, identified 7 risk alleles. Among these populations, African American exhibited the highest population attributable risk (PAR)for the combination effect of the 7 alleles (68%), indicating that genetic factors contribute to the racial disparities in prostate cancer. 94 Moreover, evidence suggested that ancestral genome was also directly linked to prostate cancer risk. A prospective study on ancestry and prostate cancer risk among men with a family history and intensive screening frequency found that the risk of prostate cancer increased with Western African ancestry among men with self-reported race of European or African American. 95 In light of the evidence above, genetics plays a pivotal role in prostate carcinogenesis. Understanding the genetic basis of prostate cancer could advance our knowledge of the mechanism of prostate cancer initiation and development. More importantly, in the context of precision medicine, it could help develop targeted screening and prevention strategies for high-risk men and 12 design tailored treatments for each individual patient to minimize over-diagnosis and unnecessary treatment. 1.3.2 Genetic susceptibility loci Before the Human Genome Project lifted the veil on single-nucleotide polymorphisms (SNPs) and linkage disequilibrium along the human genome, linkage analysis within disease enriched families was the primary approach to search for genetic susceptibility loci for prostate cancer. Such studies are designed to evaluate genetic markers in high-risk families and compare the segregation and recombination events of the markers among all members. 96 The first prostate cancer gene identified by linkage analysis was the HPC1 allele on chromosome 1q24-25 97 , which is later linked to the RNASEL gene. 98 A number of risk loci were then reported by a series of linkage analyses, covering 16 of the 24 chromosomes. 99 However, studies with high-risk families are usually limited to findings on heritable or familial prostate cancer, whereas sporadic prostate cancer cases account for a greater proportion of all the cases. 100 Moreover, risk loci identified by linkage-based analysis often failed to be validated in other study populations. Therefore, the genetic basis for prostate cancer, including the familial component of the disease, is more likely to be explained by many common but low penetrance variants. Genome-wide association studies (GWAS) connect common genetic variants to disease risk. Following the availability of reference panels of genetic variation from Human Genome Project and the International HapMap Project, the reduced cost of genotyping, rapid development of genotype imputation, and the availability of large biobank resources has led to a dramatic expansion of SNP-trait discoveries by GWAS. To date, the largest multi-ancestry GWAS meta- 13 analysis of prostate cancer, combining 107,247 prostate cancer cases and 127,006 controls from European, African, East Asian, and Hispanic populations, has identified 269 risk loci for prostate cancer, and 86 of them were not reported in previous GWAS. 101 In total, the 269 risk loci contributes to 33.6% of excessive familial attributable risk among East Asian, 38.5% among Hispanic, 42.6% among European ancestry men, and 43.2% among African ancestry men. 101 Although this meta-analysis provided a comprehensive understanding on the genetic susceptibility loci for prostate cancer across ancestry groups, populations from non-European groups were still under-represented. In particular, the African ancestry population has a more diverged genetic architecture, and thus would need larger sample size to comprehensively study all genetic variation. A recent simulation study estimated that, in order to explain an 80% of the total prostate cancer heritability, a GWAS would need at least 500,000 men (assuming a 1:1 case:control ratio). This number is far beyond the current sample size from the meta-analyses. 102 In addition, limited by genotyping/imputation accuracy of rare alleles, most GWAS have restricted analyses to variants with a minor allele frequency (MAF) more than 1%, or 5% in earlier studies. Given the improvement of genotyping and imputation platforms, SNPs with lower MAF can now be genotyped/imputed and tested with greater accuracy. Those rare variants, may contribute to the missing heritability from current GWAS. 103 In addition to total prostate cancer, efforts from GWAS have also been made to explore genetic predisposition of more aggressive subtypes of prostate cancer. Limited by sample size and power to detect low-penetrance risk loci, most case-only GWAS were not able to identify genome-wide significant loci. 104-106 A GWAS with 4,545 prostate cancer cases observed three loci, rs35148638, rs35148638, and rs62113212 associated with Gleason score, which can be used to differentiate 14 aggressiveness of prostate cancer. 107 Among the three loci, rs62113212 is a known risk loci for total prostate cancer, and located on the gene KLK3, which is the PSA coding gene. 108 Although this significant association might be confounded by serum PSA level, a replication study with 18,343 prostate cancer patients further supported this finding. 109 1.3.3 Polygenic risk score With genome-wide level estimation of SNP-trait associations, a polygenic risk score (PRS), or genetic risk score (GRS) or polygenic score (PGS), can be calculated to obtain an aggregated estimate of an individual-level genetic heritability across all disease risk loci. The most commonly used method to calculate a PRS is to sum the weighted dosage for each risk locus within each individual. 110 In addition to prediction of individual-level risk of prostate cancer, PRS could also serve as a surrogate for one’s genetic heritability of the disease. Prostate cancer PRS has been used to evaluate putative causal relationships in Mendelian Randomization studies 111 , combining with other traits to inform biopsy or treatment decisions, and stratify patients by their genetic risks 112- 115 . In addition to providing risk prediction, GWAS-based PRS also helps to understand the relationship of common variants and prostate cancer mechanisms. The PRS of the 269 risk loci for prostate cancer was shown to be associated with age at diagnosis, and family history of prostate cancer, but was not associated with disease aggressiveness, suggesting the common variants in aggregation may contribute to the etiology of familial prostate cancer but rarely to aggressive prostate cancer. 101 15 1.4 Post-diagnostic factors and prostate cancer survival Given the high incidence of prostate cancer and the highly heritable nature of prostate cancer, identifying modifiable risk factors for prostate cancer progression and death can help prolong survival and potentially improve quality of life after diagnosis. A large body of work has focused on identifying risk factors prior to the onset of prostate cancer, while evidence for post-diagnostic lifestyle factors on prostate cancer progression and overall survival is relatively scarce. 1.4.1 Common medications Many studies have looked at the impact of post-diagnostic medications on prostate cancer outcomes. As prostate cancer is common among older men, many prostate cancer patients have chronic conditions, such as cardiovascular diseases, diabetes, hyperlipidemia ，and others. Many of the medications prescribed for these chronic diseases have been shown to reduce prostate cancer incidence and mortality and could be potentially influence prostate cancer survival and prognosis. Beta-blockers, a common type of anti-hypertensive drug that has a potential anti-carcinogenesis effect on prostate cancer incidence and mortality, 116-120 was found to be not associated with prostate cancer specific mortality or all-cause death after diagnosis. 121-124 A meta-analysis with 13 studies (published before 2018) on post-diagnostic aspirin use showed a null association with prostate cancer specific mortality. 125 Another meta-analysis with 6 studies (published before 2016) of non-steroidal anti-inflammatory drugs (NSAIDs) exposure observed a protective effect on distant metastasis, with post-diagnostic use having an even stronger impact that pre-diagnostic use (RRpre = 0.874, 95% CI: 0.787–0.970; RRpost = 0.482, 95% CI: 0.359–0.647). 126 Statins use after diagnosis was also found to be a protective factor for metastases (HR=0.78, 95%CI:0.70-0.88), 16 prostate cancer-specific mortality (HR=0.76, 95%CI:0.64-0.89), and all-cause mortality (HR=0.76, 95%CI:0.63-0.91), in a meta-analysis study of 34 prospective cohort studies. 127 As users and non-users of medications may differ in terms of underlying conditions, some issues need to be considered when studying the post-diagnostic effect of medication use on prostate cancer survival. First, the association detected between a certain type of medication and prostate cancer progression use may be confounded by the indication of the medication. Stratifying by or adjusting for the indication might avoid the confounding effect, but should be done with caution as it may introduce a conditional association due to collider bias. 128 Another issue of note is the competing risk of other underlying diseases when the outcome of interest is all-cause mortality among prostate cancer cases. For example, anti-hypertensive drugs may prevent cardiovascular diseases progression and thus lead to a lower risk of all-cause mortality, whilst those drugs may not have a direct impact on prostate cancer progression. 1.4.2 Dietary factors Dietary factors after prostate cancer diagnosis have also been studied. A series of survival analysis of post-diagnostic dietary patterns have been conducted among men with localized/regional prostate cancer in the Health Professionals Follow-Up Study. Specifically, higher fish (HRhighest quartile vs. lowest quartile =0.73, 95% CI: 0.52–1.02) and tomato sauce (HR highest quartile vs. lowest quartile = 0.56, 95% CI: 0.38–0.82) consumption levels were found to be associated with prostate cancer progression, which was defined as any clinical evidence of recurrence or metastasis or prostate cancer death; 129 a consumption of vegetable fat of 10% of total energy intake was associated with a 29% reduced risk of lethal prostate cancer (HR=0.71, 95% CI: 0.51-0.98) and a 26% reduced 17 risk of all-cause mortality (HR=0.74, 95% CI: 0.61-0.88); 130 a 5% energy intake of saturated and a 1% intake of trans fats were associated with a 30% and 25% increased risk of all-cause mortality (HRsaturated fat=1.30, 95% CI:1.05-1.60; HRtrans fats =1.25, 95% CI:1.05-1.49), respectively; 130 any red wine consumption is associated with an decreased risk of lethal prostate cancer (HR=0.50, 95% CI: 0.29- 0.86) and all-cause mortality (HR=0.74, 95%CI: 0.57-0.97). 131 Other large cohorts, with slightly smaller sample size of post-diagnostic cases, also yielded similar results. The Cancer of the Prostate Strategic Urologic Research Endeavor with 1,557 non-metastatic cases observed that cruciferous vegetable intake tended to reduce the risk of prostate cancer progression, while coffee intake among current smokers tended to have an increased risk of progression. 132,133 Physicians’ Health Study with 926 non-metastatic cases suggested that saturated fat and a “Western” pattern of diet were risk factors of prostate cancer specific mortality. 134 Another cohort study examined the alcohol drinking pattern among 829 men diagnosed with at least T2 prostate cancer in 2000 in Alberta, Canada. Post-diagnosis alcohol consumption of more than 2 drinks/day (the recommendation borderline in Canada) was associated with prostate cancer specific death after accounting for competing risk of death due to other causes; and a pre-post diagnosis reduction pattern of alcohol drinking was associated with a reduced risk of prostate cancer specific mortality and recurrence or progression, but not associated with all-cause death. 135 In addition, dietary patterns that represent inflammation and insulin metabolic pathways may also associated with the survival and progression of prostate cancer, as altering systematic inflammatory and metabolic balance may subsequently influence the disease’s course 8-10,136 . Several indices have been developed to quantify the inflammatory potential of the diet, including the Dietary Inflammatory Index (DII) and the empirical dietary index for inflammatory pattern 18 (EDIP). Furthermore, the empirical dietary index for hyperinsulinemia (EDIH) and empirical lifestyle index for hyperinsulinemia (ELIH) have been developed to measure hyperinsulinemia, while the empirical dietary index for insulin resistance (EDIR) and empirical lifestyle index for insulin resistance (ELIR) have been developed to measure insulin resistance. A recent study with non-metastatic prostate cancer cases reported a positive association between these indices and prostate cancer progression, but not with prostate cancer-specific mortality 137 . When interpreting the results of survival analyses examining post-diagnostic risk factors for prostate cancer, it is crucial to consider the endpoint of interest. For studies focusing on lethal prostate cancer (i.e., prostate cancer specific mortality), the identified risk factors are likely to represent the progression etiology for prostate cancer; while for studies focusing on all-cause mortality, the identified risk factors related to other conditions that may contribute to mortality. For example, saturated and trans fats may increase the etiology of cardiovascular diseases, which are a common cause of death among prostate cancer cases. 138 1.4.3 Obesity Evidence of the impact of pre-diagnosis BMI on prostate cancer incidence are mixed but suggests a positive link to aggressive and lethal prostate cancer. 139-141 Potentially, obesity may promote the progression of prostate cancer after tumor initialization, and thus obesity after diagnosis continue to function as a promotor of tumor growth, leading to poor survival. Extensive research on post-diagnostic BMI and prostate cancer progression/survival has had mixed results. 142-147 Langlais et al. suggested that the positive findings might be due to residual confounding caused by misclassification of pathology factors at diagnosis with biopsy – obese 19 patients may more likely to be misclassified regarding their disease severity (clinical Gleason score, stage) than men with normal weight. They found that when adjusting for surgical pathology factors (pathology Gleason score, stage) determined at surgery, the positive association became null. 143 This hypothesis was supported by the fact that other studies without adjusting for pathologic factors reported positive associations 146,147 while those studies controlled for pathologic factors reported null association 142-145 . In addition to residual confounding, studies examining post-diagnostic obesity may also be limited by other deficiencies, complicating the interpretation of the study results. First, higher BMI is significantly associated with reduced PSA level, which may delay the detection of prostate cancer at an early stage. Thus, obese patients may have more advanced disease than the normal weight patients. However, as mentioned above, some large population-based cohorts, may lack surgical, pathological and treatment covariates. Whereas hospital-based studies with more comprehensive data on clinical and pathological covariates, are usually lacking in information on important lifestyle factors, such as smoking. Third, competing risk due to comorbidities of obesity is another issue to be considered, while was rarely adjusted for in previous studies. 139 1.4.4 Physical activity A modest amount of exercise is considered to be a healthy behavior contributes to a patient's physical and mental health and quality of life. In general, physical activity has been consistently found to inversely associated with all-cause mortality, and to be inversely associated with prostate cancer progression or prostate cancer-specific mortality in only some studies. Specificlally, vigorous activities (metabolic equivalent task[MET] ≥ 6) after diagnosis were shown to be 20 associated with a lower risk of prostate cancer specific death whereas no association with non- vigorous activities. 148-151 Amongst different types of activities, exercising, walking and biking were associated with lower risk of prostate cancer specific mortality. 130,152 Recreational activities were found to be protective in some cohorts, 151,153 but to be null in others. 150,152 It is difficult to rule out reverse causality in the association as men with a less severe disease are more physically capable of exercising., Studies on the change patterns of physical activity have shown that patients with a poorer quality of life were less likely to increase or maintain their physical activity level after diagnosis. 154 This might be addressed by introducing a lag time after the prostate cancer diagnosis and before death, but results may vary depends on different choice of lag time. 1.4.5 Smoking Evidence on the effect of smoking on the incidence of total prostate cancer is mixed, but studies have shown a consistent deleterious effect on metastatic prostate cancer or prostate cancer mortality, suggesting that smoking is likely to promote disease progression after tumor initiation (section 1.2.3). Studies on post-diagnostic smoking were mostly conducted within hospital-based prostate cancer patients after radical prostatectomy or radiation therapy. Only two studies were populational-based, and both reported a positive association between smoking at diagnosis (current smokers) as well as a null association between smoking prior to diagnosis (former smokers) and prostate cancer specific mortality. 57,155 These findings were supported by other hospital-based studies. 156-158 21 Similar results were found for prostate cancer recurrence or progression, where current smokers showed elevated risk but former smokers did not. 156,157,159-162 Two issues need to be considered regarding the results of smoking on prostate cancer mortality. First, the assessment of smoking might be subject to measurement error. As cancer patients are likely to cease smoking after cancer diagnosis, most of the above studies assessed smoking status at diagnosis or right after completion of treatment and may miss later changes in smoking behavior. 163 This may lead to a non-differential misclassification of current smokers among men died from prostate cancer and those who did not, resulting in a diluted estimation of the association. Additionally, results from the hospital-based studies may not be generalizable to patients who have not undergone surgery, as the results could be influenced by the treatment effect of surgery or may not apply to patients with different characteristics or disease severity (e.g., those under active surveillance). 1.4.6 Summary Overall, the evidence above provides insights into how certain behavior and lifestyle factors may alter prostate cancer outcomes following a prostate cancer diagnosis. Additionally, it suggests that the risk factors for prostate cancer development and progression might be different. Nonetheless, most of the forementioned studies were conducted in a single race/ethnicity setting and mostly within white populations, with hospital-based studies, or were small in sample size. Therefore, our investigation of post-diagnostic factors among prostate cancer cases in a large, population-based, multi-ethnic study could provide a more comprehensive understanding of the role of lifestyle factors in prostate cancer progression and survival. 22 1.5 References 1. Global Burden of Disease Cancer C, Fitzmaurice C, Abate D, et al. Global, Regional, and National Cancer Incidence, Mortality, Years of Life Lost, Years Lived With Disability, and Disability-Adjusted Life-Years for 29 Cancer Groups, 1990 to 2017: A Systematic Analysis for the Global Burden of Disease Study. JAMA Oncol. 2019;5(12):1749-1768. 2. American Cancer Society. Cancer Facts & Figures 2021. Atlanta: American Cancer Society;2021. 3. Rebbeck TR. Prostate Cancer Disparities by Race and Ethnicity: From Nucleotide to Neighborhood. Cold Spring Harb Perspect Med. 2018;8(9). 4. Mahal BA, Gerke T, Awasthi S, et al. Prostate Cancer Racial Disparities: A Systematic Review by the Prostate Cancer Foundation Panel. Eur Urol Oncol. 2021. 5. Jahn JL, Giovannucci EL, Stampfer MJ. The high prevalence of undiagnosed prostate cancer at autopsy: implications for epidemiology and treatment of prostate cancer in the Prostate-specific Antigen-era. Int J Cancer. 2015;137(12):2795-2802. 6. Howlader N, Noone A, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2017, National Cancer Institute. Bethesda, MD. In:2020. 7. Pernar CH, Ebot EM, Wilson KM, Mucci LA. The Epidemiology of Prostate Cancer. Cold Spring Harb Perspect Med. 2018;8(12):a030361. 8. De Marzo AM, Platz EA, Sutcliffe S, et al. Inflammation in prostate carcinogenesis. Nat Rev Cancer. 2007;7(4):256-269. 9. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell. 2010;140(6):883-899. 10. de Bono JS, Guo C, Gurel B, et al. Prostate carcinogenesis: inflammatory storms. Nat Rev Cancer. 2020;20(8):455-469. 11. De Marzo AM, Marchi VL, Epstein JI, Nelson WG. Proliferative inflammatory atrophy of the prostate: implications for prostatic carcinogenesis. Am J Pathol. 1999;155(6):1985- 1992. 12. van Leenders GJ, Gage WR, Hicks JL, et al. Intermediate cells in human prostate epithelium are enriched in proliferative inflammatory atrophy. Am J Pathol. 2003;162(5):1529-1537. 13. Putzi MJ, De Marzo AM. Morphologic transitions between proliferative inflammatory atrophy and high-grade prostatic intraepithelial neoplasia. Urology. 2000;56(5):828-832. 14. Bostwick DG, Brawer MK. Prostatic intra-epithelial neoplasia and early invasion in prostate cancer. Cancer. 1987;59(4):788-794. 23 15. Steiner GE, Djavan B, Kramer G, et al. The picture of the prostatic lymphokine network is becoming increasingly complex. Reviews in urology. 2002;4(4):171. 16. Steiner GE, Newman ME, Paikl D, et al. Expression and function of pro-inflammatory interleukin IL-17 and IL-17 receptor in normal, benign hyperplastic, and malignant prostate. Prostate. 2003;56(3):171-182. 17. Steiner GE, Stix U, Handisurya A, et al. Cytokine expression pattern in benign prostatic hyperplasia infiltrating T cells and impact of lymphocytic infiltration on cytokine mRNA profile in prostatic tissue. Lab Invest. 2003;83(8):1131-1146. 18. Dehne N, Mora J, Namgaladze D, Weigert A, Brune B. Cancer cell and macrophage cross-talk in the tumor microenvironment. Curr Opin Pharmacol. 2017;35:12-19. 19. Chen Y, Song Y, Du W, Gong L, Chang H, Zou Z. Tumor-associated macrophages: an accomplice in solid tumor progression. J Biomed Sci. 2019;26(1):78. 20. Sato E, Olson SH, Ahn J, et al. Intraepithelial CD8+ tumor-infiltrating lymphocytes and a high CD8+/regulatory T cell ratio are associated with favorable prognosis in ovarian cancer. Proc Natl Acad Sci U S A. 2005;102(51):18538-18543. 21. Curiel TJ, Coukos G, Zou L, et al. Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nat Med. 2004;10(9):942-949. 22. Galon J, Costes A, Sanchez-Cabo F, et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science. 2006;313(5795):1960- 1964. 23. Liu Y, Mikrani R, Xie D, et al. Chronic prostatitis/chronic pelvic pain syndrome and prostate cancer: study of immune cells and cytokines. Fundam Clin Pharmacol. 2020;34(2):160-172. 24. Muthuswamy R, Corman JM, Dahl K, Chatta GS, Kalinski P. Functional reprogramming of human prostate cancer to promote local attraction of effector CD8(+) T cells. Prostate. 2016;76(12):1095-1105. 25. Gallimore AM, Simon AK. Positive and negative influences of regulatory T cells on tumour immunity. Oncogene. 2008;27(45):5886-5893. 26. Dennis LK, Lynch CF, Torner JC. Epidemiologic association between prostatitis and prostate cancer. Urology. 2002;60(1):78-83. 27. Platz EA, Kulac I, Barber JR, et al. A Prospective Study of Chronic Inflammation in Benign Prostate Tissue and Risk of Prostate Cancer: Linked PCPT and SELECT Cohorts. Cancer Epidemiol Biomarkers Prev. 2017;26(10):1549-1557. 24 28. Gurel B, Lucia MS, Thompson IM, Jr., et al. Chronic inflammation in benign prostate tissue is associated with high-grade prostate cancer in the placebo arm of the prostate cancer prevention trial. Cancer Epidemiol Biomarkers Prev. 2014;23(5):847-856. 29. Davidsson S, Fiorentino M, Andren O, et al. Inflammation, focal atrophic lesions, and prostatic intraepithelial neoplasia with respect to risk of lethal prostate cancer. Cancer Epidemiol Biomarkers Prev. 2011;20(10):2280-2287. 30. Vasavada SR, Dobbs RW, Kajdacsy-Balla AA, Abern MR, Moreira DM. Inflammation on Prostate Needle Biopsy is Associated with Lower Prostate Cancer Risk: A Meta- Analysis. J Urol. 2018;199(5):1174-1181. 31. Langston ME, Sfanos KS, Khan S, et al. Why Do Epidemiologic Studies Find an Inverse Association Between Intraprostatic Inflammation and Prostate Cancer: A Possible Role for Colliding Bias? Cancer Epidemiol Biomarkers Prev. 2021;30(2):255-259. 32. De Marzo AM, Platz EA, Sutcliffe S, et al. Inflammation in prostate carcinogenesis. Nat Rev Cancer. 2007;7(4):256-269. 33. Sfanos KS, De Marzo AM. Prostate cancer and inflammation: the evidence. Histopathology. 2012;60(1):199-215. 34. Kay AB. Allergy and allergic diseases. First of two parts. N Engl J Med. 2001;344(1):30- 37. 35. Filella X, Alcover J, Zarco MA, Beardo P, Molina R, Ballesta AM. Analysis of type T1 and T2 cytokines in patients with prostate cancer. Prostate. 2000;44(4):271-274. 36. Van Hemelrijck M, Karagiannis SN, Rohrmann S. Atopy and prostate cancer: Is there a link between circulating levels of IgE and PSA in humans? Cancer Immunol Immunother. 2017;66(12):1557-1562. 37. Li W, Mao S, Tu M, et al. Asthma and the risk of prostate cancer. J Cancer Res Ther. 2018;14(Supplement):S571-S575. 38. Platz EA, Drake CG, Wilson KM, et al. Asthma and risk of lethal prostate cancer in the Health Professionals Follow-Up Study. Int J Cancer. 2015;137(4):949-958. 39. Talbot-Smith A, Fritschi L, Divitini ML, Mallon DF, Knuiman MW. Allergy, atopy, and cancer: a prospective study of the 1981 Busselton cohort. Am J Epidemiol. 2003;157(7):606-612. 40. Severi G, Baglietto L, Muller DC, et al. Asthma, asthma medications, and prostate cancer risk. Cancer Epidemiol Biomarkers Prev. 2010;19(9):2318-2324. 41. Su YL, Chou CL, Rau KM, Lee CT. Asthma and Risk of Prostate Cancer: A Population- Based Case-Cohort Study in Taiwan. Medicine (Baltimore). 2015;94(36):e1371. 25 42. Zhu J, Song J, Liu Z, et al. Association between allergic conditions and risk of prostate cancer: A Prisma-Compliant Systematic Review and Meta-Analysis. Sci Rep. 2016;6:35682. 43. Weiss D, El-Zein M, Rousseau MC, Richard H, Karakiewicz PI, Parent ME. Asthma, allergy and the risk of prostate cancer: results from the Montreal PROtEuS study. Cancer Epidemiol. 2014;38(6):695-699. 44. Arnson Y, Shoenfeld Y, Amital H. Effects of tobacco smoke on immunity, inflammation and autoimmunity. Journal of Autoimmunity. 2010;34(3):J258-J265. 45. Dwivedi S, Goel A, Khattri S, Mandhani A, Sharma P, Pant KK. Tobacco exposure by various modes may alter proinflammatory (IL-12) and anti-inflammatory (IL-10) levels and affects the survival of prostate carcinoma patients: an explorative study in North Indian population. Biomed Res Int. 2014;2014:158530. 46. Moreira DM, Nickel JC, Gerber L, et al. Smoking Is Associated with Acute and Chronic Prostatic Inflammation: Results from the REDUCE Study. Cancer Prevention Research. 2015;8(4):312-317. 47. Dwivedi S, Goel A, Mandhani A, Khattri S, Pant KK. Tobacco exposure may enhance inflammation in prostate carcinoma patients: an explorative study in north Indian population. Toxicol Int. 2012;19(3):310-318. 48. Dwivedi S, Goel A, Khattri S, Mandhani A, Sharma P, Pant KK. Tobacco Exposure by Various Modes May Alter Proinflammatory (IL-12) and Anti-Inflammatory (IL-10) Levels and Affects the Survival of Prostate Carcinoma Patients: An Explorative Study in North Indian Population. BioMed Research International. 2014;2014:158530. 49. Islami F, Moreira DM, Boffetta P, Freedland SJ. A Systematic Review and Meta-analysis of Tobacco Use and Prostate Cancer Mortality and Incidence in Prospective Cohort Studies. European Urology. 2014;66(6):1054-1064. 50. Zu K, Giovannucci E. Smoking and aggressive prostate cancer: a review of the epidemiologic evidence. Cancer Causes Control. 2009;20(10):1799-1810. 51. Park SY, Haiman CA, Cheng I, et al. Racial/ethnic differences in lifestyle-related factors and prostate cancer risk: the Multiethnic Cohort Study. Cancer Causes Control. 2015;26(10):1507-1515. 52. Kenfield SA, Stampfer MJ, Chan JM, Giovannucci E. Smoking and prostate cancer survival and recurrence. Jama. 2011;305(24):2548-2555. 53. Murta-Nascimento C, Romero AI, Sala M, et al. The effect of smoking on prostate cancer survival: a cohort analysis in Barcelona. Eur J Cancer Prev. 2015;24(4):335-339. 54. Polesel J, Gini A, Dal Maso L, et al. The negative impact of tobacco smoking on survival after prostate cancer diagnosis. Cancer Causes Control. 2015;26(9):1299-1305. 26 55. Wilson KM, Markt SC, Fang F, et al. Snus use, smoking and survival among prostate cancer patients. Int J Cancer. 2016;139(12):2753-2759. 56. Gansler T, Shah R, Wang Y, et al. Smoking and Prostate Cancer-Specific Mortality after Diagnosis in a Large Prospective Cohort. Cancer Epidemiol Biomarkers Prev. 2018;27(6):665-672. 57. Riviere P, Kumar A, Luterstein E, et al. Tobacco smoking and death from prostate cancer in US veterans. Prostate Cancer Prostatic Dis. 2020;23(2):252-259. 58. Tchir D, Farag M, Szafron M. Prostate-specific antigen (PSA) screening rates and factors associated with screening in Eastern Canadian men: Findings from cross-sectional survey data. Can Urol Assoc J. 2020;14(7):E319-e327. 59. Li J, Thompson T, Joseph DA, Master VA. Association between smoking status, and free, total and percent free prostate specific antigen. J Urol. 2012;187(4):1228-1233. 60. Seltzer J, Bigby BG, Stulbarg M, et al. O3-induced change in bronchial reactivity to methacholine and airway inflammation in humans. Journal of Applied Physiology. 1986;60(4):1321-1326. 61. Poynter ME, Persinger RL, Irvin CG, et al. Nitrogen dioxide enhances allergic airway inflammation and hyperresponsiveness in the mouse. American Journal of Physiology- Lung Cellular and Molecular Physiology. 2006;290(1):L144-L152. 62. Frampton MW, Boscia J, Roberts Jr NJ, et al. Nitrogen dioxide exposure: effects on airway and blood cells. American Journal of Physiology-Lung Cellular and Molecular Physiology. 2002;282(1):L155-L165. 63. Friedman MS, Powell KE, Hutwagner L, Graham LM, Teague WG. Impact of changes in transportation and commuting behaviors during the 1996 Summer Olympic Games in Atlanta on air quality and childhood asthma. Jama. 2001;285(7):897-905. 64. Weinmayr G, Romeo E, De Sario M, Weiland SK, Forastiere F. Short-term effects of PM10 and NO2 on respiratory health among children with asthma or asthma-like symptoms: a systematic review and meta-analysis. Environmental health perspectives. 2010;118(4):449-457. 65. Nishimura KK, Galanter JM, Roth LA, et al. Early-life air pollution and asthma risk in minority children. The GALA II and SAGE II studies. American journal of respiratory and critical care medicine. 2013;188(3):309-318. 66. Jacquemin B, Sunyer J, Forsberg B, et al. Home outdoor NO 2 and new onset of self- reported asthma in adults. Epidemiology. 2009:119-126. 67. Jerrett M, Shankardass K, Berhane K, et al. Traffic-related air pollution and asthma onset in children: a prospective cohort study with individual exposure measurement. Environmental health perspectives. 2008;116(10):1433-1438. 27 68. Brunst KJ, Leung Y-K, Ryan PH, et al. Forkhead box protein 3 (FOXP3) hypermethylation is associated with diesel exhaust exposure and risk for childhood asthma. Journal of allergy and clinical immunology. 2013;131(2):592-594. e593. 69. Nadeau K, McDonald-Hyman C, Noth EM, et al. Ambient air pollution impairs regulatory T-cell function in asthma. Journal of allergy and clinical immunology. 2010;126(4):845-852. e810. 70. Kohli A, Garcia MA, Miller RL, et al. Secondhand smoke in combination with ambient air pollution exposure is associated with increasedx CpG methylation and decreased expression of IFN-γ in T effector cells and Foxp3 in T regulatory cells in children. Clinical epigenetics. 2012;4(1):1-16. 71. Brandt EB, Kovacic MB, Lee GB, et al. Diesel exhaust particle induction of IL-17A contributes to severe asthma. Journal of Allergy and Clinical Immunology. 2013;132(5):1194-1204. e1192. 72. Al-Ahmadi K, Al-Zahrani A. NO(2) and cancer incidence in Saudi Arabia. Int J Environ Res Public Health. 2013;10(11):5844-5862. 73. Parent M, Goldberg MS, Crouse DL, et al. Traffic-related air pollution and prostate cancer risk: a case-control study in Montreal, Canada. Occup Environ Med. 2013;70(7):511-518. 74. Shekarrizfard M, Valois MF, Goldberg MS, et al. Investigating the role of transportation models in epidemiologic studies of traffic related air pollution and health effects. Environ Res. 2015;140:282-291. 75. Cohen G, Levy I, Yuval, et al. Long-term exposure to traffic-related air pollution and cancer among survivors of myocardial infarction: A 20-year follow-up study. Eur J Prev Cardiol. 2017;24(1):92-102. 76. Datzmann T, Markevych I, Trautmann F, Heinrich J, Schmitt J, Tesch F. Outdoor air pollution, green space, and cancer incidence in Saxony: a semi-individual cohort study. BMC Public Health. 2018;18(1):715. 77. Cohen G, Levy I, Yuval, et al. Chronic exposure to traffic-related air pollution and cancer incidence among 10,000 patients undergoing percutaneous coronary interventions: A historical prospective study. Eur J Prev Cardiol. 2018;25(6):659-670. 78. Cong X. Air pollution from industrial waste gas emissions is associated with cancer incidences in Shanghai, China. Environ Sci Pollut Res Int. 2018;25(13):13067-13078. 79. Morganti G, Gianferrari L, Cresseri A, Arrigoni G, Lovati G. [Clinico-statistical and genetic research on neoplasms of the prostate]. Acta Genet Stat Med. 1956;6(2):304-305. 80. Ostrander EA, Stanford JL. Genetics of prostate cancer: too many loci, too few genes. Am J Hum Genet. 2000;67(6):1367-1375. 28 81. Grönberg H, Damber L, Damber JE, Iselius L. Segregation analysis of prostate cancer in Sweden: support for dominant inheritance. Am J Epidemiol. 1997;146(7):552-557. 82. Schaid DJ, McDonnell SK, Blute ML, Thibodeau SN. Evidence for autosomal dominant inheritance of prostate cancer. Am J Hum Genet. 1998;62(6):1425-1438. 83. Verhage BA, Baffoe-Bonnie AB, Baglietto L, et al. Autosomal dominant inheritance of prostate cancer: a confirmatory study. Urology. 2001;57(1):97-101. 84. Gong G, Oakley-Girvan I, Wu AH, et al. Segregation analysis of prostate cancer in 1,719 white, African-American and Asian-American families in the United States and Canada. Cancer Causes Control. 2002;13(5):471-482. 85. Valeri A, Briollais L, Azzouzi R, et al. Segregation analysis of prostate cancer in France: evidence for autosomal dominant inheritance and residual brother-brother dependence. Ann Hum Genet. 2003;67(Pt 2):125-137. 86. Kiciński M, Vangronsveld J, Nawrot TS. An epidemiological reappraisal of the familial aggregation of prostate cancer: a meta-analysis. PLoS One. 2011;6(10):e27130. 87. Ren ZJ, Cao DH, Zhang Q, et al. First-degree family history of breast cancer is associated with prostate cancer risk: a systematic review and meta-analysis. BMC Cancer. 2019;19(1):871. 88. Abdel-Rahman O. Prostate Cancer Incidence and Mortality in Relationship to Family History of Prostate Cancer; Findings From The PLCO Trial. Clin Genitourin Cancer. 2019;17(4):e837-e844. 89. Chattopadhyay S, Hemminki O, Försti A, Sundquist K, Sundquist J, Hemminki K. Impact of family history of cancer on risk and mortality of second cancers in patients with prostate cancer. Prostate Cancer Prostatic Dis. 2019;22(1):143-149. 90. Brandt A, Bermejo JL, Sundquist J, Hemminki K. Age at diagnosis and age at death in familial prostate cancer. Oncologist. 2009;14(12):1209-1217. 91. Lichtenstein P, Holm NV, Verkasalo PK, et al. Environmental and heritable factors in the causation of cancer--analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med. 2000;343(2):78-85. 92. Mucci LA, Hjelmborg JB, Harris JR, et al. Familial Risk and Heritability of Cancer Among Twins in Nordic Countries. Jama. 2016;315(1):68-76. 93. Freedman ML, Haiman CA, Patterson N, et al. Admixture mapping identifies 8q24 as a prostate cancer risk locus in African-American men. Proc Natl Acad Sci U S A. 2006;103(38):14068-14073. 94. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24 independently affect risk for prostate cancer. Nat Genet. 2007;39(5):638-644. 29 95. Giri VN, Egleston B, Ruth K, et al. Race, Genetic West African Ancestry, and Prostate Cancer Prediction by Prostate-Specific Antigen in Prospectively Screened High-Risk Men. Cancer Prevention Research. 2009;2(3):244-250. 96. Pulst SM. Genetic Linkage Analysis. Archives of Neurology. 1999;56(6):667-672. 97. Smith JR, Freije D, Carpten JD, et al. Major susceptibility locus for prostate cancer on chromosome 1 suggested by a genome-wide search. Science. 1996;274(5291):1371-1374. 98. Carpten J, Nupponen N, Isaacs S, et al. Germline mutations in the ribonuclease L gene in families showing linkage with HPC1. Nat Genet. 2002;30(2):181-184. 99. Saunders EJ, Kote-Jarai Z, Eeles RA. Identification of Germline Genetic Variants that Increase Prostate Cancer Risk and Influence Development of Aggressive Disease. Cancers (Basel). 2021;13(4). 100. Ni Raghallaigh H, Eeles R. Genetic predisposition to prostate cancer: an update. Fam Cancer. 2021. 101. Conti DV, Darst BF, Moss LC, et al. Trans-ancestry genome-wide association meta- analysis of prostate cancer identifies new susceptibility loci and informs genetic risk prediction. Nature Genetics. 2021;53(1):65-75. 102. Zhang YD, Hurson AN, Zhang H, et al. Assessment of polygenic architecture and risk prediction based on common variants across fourteen cancers. Nature communications. 2020;11(1):3353-3353. 103. Momozawa Y, Mizukami K. Unique roles of rare variants in the genetics of complex diseases in humans. Journal of Human Genetics. 2021;66(1):11-23. 104. Penney KL, Pyne S, Schumacher FR, et al. Genome-wide association study of prostate cancer mortality. Cancer Epidemiol Biomarkers Prev. 2010;19(11):2869-2876. 105. Szulkin R, Karlsson R, Whitington T, et al. Genome-wide association study of prostate cancer-specific survival. Cancer Epidemiol Biomarkers Prev. 2015;24(11):1796-1800. 106. Wiklund FE, Adami HO, Zheng SL, et al. Established prostate cancer susceptibility variants are not associated with disease outcome. Cancer Epidemiol Biomarkers Prev. 2009;18(5):1659-1662. 107. Berndt SI, Wang Z, Yeager M, et al. Two susceptibility loci identified for prostate cancer aggressiveness. Nature communications. 2015;6:6889-6889. 108. Eeles RA, Kote-Jarai Z, Giles GG, et al. Multiple newly identified loci associated with prostate cancer susceptibility. Nat Genet. 2008;40(3):316-321. 30 109. Helfand BT, Roehl KA, Cooper PR, et al. Associations of prostate cancer risk variants with disease aggressiveness: results of the NCI-SPORE Genetics Working Group analysis of 18,343 cases. Hum Genet. 2015;134(4):439-450. 110. Dudbridge F. Power and predictive accuracy of polygenic risk scores. PLoS Genet. 2013;9(3):e1003348. 111. Graff RE, Cavazos TB, Thai KK, et al. Cross-cancer evaluation of polygenic risk scores for 16 cancer types in two large cohorts. Nature Communications. 2021;12(1):970. 112. Chen H, Liu X, Brendler CB, et al. Adding genetic risk score to family history identifies twice as many high-risk men for prostate cancer: Results from the prostate cancer prevention trial. Prostate. 2016;76(12):1120-1129. 113. Aly M, Wiklund F, Xu J, et al. Polygenic risk score improves prostate cancer risk prediction: results from the Stockholm-1 cohort study. Eur Urol. 2011;60(1):21-28. 114. Pashayan N, Pharoah PDP, Schleutker J, et al. Reducing overdiagnosis by polygenic risk- stratified screening: findings from the Finnish section of the ERSPC. British Journal of Cancer. 2015;113(7):1086-1093. 115. Na R, Labbate C, Yu H, et al. Single-Nucleotide Polymorphism–Based Genetic Risk Score and Patient Age at Prostate Cancer Diagnosis. JAMA Network Open. 2019;2(12):e1918145-e1918145. 116. Perron L, Bairati I, Harel F, Meyer F. Antihypertensive drug use and the risk of prostate cancer (Canada). Cancer Causes and Control. 2004;15(6):535-541. 117. Grytli HH, Fagerland MW, Fosså SD, Taskén KA. Association between use of β- blockers and prostate cancer-specific survival: A cohort study of 3561 prostate cancer patients with high-risk or metastatic disease. European Urology. 2014;65(3):635-641. 118. Chang PY, Huang WY, Lin CL, et al. Propranolol reduces cancer risk. Medicine (United States). 2015;94(27):e1097. 119. Kaapu KJ, Ahti J, Tammela TLJ, Auvinen A, Murtola TJ. Sotalol, but not digoxin is associated with decreased prostate cancer risk: A population-based case-control study. International Journal of Cancer. 2015;137(5):1187-1195. 120. Zahalka AH, Fram E, Lin W, et al. Use of beta-blocker types and risk of incident prostate cancer in a multiethnic population. Urologic Oncology: Seminars and Original Investigations. 2020;38(10):794.e711-794.e716. 121. Assayag J, Pollak MN, Azoulay L. Post-diagnostic use of beta-blockers and the risk of death in patients with prostate cancer. Eur J Cancer. 2014;50(16):2838-2845. 31 122. Cardwell CR, Coleman HG, Murray LJ, O'Sullivan JM, Powe DG. Beta-blocker usage and prostate cancer survival: a nested case-control study in the UK Clinical Practice Research Datalink cohort. Cancer Epidemiol. 2014;38(3):279-285. 123. Santala EE, Rannikko A, Murtola TJ. Antihypertensive drugs and prostate cancer survival after radical prostatectomy in Finland-A nationwide cohort study. Int J Cancer. 2019;144(3):440-447. 124. Siltari A, Murtola TJ, Talala K, Taari K, Tammela TLJ, Auvinen A. Antihypertensive drug use and prostate cancer-specific mortality in Finnish men. PLoS One. 2020;15(6):e0234269. 125. Fan LL, Xie CP, Wu YM, Gu XJ, Chen YH, Wang YJ. Aspirin Exposure and Mortality Risk among Prostate Cancer Patients: A Systematic Review and Meta-Analysis. Biomed Res Int. 2019;2019:9379602. 126. Zhao X, Xu Z, Li H. NSAIDs Use and Reduced Metastasis in Cancer Patients: results from a meta-analysis. Sci Rep. 2017;7(1):1875. 127. Raval AD, Thakker D, Negi H, Vyas A, Kaur H, Salkini MW. Association between statins and clinical outcomes among men with prostate cancer: a systematic review and meta-analysis. Prostate Cancer Prostatic Dis. 2016;19(2):151-162. 128. Stopsack KH, Greenberg AJ, Mucci LA. Common medications and prostate cancer mortality: a review. World J Urol. 2017;35(6):875-882. 129. Chan JM, Holick CN, Leitzmann MF, et al. Diet after diagnosis and the risk of prostate cancer progression, recurrence, and death (United States). Cancer Causes Control. 2006;17(2):199-208. 130. Richman EL, Kenfield SA, Chavarro JE, et al. Fat intake after diagnosis and risk of lethal prostate cancer and all-cause mortality. JAMA Intern Med. 2013;173(14):1318-1326. 131. Downer MK, Kenfield SA, Stampfer MJ, et al. Alcohol Intake and Risk of Lethal Prostate Cancer in the Health Professionals Follow-Up Study. J Clin Oncol. 2019;37(17):1499-1511. 132. Richman EL, Carroll PR, Chan JM. Vegetable and fruit intake after diagnosis and risk of prostate cancer progression. Int J Cancer. 2012;131(1):201-210. 133. Langlais CS, Chan JM, Kenfield SA, et al. Post-diagnostic coffee and tea consumption and risk of prostate cancer progression by smoking history. Cancer Causes Control. 2021;32(6):635-644. 134. Van Blarigan EL, Kenfield SA, Yang M, et al. Fat intake after prostate cancer diagnosis and mortality in the Physicians' Health Study. Cancer Causes Control. 2015;26(8):1117- 1126. 32 135. Farris MS, Courneya KS, Kopciuk KA, McGregor SE, Friedenreich CM. Post-diagnosis alcohol intake and prostate cancer survival: A population-based cohort study. Int J Cancer. 2018;143(2):253-262. 136. Albanes D, Weinstein SJ, Wright ME, et al. Serum Insulin, Glucose, Indices of Insulin Resistance, and Risk of Prostate Cancer. JNCI: Journal of the National Cancer Institute. 2009;101(18):1272-1279. 137. Langlais CS, Graff RE, Van Blarigan EL, et al. Postdiagnostic Inflammatory, Hyperinsulinemic, and Insulin-Resistant Diets and Lifestyles and the Risk of Prostate Cancer Progression and Mortality. Cancer Epidemiol Biomarkers Prev. 2022;31(9):1760- 1768. 138. Mozaffarian D, Aro A, Willett WC. Health effects of trans-fatty acids: experimental and observational evidence. Eur J Clin Nutr. 2009;63 Suppl 2:S5-21. 139. Cao Y, Ma J. Body Mass Index, Prostate Cancer–Specific Mortality, and Biochemical Recurrence: a Systematic Review and Meta-analysis. Cancer Prevention Research. 2011;4(4):486-501. 140. Zhong S, Yan X, Wu Y, et al. Body mass index and mortality in prostate cancer patients: a dose-response meta-analysis. Prostate Cancer Prostatic Dis. 2016;19(2):122-131. 141. Harrison S, Tilling K, Turner EL, et al. Systematic review and meta-analysis of the associations between body mass index, prostate cancer, advanced prostate cancer, and prostate-specific antigen. Cancer Causes Control. 2020;31(5):431-449. 142. Schiffmann J, Salomon G, Tilki D, et al. Radical prostatectomy neutralizes obesity- driven risk of prostate cancer progression. Urologic Oncology: Seminars and Original Investigations. 2017;35(5):243-249. 143. Langlais CS, Cowan JE, Neuhaus J, et al. Obesity at Diagnosis and Prostate Cancer Prognosis and Recurrence Risk Following Primary Treatment by Radical Prostatectomy. Cancer Epidemiology Biomarkers & Prevention. 2019;28(11):1917-1925. 144. Leal-García M, Canto P, Cárdenas-Cárdenas E, Feria-Bernal G, García-García E, Méndez JP. Overweight and obesity in men with prostate cancer do not constitute risk factors for biochemical recurrence. The Aging Male. 2020;23(5):1283-1288. 145. Vidal AC, Oyekunle T, Howard LE, et al. Obesity, race, and long-term prostate cancer outcomes. Cancer. 2020;126(16):3733-3741. 146. Vidal AC, Howard LE, Sun SX, et al. Obesity and prostate cancer-specific mortality after radical prostatectomy: results from the Shared Equal Access Regional Cancer Hospital (SEARCH) database. Prostate Cancer and Prostatic Diseases. 2017;20(1):72-78. 33 147. Wang LS, Murphy CT, Ruth K, et al. Impact of obesity on outcomes after definitive dose-escalated intensity-modulated radiotherapy for localized prostate cancer. Cancer. 2015;121(17):3010-3017. 148. Kenfield SA, Stampfer MJ, Giovannucci E, Chan JM. Physical activity and survival after prostate cancer diagnosis in the health professionals follow-up study. J Clin Oncol. 2011;29(6):726-732. 149. Dai JY, Wang B, Wang X, et al. Vigorous Physical Activity Is Associated with Lower Risk of Metastatic-Lethal Progression in Prostate Cancer and Hypomethylation in the CRACR2A Gene. Cancer Epidemiol Biomarkers Prev. 2019;28(2):258-264. 150. Guy DE, Vandersluis A, Klotz LH, et al. Total energy expenditure and vigorous-intensity physical activity are associated with reduced odds of reclassification among men on active surveillance. Prostate Cancer Prostatic Dis. 2018;21(2):187-195. 151. Friedenreich CM, Wang Q, Neilson HK, Kopciuk KA, McGregor SE, Courneya KS. Physical Activity and Survival After Prostate Cancer. Eur Urol. 2016;70(4):576-585. 152. Bonn SE, Sjölander A, Lagerros YT, et al. Physical activity and survival among men diagnosed with prostate cancer. Cancer Epidemiol Biomarkers Prev. 2015;24(1):57-64. 153. Wang Y, Jacobs EJ, Gapstur SM, et al. Recreational Physical Activity in Relation to Prostate Cancer-specific Mortality Among Men with Nonmetastatic Prostate Cancer. Eur Urol. 2017;72(6):931-939. 154. Stone CR, Courneya KS, McGregor SE, Li H, Friedenreich CM. Determinants of changes in physical activity from pre-diagnosis to post-diagnosis in a cohort of prostate cancer survivors. Support Care Cancer. 2019;27(8):2819-2828. 155. Gong Z, Agalliu I, Lin DW, Stanford JL, Kristal AR. Cigarette smoking and prostate cancer-specific mortality following diagnosis in middle-aged men. Cancer Causes Control. 2008;19(1):25-31. 156. Steinberger E, Kollmeier M, McBride S, Novak C, Pei X, Zelefsky MJ. Cigarette smoking during external beam radiation therapy for prostate cancer is associated with an increased risk of prostate cancer-specific mortality and treatment-related toxicity. BJU International. 2015;116(4):596-603. 157. Pantarotto J, Malone S, Dahrouge S, Gallant V, Eapen L. Smoking is associated with worse outcomes in patients with prostate cancer treated by radical radiotherapy. BJU International. 2007;99(3):564-569. 158. Moreira DM, Aronson WJ, Terris MK, et al. Cigarette smoking is associated with an increased risk of biochemical disease recurrence, metastasis, castration-resistant prostate cancer, and mortality after radical prostatectomy. Cancer. 2014;120(2):197-204. 34 159. Joshu CE, Mondul AM, Meinhold CL, et al. Cigarette smoking and prostate cancer recurrence after prostatectomy. J Natl Cancer Inst. 2011;103(10):835-838. 160. Rieken M, Shariat SF, Kluth LA, et al. Association of Cigarette Smoking and Smoking Cessation with Biochemical Recurrence of Prostate Cancer in Patients Treated with Radical Prostatectomy. European Urology. 2015;68(6):949-956. 161. PICKLES T, LIU M, BERTHELET E, KIM-SING C, KWAN W, TYLDESLEY S. THE EFFECT OF SMOKING ON OUTCOME FOLLOWING EXTERNAL RADIATION FOR LOCALIZED PROSTATE CANCER. Journal of Urology. 2004;171(4):1543-1546. 162. Merrick GS, Butler WM, Wallner KE, Galbreath RW, Lief JH, Adamovich E. Effect of cigarette smoking on biochemical outcome after permanent prostate brachytherapy. International Journal of Radiation Oncology*Biology*Physics. 2004;58(4):1056-1062. 163. Westmaas JL, Newton CC, Stevens VL, Flanders WD, Gapstur SM, Jacobs EJ. Does a Recent Cancer Diagnosis Predict Smoking Cessation? An Analysis From a Large Prospective US Cohort. J Clin Oncol. 2015;33(15):1647-1652. 35 Chapter Two: Specific Aims Aim 1: To identify novel genetic risk variants for prostate cancer in multi-ancestry populations: Aim 1a: To identify novel genetic risk variants for prostate cancer by conducting GWAS in UK Biobank European ancestry population. Aim 1b: To identify ancestry-specific novel genetic variants for prostate cancer by conducting ancestry-specific and multi-ancestry GWAS meta-analyses. Aim 1c: To develop a genetic risk score to characterize prostate cancer risk in multi- ancestry populations. Aim 2: To assess the association between clonal hematopoiesis and prostate cancer. Aim 3: To examine the effects of atopic allergic conditions on the risk of prostate cancer in multi- ethnic population, in particular on disease severity and survival. 36 Chapter Three: Characterizing prostate cancer risk through multi- ancestry genome-wide discovery of 187 novel risk variants 3.1 Background In men, prostate cancer (PCa) is the most frequently diagnosed non-skin cancer globally 1 . Variation in PCa incidence is observed across populations globally, with the highest rates observed in men of African ancestry 1 . PCa risk is heavily influenced by genetic factors, with 278 genetic risk variants identified through GWAS 2-13 . While the majority of samples in PCa GWAS have been of European ancestry, multi-ancestry analysis has been demonstrated to improve discovery of novel risk variants 14 and enhance genetic risk prediction for PCa across populations 2 . Previously, in a multi-ancestral genome-wide association (GWA) meta-analysis of 107,247 prostate cancer cases and 127,006 controls, we have identified 269 independent genetic risk variants of prostate cancer and developed a genetic risk scores (GRS) based on these risk variants. 2 However, the transferability and clinical value of GRS across populations remains limited due to an imbalance in genetic studies across ancestrally diverse populations. In this study, we conducted a multi- ancestry genome-wide association study (GWAS) of 156,319 prostate cancer (PCa) cases and 788,443 controls of European, African, Asian, and Hispanic men, reflecting a 57% increase in the number of non-European cases over previous PCa GWAS. We aimed to identify novel susceptibility loci for prostate cancer, and to develop an improved GRS to characterize prostate cancer risks across ancestry groups. 37 3.2 Methods Study subjects in the multi-ancestry GWAS. The multi-ancestry meta-analysis included 107,247 cases and 127,006 controls that were part of a previous multi-ancestry meta-analysis of the following studies/consortia: Prostate Cancer Association Group to Investigate Cancer-Associated Alterations in the Genome and Collaborative Oncological Gene-Environment Study Consortium [PRACTICAL iCOGS], the Elucidating Loci Involved in Prostate Cancer Susceptibility OncoArray Consortium [ELLIPSE OncoArray], the United Kingdom GWAS [UK GWAS1 and UK GWAS2], Cancer of the Prostate in Sweden [CAPS1 and CAP2], the National Cancer Institute [NCI] Prostate cancer Genome-wide Association Study of Uncommon Susceptibility loci study [PEGASUS], the NCI Breast and Prostate Cancer Cohort Consortium [BPC3], the ProHealth GWAS Study within the Research Program on Genes, Environment and Health Kaiser Permanente cohort [ProHealth Kaiser GWAS], the African Ancestry Prostate Cancer Consortium [AAPC GWAS], BioBank Japan, GWAS of prostate cancer in Latinos [LAPC GWAS] and Japanese [JAPC GWAS] in the Multiethnic Cohort Study [MEC] and the Ghana Prostate Study [GPS]) 2 . The present study included an additional 49,072 cases and 661,437 controls from the UK Biobank, the FinnGen study, the Electronic Medical Records and Genomics (eMERGE) Network, the BioVU Biobank, the BioMe Biobank, the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO), the MD Anderson prostate cancer study (MD Anderson), the California and Uganda Prostate Cancer Study (CA UG), the VA Million Veteran Program (MVP), and the Maryland Prostate Cancer Case-Control Study (NCI-MD) (Supplementary Table 1). In total, there were 122,188 cases and 604,640 controls of European ancestry, 19,391 cases and 61,608 controls of African ancestry, 10,809 cases and 95,790 controls of Asian ancestry, and 3,931 cases and 26,405 controls of Hispanic ethnicity. The 38 effective sample size for each population was calculated using the formula Neff = 4/(1/Ncases + 1/Ncontrols). Genotyping and imputation in the multi-ancestry GWAS. The details of study design, inclusion and exclusion criteria, genotyping, imputation and quality control procedures in each study/consortium are provided in Supplementary Tables 1 and 2 in the previous paper. All participants provided written informed consent, and study protocols were approved by the Institutional Review Board at each study site. Briefly, samples and variants were excluded with a corresponding study-specific sample or genotyping call rate <95%. Imputation in each study was performed using Minimac3/Minimac4 15 , Impute2 16 , Eagle2 17 , or Beagle 4.1 18 under the 1000 Genome phase 3 19 , the NHLBI Trans-Omics for Precision Medicine (TOPMed) Consortium freeze 5 20 , Haplotype Reference Consortium (HRC), UK10K 21 , or SISu v3 imputation 18 panels. For most studies, single nucleotide polymorphisms (SNPs) and small insertion/deletions (indels) with MAF≥0.1% and imputation quality scores ≥0.3 were included in the association analysis. A higher cutoff of imputation quality score was applied in FinnGen (>0.6) and BioMe (≥0.8). Statistical analysis for GWAS. Genetic ancestry was estimated using principal component analysis in each study based on uncorrelated SNPs. In total, 42,428,922 variants (SNPs and indels) were examined for association with PCa risk using logistic regression adjusting for age, sub-study (if applicable, see Supplementary Table 1) and up to 10 principal components. Statistical software or programs used were R, PLINK, or C++ for different studies. Per-allele odds ratios (OR) and standard errors from 39 individual studies were combined by a fixed-effects inverse-variance weighted meta-analysis using METAL in ancestry-specific analyses as well as across all four ancestry groups to obtain multiethnic estimates of effects. Heterogeneity of effect sizes across ancestries were examined by the statistic I 2 with corresponding tests of significance. The genomic inflation factors (λ) were calculated in each study/consortium and within each population (Supplementary Table 1). Each inflation factor was then rescaled to λ 1000 , which represents the inflation factor for an equivalent study of 1,000 cases and 1,000 controls: λ 1000 = 1 + 500 (λ−1) ∑ ( 1 n k k + 1 m k ) −1 , where n k : the number of cases for study k, m k : the number of controls for study k. Risk variants identification. Genome-wide significant associations were defined as variants with P<5x10 -8 in the multi- ancestry meta-analysis. To identify and update independent index risk variants for PCa in the newly identified and previously known risk regions, we implemented a forward-selection conditional analysis approach using a multi-population Joint Analysis of Marginal summary statistic (mJAM). Within each region, the forward selection process started with a model containing the variants with the most significant muti-ancestry marginal P value, and additional variants were added if they were independent of the selected variants (LD R 2 <0.1 in all four populations). Variants with a conditional multi-ancestry P<5x10 -8 were retained in the model. Imputation quality scores of all individual studies were checked for all selected risk variants. 40 Genome-wide significant variants were considered “novel” if they were not in linkage disequilibrium (LD) with any previously known risk variants in any of the four populations, and remained genome-wide significant after conditioning on nearby known risk variants. Previously known variants were 1) dropped if their marginal P values were below the genome-wide significance threshold, 2) replaced by a correlated new lead variant with a more significant conditional P value, or 3) not replaced. Genetic Risk Score (GRS) construction. Based on the assumption of a log-additive model, we constructed a genetic risk score (GRS) from the summed risk allelic dosages weighted by the per-allele log-odds ratios in the marginal model for independent variants and in the conditional model for the variants in the same region. Thus, for each individual j we derived: where N : Number of variants, ij g : Allele dose at variant i for individual j, i  : Per-allele log-odds ratio of variant i. GRS was constructed for the 451 risk variants, and also for risk variant sets reported in previous PCa GWAS meta-analyses: (1) N=269 variants reported in a multi-ancestry study (107,247 cases / 127,006 controls) 2 , (2) N=181 variants reported in European (25,723 cases / 26,274 controls) 22 , African (10,202 cases/ 10,810 controls) 23 and Asian (3,000 cases/ 4,394 controls) 6 ancestry- specific studies, respectively, and (3) N=100 variants reported in a multi-ancestry study (43,303 cases / 43,737 controls) 8 . ij N i i j g Score  = = 1  41 Discriminative improvement of GRS. To visualize the improvement of predictive ability of PCa GRS over time with the increasing number of risk variants included, we categorized the distributions of previous GRS (GRS100, GRS 181, GRS269) and the current GRS (GRS451) into quintiles ([0-20%], (20-40%], (40-60%], (60-80%], and (80-100%]) based on the distribution of the score in controls for each study or consortium. We included studies of European ancestry (OncoArray study, PRACTICAL iCOGS, UK Biobank, and MVP), African ancestry (OncoArray study, AAPC GWAS, CA UG, Ghana Prostate Study, NCI- MD, and MVP), Asian ancestry (OncoArray study and JAPC GWAS), and Hispanic ethnicity (OncoArray study and LAPC GWAS). In the combined sample from all four populations, we used Sankey diagrams to visualize the change in risk categorization from the previous GRS to the subsequent GRS among controls and cases, respectively. In the Sankey diagrams, we used each node to represent a GRS category, with the height of each node proportional to the percentage of that GRS category among all cases or all controls. The flow widths connecting GRS categories were proportional to the percentage of individuals from a specific quintile in the previous GRS who were then categorized into the corresponding quintile in the next GRS. To quantify the discriminative ability improvement of GRS offered by inclusion of additional risk variants, we calculated continuous-based net reclassification improvement (NRI) in our GWAS discovery sample 24 . For each study, we calculated NRI comparing a risk model with age only (adjusted for sub-studies, and top 10 principal components) to risk models with additional inclusion of GRS100, GRS181, GRS269, and GRS451, respectively. Additionally, we calculated NRI comparing the GRS451 model to the GRS269 model to show the discriminative ability improvement 42 of the current GRS relative to last GRS. The 95% CIs for NRI were estimated using 1,000 bootstrap replications. Combined estimates of NRI for each population were estimated using the inverse variance weighted meta-analysis. GRS association analysis. The risk of PCa was estimated for the per standard deviation (SD) GRS change and for each percentile category of the GRS: [0-10%], (10-20%], (20-30%], (30-40%], (40-60%], (60-70%], (70-80%], (80-90%], and (90-100%]. Additional analysis was performed to obtain the risk of PCa for the top 1% ((99-100%]). We reported the GRS associations using the median quintile (40-60%] category (Supplementary Table 4). The mean and SD, and the GRS categories were determined by the observed distribution among controls for each population. We applied the conditional multi- ancestry effect estimates from the overall meta-analysis to calculate GRS for individuals from studies mentioned above. In each study, logistic regression was performed to estimate the odds ratio (OR) and 95% confidence interval (CI) corresponding to per SD change of GRS or each GRS category, adjusted for age, sub-study (if applicable), and up to 10 principal components. All GRS analyses were adjusted for age since age is an important PCa risk factor, and the GRS is associated with PCa diagnosis age. Within each population, the associations of GRS with PCa risk were meta- analyzed across individual studies using a fixed-effect inverse-variance-weighted method. GRS association in replication and overall samples. We validated the GRS performance in independent samples that were not part of the GWAS discovery, including the Michigan Genomics Initiative (MGI; European: 3,244 cases, 10,537 controls; African: 189 cases, 450 controls), Mass General Brigham Biobank (MGB; European: 1868 cases, 10,980 controls; African: 85 cases, 471 43 controls), Men of African Descent and Carcinoma of the Prostate (MADCaP; African: 2,505 cases, 2,160 controls), and Estonian Biobank (EstBB; European: 2,352 cases, 28,546 controls). Details of study population, genotyping and imputation were described in Supplementary Tables 1 and 2. GRS451 and GRS269 were constructed and weighted by the multi-ancestry conditional weights. Odds ratios per SD and for each decile were estimated within study population using logistic regression adjusted for age, sub-study (if applicable), and up to 10 principal components. To illustrate the variation in the GRS451 association with PCa risk across studies, we estimated the OR per SD within each sub-study by ancestry. The sub-studies included in this analysis were limited to those included in our GWAS discovery or validation sample with case or control counts of at least 100. The combined estimates for each ancestry group and in the multi-ancestry overall sample (both GWAS discovery and replication studies) were obtained using the inverse-variance weighted fixed effects meta-analysis. To account for the potential over-fitting of the GRS in the GWAS discovery samples and to maximize statistical power, these estimates were reported as main GRS effects and used for subsequent calculations for absolute risk by population. Bias correction for GRS. The GRS weights for the 451 risk variants are possibly inflated because they are the effect estimates from the original GWAS that detected them, sometimes referred to as “winner’s curse”. To account for the potential bias, we applied the bias-reduction procedure based on conditional likelihood 25 to obtain the corrected effect estimates of those selected risk variants. GRS by Age and Disease Aggressiveness. 44 We investigated the association of GRS with PCa risk stratified by age and its association with disease aggressiveness in GWAS discovery and replication studies, respectively. In age-stratified analysis, cases and controls were both stratified into two age groups (age ≤55 vs. age >55 years). PCa was defined as aggressive if one or more of the following criteria were met: tumor stage T3/T4, regional lymph node involvement, metastatic disease (M1), Gleason score ≥ 8, PSA level ≥ 20 ng/mL or PCa as the underlying cause of death. Non-aggressive PCa was defined as PCa without aggressive features and meeting one or more of the following criteria: Gleason score ≤ 7.0, PSA < 20 ng/mL, and stage ≤ T2. Among men aged ≤ 55 years, proportions of aggressive prostate cancer cases of all cases were 36.7%, 32.1%, 53.8%, and 36.3% in European, African, Asian, and Hispanic men, respectively; while the proportions for men aged < 55 years were 35.6%, 39.9%, 51.5%, and 36.6%. Logistic regressions were performed with PCa status (non-aggressive vs. control, aggressive vs. control, or aggressive vs. non-aggressive) as the outcome and per SD GRS or GRS categories as the independent predictors, adjusting for age, sub-study (if applicable), and up to 10 principal components. Ancestry-specific GRS estimates were obtained via an inverse- variance weighted fixed effects meta-analysis performed within each population. Heterogeneity between stratum was assessed via a Q-statistic between effect estimates with corresponding tests of significance. Impact of prostate-specific antigen (PSA) screening on PCa GWAS. We compared the 128 PSA variant reported in the latest PSA GWAS 26 to the 451 PCa risk variants and found 50 overlapping variants [in high LD (R 2 >0.8) or identical index variant]. Three of the variants (2 of which overlapped with the PSA variants) are near the KLK3 gene, which encodes the PSA protein and are very strongly associated with PSA level. For the 48 overlapping variants 45 (removing KLK3), it is currently difficult to differentiate whether they are PCa risk variants, PSA variants or both. To better understand the likelihood of these variants being identified as the result of altering PSA levels, leading to biopsy and a PCa diagnosis, we examined their aggregate effect on disease aggressiveness in our GWAS discovery samples. Additionally, we removed the 48 potential PSA variants (and 3 KLK3 variants) from the PCa GRS (with 400 variants) and examine the association with aggressive versus non-aggressive PCa in the multi-ancestry sample. To account for the multiple comparisons being made in our sub-group analyses (in total 20 independent tests), we applied Bonferroni correction to the significance level (0.05/20=0.0025). Age-specific absolute risk estimation. Absolute risk for a given age for each GRS percentile and each population has been described previously. 2,27-30 The approach constrains the GRS-specific absolute risks for a given age to be equivalent to the age-specific incidence for the entire population while accounting for competing causes of death. Specifically, for a given population and a given GRS risk category k (e.g. 50%, 80%), the absolute risk by age t is computed as: 𝐴𝑅 𝑘 (𝑡 ) = ∑ 𝑃 𝑁𝐷 (𝑡 ) 𝑡 0 𝑆 𝑘 (𝑡 )𝐼 𝑘 (𝑡 ). This calculation consists of three components: (1) 𝑃 𝑁𝐷 (𝑡 ) is the probability of not dying from another cause of death by age t using age-specific mortality rates, 𝜇 𝐷 (𝑡 ): 𝑃 𝑁𝐷 (𝑡 ) = exp[− ∑ 𝜇 𝐷 (𝑡 − 1) 𝑡 0 ]. Age-specific mortality rates are provided from a reference cohort. (2) 𝑆 𝑘 (𝑡 ) is the probability of surviving PCa by age t in the GRS category k and uses the PCa incidence by age t for category k: 𝑆 𝑘 (𝑡 ) = exp[− ∑ 𝐼 𝑘 (𝑡 − 1) 𝑡 0 ]. 46 (3) The PCa incidence by age t for GRS category k is 𝐼 𝑘 (𝑡 ) and is calculated by multiplying the population PCa incidence for the reference category, 𝐼 0 (𝑡 ) and the corresponding risk ratio for GRS category k, as estimated from the odds ratio obtained from the population-specific individual- level GRS analysis as described above: 𝐼 𝑘 (𝑡 ) = 𝐼 0 (𝑡 )exp (𝛽 𝑘 ). PCa incidence for age t for the reference category, 𝐼 0 (𝑡 ), is obtained by constraining the weighted average of the population cancer incidences for the GRS categories to the population age-specific PCa incidence, 𝜇 (𝑡 ). 𝐼 0 (𝑡 ) = 𝜇 (𝑡 ) ∑ 𝑓 𝑘 𝑆 𝑘 (𝑡 −1) 𝐾 ∑ 𝑓 𝑘 𝑆 𝑘 (𝑡 −1)exp (𝛽 𝑘 ) 𝐾 , where 𝑓 𝑘 is the frequency of the GRS category k with 𝑓 𝑘 = 0.1 for all non-reference categories in our primary GRS analysis by deciles (e.g. [0-1%], (1-2%], (2-3%], etc.). Absolute risks were calculated iteratively following the above steps. For each ethnic group, absolute risks by age t were calculated using age-specific PCa incidence, 𝜇 (𝑡 ), and age-specific mortality rates, 𝜇 𝐷 (𝑡 ), from the Surveillance, Epidemiology, and End Results (SEER) Program (2014-2018) 31,32 . Variant annotation. Lead variants were annotated for of indicators of functionality according to a framework described previously 2 , and incorporating additional datasets. Gene-based information was obtained using wANNOVAR 33 , with additional manual review of putative exonic variants. Chromatin Immunoprecipitation Sequencing (ChIP-Seq) peaks analysed through a harmonised pipeline were obtained from the Cistrome Data Browser (http://cistrome.org/db/) 34 for the PCa cell-lines LNCaP, PC3 and VCaP and prostate epithelium cell-line PrEC, and converted to GRCh37 reference assembly co-ordinates in R using rtracklayer v.1.42.2 liftOver 35 . Replicate datasets for individual marks were included where sufficient high-quality datasets were available. Peak data 47 were obtained for open chromatin (DNase-Seq and ATAC-seq), histone modifications (H3K27Ac, H3K9Ac, H3K4me1, H3K4me2 and H3K4me3), and transcription factor binding (AR, ARv7, ARID1A, ASH2L, BRD2, BRD3, BRD4, CREB1, CTCF, DAXX, ETF1, E2F1, ERG, ETS1, ETV1, FOXA1, GABPA, GATA2, GRHL2, HIF1A, HNF4G, HOXB13, JUND, KDM1A, MYC, NKX3.1, PIAS1, POLR2A, POU2F1, RUNX2, SUMO2, TCF7L2, TLE3, TRIM24, TRIM28, VDR, ZFX, ZMYND8). Data for significant variant-gene pairs for differential gene expression (eQTLs) in three prostate tissue cohorts (GTEx v8 36 , https://gtexportal.org/home/datasets, normal prostate tissue, n=221; TCGA PRAD, https://portal.gdc.cancer.gov/, prostate adenocarcinoma, n=359; refZ 37 , https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000985.v1.p1, tumor-adjacent normal prostate tissue, n=471) were obtained as described previously 2 . All significantly associated genes at False Discovery Rate (FDR) ≤0.05 identified were reported for each lead variant. Data for significant variant-gene pairs for differential gene splicing (sQTLs) were obtained for two prostate tissue cohorts. sQTLs for normal prostate tissue (n=221) were downloaded for GTEx v8 from the GTEx portal and converted to GRCh37 reference assembly co-ordinates in R using rtracklayer v1.42.2 liftOver. sQTLs for TCGA PRAD (n=485) were obtained from the CancerSplicingQTL database (http://www.cancersplicingqtl-hust.com/) 38 . All genes significantly associated with alternative splicing in the respective datasets were reported for each lead variant. 48 Fitting prediction models of gene expression in prostate tissues. To perform a transcriptome-wide association study (TWAS) using predicted gene expression, we fitted predictive models using genotype and mRNA measurements from samples of normal prostate in GTEx v8 (n=221) 39 and histologically normal prostate in refZ (n=471) 37 . We performed quality control on genotype data and kept only biallelic SNPs with MAF ≥0.01, HWE P >5 x e -5 , imputation quality INFO >0.6, and were annotated in HapMap3. Using the FUSION pipeline, we estimated cis-h2g using QC’d genotypes within 1 Mb flanking the gene body (i.e., ±500 kb transcription start and stop sites) 40 . For GTEx expression data, we adjusted expression models using eQTL covariates described in reference 39 , which included 5 genotyping principal components, 30 PEER factors 41 , and two binary indicators for sequencing protocol and platform. For expression data in refZ 37 , we adjusted expression models for histologic characteristics, percent lymphocytic population, percent epithelium present, and 14 gene expression principal components, which were defined in refZ. We limited downstream model fitting to genes whose expression levels exhibited evidence of genetic control by testing for nonzero cis-heritability (P <0.01) using GCTA 42 . To build prediction models of expression, we fit penalized linear models using a modified version of the FUSION software which included SuSiE 43 . TWAS and PWAS using predicted gene and protein expression levels. To perform downstream TWAS, we used the FUSION software 40 to integrate our fitted prostate expression models together with meta-analyzed GWAS summary statistics. In addition to our fitted models of prostate expression, we also downloaded previously described prediction models of gene expression in prostate adenocarcinoma samples from TCGA (n=468; http://gusevlab.org/projects/fusion/) 44 . To test for association with genetically predicted levels of 49 protein expression in plasma with PCa risk, we downloaded previously described prediction models fitted using the INTERVAL study (N=3301; https://www.mancusolab.com/pwas/) 45 . In total we performed m=19,352 association tests (m_GTEx=5063, m_Mayo=8632, m_TCGA=4664, m_INTERVAL=993). We used a per-reference panel Bonferroni adjustment to determine transcriptome- or proteome-wide significance (TWAS P < 0.05 / m_study). To quantify the extent to which novel risk regions identify from TWAS replicate in larger GWAS, we also performed TWAS and PWAS using a smaller, previously published meta-analyzed GWAS summary statistics of PCa (N=234,253) 2 . A region exhibiting TWAS/PWAS significant signal was determined to be novel if it did fall within 250kb of a sentinel GWAS variant. 3.3 Results We included 122,188/604,640 (cases/controls) men of European ancestry, 19,391/61,608 of African ancestry, 10,809/95,790 of Asian ancestry, and 3,931/26,405 of Hispanic ethnicity in the multi-ancestry GWAS meta-analysis,. Studies, genotyping, quality control and association testing methods are described in Methods. Case sample size was increased by 43% in European, 87% in African, 26% in Asian and 45% in Hispanic groups (with a corresponding effective sample size 128% in each population accounting for controls), compared to previous multi-ancestry GWAS analyses 2 . We performed a fixed-effect meta-analysis within each ancestry group and meta- analyzed the ancestry-specific GWAS to obtain multi-ancestry GWAS estimators. The genomic inflation statistic (λ) was 1.158 in the multi-ancestry GWAS and ranged from 1.053 in Asian to 1.169 in European ancestry studies (Supplementary Table 1); the corresponding meta-analysis λ1000 (scaled to a sample size of 1,000 cases and 1,000 controls) was 1.001. 50 Overall, 42,428,922 variants were examined for association with PCa risk, with 55,241 variants reaching genome-wide significance (P<5.0x10 -8 ). To identify independent risk variants, we implemented a forward-selection conditional analysis using multi-population Joint Analysis of Marginal summary statistics (mJAM) 2,46 . We identified 451 independent risk variants for PCa that were genome-wide significant in multi-ancestry or ancestry-specific analyses, including 187 that were previously unreported (Fig. 1, Supplementary Tables 2), of which 61 were within 800 kb of a known variants but remained genome-wide significant after conditioning on nearby known variants. Of the 451 variants, 150 were known risk variants that were replaced by a more significant lead variant, while 114 remained the lead risk variant in the region. Eighteen variants previously reported as PCa risk variants were dropped because they did not reach genome-wide significance (Supplementary Table 2). Of the 451 risk variants, 429(95%) in European, 411(91%) in African, 377(84%) in Asian and 424(94%) in Hispanic populations had minor allele frequencies (MAF)>1% (eFig. 1) and 339(75%), 47(10%), 42(9%), and 9(2%) were genome-wide significant, respectively (Fig. 2a). Of these, nineteen (European), five (African) and three (Asian) were population-specific risk variants with MAF 1% in all other populations (eFig. 1). For variants with a MAF>1% in all populations (n=370), 369, 247, 208, and 125 were nominally significant in European, African, Asian and Hispanic populations, respectively (Fig. 2b). The correlation between effect sizes for common variants (MAF>1%) was observed across populations in Fig. 2c, with an R=0.73 for European versus African ancestry (398 variants), R=0.58 for European versus Asian ancestry (371 variants), and R=0.72 for European ancestry versus Hispanic men (414 variants). While the effect sizes for all variants were correlated across population, but they were not as strongly correlated as the 51 common ones (Supplementary Figure 1). Heterogeneity in effect size was nominally significant (Pheterogeneity<0.05) for 78 variants (21%) and for 23 variants after Bonferroni-correction (Pheterogeneity<1.4x10 -4 ), with the largest average effect size in Asian men (ORavg=1.11) followed by European ancestry (ORavg =1.09), African ancestry (ORavg =1.08), and Hispanic men (ORavg =1.08). Of the 451 variants, 28 (6.2%) directly alter protein structure. We detected a novel association with a population-specific frameshift deletion in the C9orf152 gene (European) and previously reported frameshift deletions in ANO7 (African 47 ) and CHEK2 (European 2 ) and a frameshift insertion in FAM111A (European 4 ). The lead variants include 24 missense substitutions representing previously reported variants within ANO7 (three lead variants 4 ), CDKN1B, CHEK2, COL23A1, HOXB13, INCENP, KLK3, POGLUT3, RASSF6, RFX7 and SUN2, replacement lead variants in FAM118A, INHBB and SPDL1, novel associations in MMAB, PIM1, RPA1, SERPINA1, SIM2, SYTL1 and ZBTB42, and a second missense risk variant in RASSF6). Among the new genes implicated in PCa risk, expression of SIM2, a transcription factor, has been shown to discriminate PCa and non-cancerous tumor tissue 48 and associated with poorer survival 49 , while PIM1 is a serine/threonine kinase overexpressed in PCa 50 , shown to modulate AR transcriptional activity through phosphorylation 51 and be a co-activator of c-MYC 52 . Many lead variants were also implicated in regulation of gene expression in prostate tissues and cell-lines (Methods). Seventy-four variants (16.4%), including 19 novel associations, were located within regions of open chromatin, chromatin modifications consistent with regulatory elements, situated within transcription factor binding sites overlapping an association for differential gene 52 expression or splicing, providing strong support for biological functionality. Candidate functional variants include rs1858800, correlated with expression of ZFXH3, a gene frequently somatically mutated in PCa 53 ; rs10499188, correlated with expression of SLC2A12, a gene encoding a glucose transporter expressed in PCa cell-lines but not benign prostatic hyperplasia 54 and regulated by AR signaling 55 , and rs79186742, correlated with expression of BARX2, a homeobox transcription factor associated with poor prognosis for a range of solid tumors 56 . Overall, 219 of the 451 lead variants (48.6%) overlap with significant associations for differential expression in prostate tissues (Methods) of 439 distinct genes (eQTLs), while 69 (15.3%) correlate with significant associations for alternative splicing of 95 unique genes (sQTLs). Of the 439 differentially expressed genes, 204 (46.5%) had not been implicated as candidate mediators of PCa risk by the previous panel of 269 PCa risk variants 2 and were established through the identification of additional novel risk variants and replacement of lead variants. To further explore the molecular mechanisms underlying PCa risk, we performed transcriptome- (TWAS) and proteome-wide association studies (PWAS) 40,57,58 using predicted gene expression and protein levels from multiple prostate tissue 37,39,44 and plasma 45 studies (Methods). Across 19,352 tests performed, we identified 746 associations across 528 genes and 230 genomic regions. Of the 746 associations, the greatest contribution was from predicted expression in histologically normal prostate tissue (351/746) 37 . However, this is likely due to the larger reference panel sample size and, thus, number of association tests performed (ANOVA P>0.05). Of the 451 genomic risk regions identified through GWAS, 237 colocalized within 250Kb of transcriptome- or proteome- wide significant associations, which is consistent with previous large-scale TWAS investigations 53 in PCa risk 59,60 . Of the 230 TWAS/PWAS genomic risk regions identified, 45 did not colocalize within 250Kb of the 451 genome-wide significant variants, suggesting that increasing GWAS sample sizes will continue to identify novel risk regions. The predictive ability of the GRS for PCa has improved with the identification of additional risk variants 2-6,8 . We compared the performance of GRSs based on past marker sets (n=100 8 , 181 5,6,22 , 269 2 ) to the current set of 451 risk variants, with GRSs constructed by summing the risk allele dosage, weighted by the multi-ethnic per-allele log-ORs estimated from the current meta-analysis (Methods). With the discovery of more risk variants, there is greater stability in the assignment of unaffected men to GRS categories; 58% of men in the top or bottom quintile remained in the same quintile between GRS100 and GRS181, whereas 69% to 70% remained between GRS269 and GRS451 (Supplementary Figures 2a). Likewise, the percentage of cases has increased for each population within higher GRS categories (e.g., from 40.5% in the top quintile of GRS100 to 51.2% in GRS451) and decreased within lower GRS categories (e.g., from 7.5% in the bottom quintile of GRS100 to 4.4% in GRS451; Fig. 3, Supplementary Figures 2b). Risk classification with the GRS in addition to age was evaluated using the net reclassification index (NRI) 24 and showed substantial improvement from GRS100 (range across populations: 30.2% in African to 49.5% in European) to GRS451 (range across populations: 58.5% in African to 69.9% in European; Supplementary Table 3), compared to an age-only model. Compared to a model with age + GRS269, the population specific improvement for a model with age + GRS451 resulted in a NRI ranging from 3.3% in Asian ancestry to 21.7% in Hispanics. The improvement in risk prediction of GRS451 over previous GRS panels was shown in the GWAS discovery samples and confirmed in replication studies among men of European and African ancestry that were not included in the GWAS (Fig. 4a-b, 54 Supplementary Table 4). In a subset of samples included in the GWAS discovery samples, compared to men in the 40-60% GRS decile, the estimated OR for men in the 90-100% category of the GRS was 4.90 [95% CI: 4.74-5.05] for men of European ancestry, 3.59 [95% CI:3.37-3.82] for men of African ancestry, 4.40 [95% CI: 3.46-5.59] for men of Asian ancestry and 3.81 [95% CI: 3.30-4.41] for Hispanic men. The OR for the 99-100% GRS category ranged from 9.99 [95% CI: 9.41-10.61] in men of European ancestry to 6.39 [95% CI: 5.62-7.25] in men of African ancestry. Compared to the GRS based on the 269 previously known risk variants, the smallest increase was observed in Asian men (-2%), while the largest increase was in African men (22%) for the top 1% category. We replicated the GRS association for prostate cancer among studies with individuals that were not included in the risk variants discovery population. For men of European ancestry, the ORs for the top decile and for the top 1% were 4.33 [95%CI: 3.64-5.15] and 10.13 [95%CI: 7.32-14.03] in MGB, 3.95 [95%CI: 3.45-4.52] and 8.66 [95%CI: 6.69-11.22] in MGI, and 3.51 [95%CI: 3.06-4.03] and 8.13 [95%CI: 6.39-10.34] in EstBB, respectively. For men of African ancestry, the ORs for the top decile and for the top 1% were 3.54 [95%CI: 2.87-4.37] and 4.25 [95%CI: 2.57-7.01] in MADCaP, 4.93 [95%CI: 2.00-12.74] and 2.56 [95%CI: 0.1-28.48] in MGB, 3.54 [95%CI: 1.82-6.86] and 2.90 [95%CI: 0.67-12.66] in MGI, respectively. Based on the high degree of variation in the association of GRS451 with PCa risk across sub-studies in the discovery and replication phases (eFig. 2), a single summary OR per SD was estimated from the overall meta-analyzed sample: 2.32 [95%CI: 2.30-2.35], 2.04 [95%CI: 2.00-2.08], 2.15 [95%CI: 1.99-2.32] and 2.12 [95%CI: 2.03-2.23] for European, African, Asian and Hispanic men, respectively (Pheterogeneity by population: 4.51x10 -50 , 7.52x10 -4 , 0.29, and 0.31, respectively). The 55 ORs in the replication studies were 2.19 [95%CI: 2.12-2.25] in European and 1.79 [95%CI:1.69- 1.90] in African ancestry men (Fig. 4b). As observed for GRS269, age modifies the association of GRS451 and PCa risk (Fig. 4c, Supplementary Table 5, Methods) 61 . In men of European ancestry, GRS451 was associated with an OR per SD of 2.90 [95 % CI: 2.80-3.00] for men ≤ 55 and 2.30 [95% CI: 2.27-2.32] for men > 55 years (Pheterogeneity = 2.0x10 -37 ). Effect modification of GRS451 by age was similarly observed in men of African ancestry: OR per SD = 2.45 [95 % CI: 2.33-2.58] for men ≤ 55 years and 2.00 [95% CI: 1.95-2.05] for men > 55 years (Pheterogeneity = 3.3x10 -12 ). In the European replication studies, GRS451 was associated with an OR per SD of 2.69 [95 % CI: 2.32-3.13] for men ≤ 55 and 2.03 [95% CI: 1.93-2.12] for men > 55 years (Pheterogeneity = 2.1x10 -16 ) in MGI, and 2.63 [95 % CI: 2.19- 3.02] for men ≤ 55 and 2.24 [95% CI: 2.12-2.38] for men > 55 years (Pheterogeneity = 3.6x10 -2 ) in EstBB. Similarly, in the African replication study MADCaP, the OR per SD were 2.24 [95 % CI: 1.73-2.90] for men ≤ 55 and 1.76 [95% CI: 1.65-1.89] for men > 55 years (Pheterogeneity = 7.9x10 -2 ). In men of European and Asian ancestry and in Hispanic men, the GRS451 was equally associated with risk of aggressive PCa (stage T3/T4, regional lymph node involvement, metastatic disease, Gleason score ≥ 8, prostate-specific antigen (PSA) level ≥ 20 ng/mL or PCa as the underlying cause of death) and non-aggressive PCa (no aggressive features) (Fig. 4c, Supplementary Table 6, Methods). For men of African ancestry with prostate cancer, GRS451 was associated with a greater risk of aggressive versus non-aggressive disease (OR per SD = 1.08, 95% CI, 1.04-1.12, P=1.1x10 -4 ; Fig. 4d). A stronger association of GRS451 with aggressive disease in African ancestry 56 men was also observed in the African PCa MADCAP replication sample (OR per SD= 1.12, 95% CI, 1.01-1.23, P = 0.03). Fifty-one of the 451 PCa risk variants have been directly or indirectly (LD R 2 >0.8) associated in GWAS of PSA at P<5x10 -8 (Methods). To assess whether the PCa risk signals for PSA-associated variants reflect an increased likelihood of PCa detection due to screening, particularly for low- stage disease, we examined their aggregate association with disease aggressiveness (Supplementary Table 7). When removing the PCa-PSA variants from the GRS analysis we found the GRS (with 400 markers) to be more strongly associated with aggressive disease (versus GRS451) in European men (OR per SD = 1.04, 95% CI, 1.03-1.06, P =3.2x10 -8 ), African men (OR per SD = 1.10, 95% CI, 1.06-1.14, P =7.0x10 -7 ) and Hispanic men (OR per SD = 1.05, 95% CI, 0.94-1.14, P =0.21) which suggests that some PCa risk variants may be over-represented in men with less aggressive disease as the result of their association with PSA levels. To illustrate the genetic effect on the cumulative risk of prostate cancer, we calculated the absolute risk for men of European and African ancestry based on the effect sizes for standardized GRS in each population. For men of European ancestry, 20% of men have a 2-fold or greater risk compared to men at the 50% of GRS451, and these men achieve an absolute risk comparable to the median risk in the population 16 years earlier. Specifically, these men reach a level of absolute risk of at least 7.8% (the risk at age 85 for men with a 50% GRS451) by age 69 or earlier (Fig. 5). For African ancestry men, 16% of men achieve a 2-fold or greater risk by age 66, with an absolute risk comparable to the risk reached by the average man by age 85 (11.6%), a full 19 years earlier. 57 3.4 Discussion In this largest multi-ancestry meta-analysis for prostate cancer GWAS, we identified 187 novel risk variants for PCa, increasing the total number of risk variants from 287 to 451. The GRS with the current 451 risk variants showed an improved ability in differentiating individuals at risk for prostate cancer across populations. An externally replicated multi-ancestry GRS was associated with risk that ranged from 1.8 (per standard deviation) in men of African ancestry to 2.2 in European ancestry men. The GRS was associated with a greater risk of aggressive versus non- aggressive disease in men of African ancestry. Our study presents novel PCa susceptibility loci and a GRS with effective risk stratification across ancestry groups. The underlying rationale for conducting a cross-ancestry meta-analysis is based on the hypothesis that true causal variants are predominantly shared across populations. Among the 451 risk variants, only 23 variants showing statistically significant heterogeneity in ancestry-specific effect sizes, while the effect sizes for common variants exhibited strong correlations across populations, thus confirming our hypothesis. Our GRS showed a substantial improvement in differentiating prostate cancer case/control status and risk stratification compared to previous risk variant sets. However, we acknowledge potential over-fitting issues that may have arisen during GRS construction and validation. First, it is possible that the GRS weights for the 451 risk variants are inflated because they are effect estimates from the original GWAS that detected them, sometimes referred to as “winner’s curse”. To address this issue, we applied the bias-reduction procedure based on conditional likelihood to obtain the corrected effect estimates of those selected risk variants. After correction, the weights are very 58 consistent with the uncorrected weights (slope=1, R=0.99) and the GRS association results remain relatively unchanged, suggesting that the weights are not likely to be biased 25 . Note that as shown in Zhong and Prentice (2008) and others 25 , the corrected estimates do not vary notably from the original estimates due to the substantial sample size that we are using for discovery. Second, to highlight the variability in estimation and precision of the GRS effect on prostate cancer, we presented effects for each of the contributing studies in our GWAS discovery and replication phases. We observed that there is not a general trend in estimation by the contributing sample size nor the time in which the study contributed to the discovery of the various GRS sets of SNPs. Although effect sizes for replication studies were generally lower, they were not consistently among the lowest studies. Therefore, we reported the meta-analyzed effect sizes for all studies, including GWAS discovery and replication, as the main effect, as it has the largest sample size and is statistically more stable. The GRS was found to be more strongly associated with aggressive than non-aggressive prostate cancer only among men of African ancestry. This observation could be attributed to differences in access to care across populations, as delayed diagnosis in some studies may result in more aggressive sub-types of prostate cancer being included. Unfortunately, we don’t have information about health care systems or access to care for the various African ancestry studies. We attempted to evaluate this question using percentage of African ancestry in the sub-studies and continent as proxy measures of health care and conducted several sensitivity analyses within African ancestry subgroups. However, neither the percentage of African ancestry (in controls) or continent was significantly associated with the ORs of aggressive vs. non-aggressive disease, which do not support the hypothesis that health care systems or access to care contribute to the GRS association 59 with disease aggressiveness in men of African ancestry. Another possible explanation could be the effect of PSA screening. PSA screening has been widely adopted in Western countries over the past few decades, leading to the identification of men with occult prostate cancer due to their elevated PSA levels in our GWAS discovery sub-studies. Consequently, this may have resulted in the identification of risk variants for PSA levels instead of prostate cancer. After removing the PSA variants, the GRS was found to be associated with aggressive prostate cancer in men of European, Asian, and Hispanic ancestry, and the association was even stronger in men of African ancestry. These findings suggest that the current risk variant list may include some prostate cancer risk variants that are overrepresented in men with less aggressive disease due to their association with PSA levels. However, it is difficult to distinguish between PSA and prostate cancer risk variants. Therefore, further studies regarding this issue are needed. A man’s cumulative risk of developing PCa, including aggressive disease, is profoundly influenced by the GRS. As shown in the absolute risk calculation, men with a more than 2-fold higher genetic risk of prostate cancer compared to the median risk men have a diagnosis of prostate cancer more than 16 years earlier in average, highlighting the importance of GRS in identifying men who may benefit from early screening and preventive measures. Moreover, a GRS-informed approach to screening may improve early detection, as over 50% of cases, including those with aggressive and lethal disease, develop among men in the top GRS quintile, while fewer than 5% of cases develop among men in the bottom 20% (Fig. 3). 60 In conclusion, in this large multi-ancestry meta-analysis of prostate cancer GWAS, we identified 451 risk variants of which 187 were novel. The GRS based on the 451 variants showed significant improvement in differentiating prostate cancer over previous GRS panels. Evidence that GRS can differentiate risk of aggressive versus non-aggressive disease, albeit modestly, for men of African ancestry suggests potential clinical utility in this high-risk population. Furthermore, a better understanding of the relationship between germline variants that influence both PSA levels and PCa risk variants is needed to accurately estimate the GRS association with PCa aggressiveness and PCa outcomes. Our findings support previous work demonstrating the ability of multi-ancestry studies to identify PCa risk variants that improve risk prediction across populations. As an effective tool for personalized risk assessment, the GRS may provide men with information about their risk of prostate cancer that could help in their decision-making process regarding when and how often to screen for early PCa detection. 61 3.5 References 1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018;68(6):394-424. 2. Conti DV, Darst BF, Moss LC, et al. Trans-ancestry genome-wide association meta- analysis of prostate cancer identifies new susceptibility loci and informs genetic risk prediction. Nature Genetics. 2021;53(1):65-75. 3. Schumacher FR, Al Olama AA, Berndt SI, et al. Association analyses of more than 140,000 men identify 63 new prostate cancer susceptibility loci. Nat Genet. 2018;50(7):928-936. 4. Dadaev T, Saunders EJ, Newcombe PJ, et al. Fine-mapping of prostate cancer susceptibility loci in a large meta-analysis identifies candidate causal variants. Nat Commun. 2018;9(1):2256. 5. Conti DV, Wang K, Sheng X, et al. Two Novel Susceptibility Loci for Prostate Cancer in Men of African Ancestry. J Natl Cancer Inst. 2017;109(8). 6. Wang M, Takahashi A, Liu F, et al. Large-scale association analysis in Asians identifies new susceptibility loci for prostate cancer. Nat Commun. 2015;6:8469. 7. Hoffmann TJ, Van Den Eeden SK, Sakoda LC, et al. A large multiethnic genome-wide association study of prostate cancer identifies novel risk variants and substantial ethnic differences. Cancer Discov. 2015;5(8):878-891. 8. Al Olama AA, Kote-Jarai Z, Berndt SI, et al. A meta-analysis of 87,040 individuals identifies 23 new susceptibility loci for prostate cancer. Nat Genet. 2014;46(10):1103- 1109. 9. Eeles RA, Olama AA, Benlloch S, et al. Identification of 23 new prostate cancer susceptibility loci using the iCOGS custom genotyping array. Nat Genet. 2013;45(4):385- 391, 391e381-382. 10. Gudmundsson J, Sulem P, Gudbjartsson DF, et al. A study based on whole-genome sequencing yields a rare variant at 8q24 associated with prostate cancer. Nat Genet. 2012;44(12):1326-1329. 11. Takata R, Akamatsu S, Kubo M, et al. Genome-wide association study identifies five new susceptibility loci for prostate cancer in the Japanese population. Nat Genet. 2010;42(9):751-754. 12. Gudmundsson J, Sulem P, Rafnar T, et al. Common sequence variants on 2p15 and Xp11.22 confer susceptibility to prostate cancer. Nat Genet. 2008;40(3):281-283. 62 13. Amundadottir LT, Sulem P, Gudmundsson J, et al. A common variant associated with prostate cancer in European and African populations. Nat Genet. 2006;38(6):652-658. 14. Wojcik GL, Graff M, Nishimura KK, et al. Genetic analyses of diverse populations improves discovery for complex traits. Nature. 2019;570(7762):514-518. 15. Das S, Forer L, Schönherr S, et al. Next-generation genotype imputation service and methods. Nat Genet. 2016;48(10):1284-1287. 16. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 2009;5(6):e1000529. 17. Loh P-R, Danecek P, Palamara PF, et al. Reference-based phasing using the Haplotype Reference Consortium panel. Nature Genetics. 2016;48(11):1443-1448. 18. Kurki MI, Karjalainen J, Palta P, et al. FinnGen: Unique genetic insights from combining isolated population and national health register data. medRxiv. 2022:2022.2003.2003.22271360. 19. Auton A, Brooks LD, Durbin RM, et al. A global reference for human genetic variation. Nature. 2015;526(7571):68-74. 20. Taliun D, Harris DN, Kessler MD, et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed Program. Nature. 2021;590(7845):290-299. 21. Bycroft C, Freeman C, Petkova D, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203-209. 22. Amin Al Olama A, Dadaev T, Hazelett DJ, et al. Multiple novel prostate cancer susceptibility signals identified by fine-mapping of known risk loci among Europeans. Hum Mol Genet. 2015;24(19):5589-5602. 23. Conti DV, Wang K, Sheng X, et al. Two Novel Susceptibility Loci for Prostate Cancer in Men of African Ancestry. J Natl Cancer Inst. 2017;109(8). 24. Pencina MJ, D'Agostino RB, Sr., Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30(1):11-21. 25. Zhong H, Prentice RL. Bias-reduced estimators and confidence intervals for odds ratios in genome-wide association studies. Biostatistics. 2008;9(4):621-634. 26. Kachuri L, Hoffmann TJ, Jiang Y, et al. Leveraging Genetic Determinants of Prostate- Specific Antigen Levels Towards Improving Prostate Cancer Screening. Nature Medicine. in press. 63 27. Amin Al Olama A, Benlloch S, Antoniou AC, et al. Risk Analysis of Prostate Cancer in PRACTICAL, a Multinational Consortium, Using 25 Known Prostate Cancer Susceptibility Loci. Cancer Epidemiol Biomarkers Prev. 2015;24(7):1121-1129. 28. Antoniou AC, Beesley J, McGuffog L, et al. Common Breast Cancer Susceptibility Alleles and the Risk of Breast Cancer for BRCA1 and BRCA2 Mutation Carriers: Implications for Risk Prediction. Cancer Res. 2010;70(23):9742-9754. 29. Antoniou AC, Pharoah PDP, McMullan G, Day NE, Ponder BAJ, Easton DE. Evidence for further breast cancer susceptibility genes in addition to BRCA1 and BRCA2 in a population-based study. Genet Epidemiol. 2001;21:1-18. 30. Kuchenbaecker KB, McGuffog L, Barrowdale D, et al. Evaluation of Polygenic Risk Scores for Breast and Ovarian Cancer Risk Prediction in BRCA1 and BRCA2 Mutation Carriers. J Natl Cancer Inst. 2017;109(7). 31. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence - SEER Research Plus Data, 18 Registries, Nov 2020 Sub (2000-2018) - Linked To County Attributes - Total U.S., 1969-2019 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, released April 2021, based on the November 2020 submission. 32. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Mortality - All COD, Aggregated With State, Total U.S. (1990- 2019) <Katrina/Rita Population Adjustment>, National Cancer Institute, DCCPS, Surveillance Research Program, released April 2021. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). 33. Chang X, Wang K. wANNOVAR: annotating genetic variants for personal genomes via the web. J Med Genet. 2012;49(7):433-436. 34. Mei S, Qin Q, Wu Q, et al. Cistrome Data Browser: a data portal for ChIP-Seq and chromatin accessibility data in human and mouse. Nucleic Acids Res. 2017;45(D1):D658- d662. 35. Lawrence M, Gentleman R, Carey V. rtracklayer: an R package for interfacing with genome browsers. Bioinformatics. 2009;25(14):1841-1842. 36. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 2013;45(6):580-585. 37. Thibodeau SN, French AJ, McDonnell SK, et al. Identification of candidate genes for prostate cancer-risk SNPs utilizing a normal prostate tissue eQTL data set. Nat Commun. 2015;6:8653. 38. Tian J, Wang Z, Mei S, et al. CancerSplicingQTL: a database for genome-wide identification of splicing QTLs in human cancer. Nucleic Acids Research. 2018;47(D1):D909-D916. 64 39. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020;369(6509):1318-1330. 40. Gusev A, Ko A, Shi H, et al. Integrative approaches for large-scale transcriptome-wide association studies. Nat Genet. 2016;48(3):245-252. 41. Stegle O, Parts L, Durbin R, Winn J. A Bayesian framework to account for complex non- genetic factors in gene expression levels greatly increases power in eQTL studies. PLoS Comput Biol. 2010;6(5):e1000770. 42. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88(1):76-82. 43. Wang G, Sarkar A, Carbonetto P, Stephens M. A simple new approach to variable selection in regression, with application to genetic fine mapping. Journal of the Royal Statistical Society: Series B (Statistical Methodology). 2020;82(5):1273-1300. 44. The Molecular Taxonomy of Primary Prostate Cancer. Cell. 2015;163(4):1011-1025. 45. Sun BB, Maranville JC, Peters JE, et al. Genomic atlas of the human plasma proteome. Nature. 2018;558(7708):73-79. 46. Newcombe PJ, Conti DV, Richardson S. JAM: A Scalable Bayesian Framework for Joint Analysis of Marginal SNP Effects. Genet Epidemiol. 2016;40(3):188-201. 47. Chen F, Madduri RK, Rodriguez AA, et al. Evidence of Novel Susceptibility Variants for Prostate Cancer and a Polygenic Risk Score that Improves Prediction of Aggressive Disease for Men of African Ancestry submitted. 48. Barfeld SJ, East P, Zuber V, Mills IG. Meta-analysis of prostate cancer gene expression data identifies a novel discriminatory signature enriched for glycosylating enzymes. BMC Med Genomics. 2014;7:513. 49. Halvorsen OJ, Rostad K, Øyan AM, et al. Increased Expression of SIM2-s Protein Is a Novel Marker of Aggressive Prostate Cancer. Clinical Cancer Research. 2007;13(3):892- 897. 50. Dhanasekaran SM, Barrette TR, Ghosh D, et al. Delineation of prognostic biomarkers in prostate cancer. Nature. 2001;412(6849):822-826. 51. Linn DE, Yang X, Xie Y, et al. Differential regulation of androgen receptor by PIM-1 kinases via phosphorylation-dependent recruitment of distinct ubiquitin E3 ligases. J Biol Chem. 2012;287(27):22959-22968. 52. Wang J, Kim J, Roh M, et al. Pim1 kinase synergizes with c-MYC to induce advanced prostate carcinoma. Oncogene. 2010;29(17):2477-2487. 65 53. Sun X, Frierson HF, Chen C, et al. Frequent somatic mutations of the transcription factor ATBF1 in human prostate cancer. Nat Genet. 2005;37(4):407-412. 54. Chandler JD, Williams ED, Slavin JL, Best JD, Rogers S. Expression and localization of GLUT1 and GLUT12 in prostate carcinoma. Cancer. 2003;97(8):2035-2042. 55. White MA, Tsouko E, Lin C, et al. GLUT12 promotes prostate cancer cell growth and is regulated by androgens and CaMKK2 signaling. Endocr Relat Cancer. 2018;25(4):453- 469. 56. Mi Y, Zhao S, Zhang W, et al. Down-regulation of Barx2 predicts poor survival in colorectal cancer. Biochemical and Biophysical Research Communications. 2016;478(1):67-73. 57. Zhang J, Dutta D, Köttgen A, et al. Plasma proteome analyses in individuals of European and African ancestry identify cis-pQTLs and models for proteome-wide association studies. Nat Genet. 2022;54(5):593-602. 58. Gamazon ER, Wheeler HE, Shah KP, et al. A gene-based association method for mapping traits using reference transcriptome data. Nat Genet. 2015;47(9):1091-1098. 59. Mancuso N, Gayther S, Gusev A, et al. Large-scale transcriptome-wide association study identifies new prostate cancer risk regions. Nat Commun. 2018;9(1):4079. 60. Liu D, Zhu J, Zhou D, et al. A transcriptome-wide association study identifies novel candidate susceptibility genes for prostate cancer risk. Int J Cancer. 2022;150(1):80-90. 61. Chen F, Darst BF, Madduri RK, et al. Validation of a multi-ancestry polygenic risk score and age-specific risks of prostate cancer: A meta-analysis within diverse populations. eLife. 2022;11:e78304. 66 Chapter Four: Clonal hematopoiesis and risk of prostate cancer in large samples of European ancestry men 4.1 Introduction Age-related clonal hematopoiesis (CH), also referred to as clonal hematopoiesis of indeterminate potential (CHIP) in the absence of a hematologic malignancy, is the expansion of hematopoietic stem cells with somatic mutations and is increasingly common with older age. In addition to CH being a risk factor for myeloid malignancy development, it has also been associated with increased risk of all-cause mortality and cardiovascular disease 1-3 . Individuals with solid tumors have been reported to be more likely to have clonal mosaicism than cancer-free participants 4 . Age-related loss of chromosome Y (LOY) in circulating leukocytes has also been associated with increased risk of non-hematological cancer mortality 5 . Further, in a two-sample Mendelian randomization analysis, genetically predicted LOY was reported to be associated with increased genetic risk of prostate cancer and other solid tumors 6 . However, little is known regarding the potential impact of age-related CH on risk of prostate cancer, the fifth leading cause of cancer death of men worldwide 7 . Pathogenic germline variants in many genes that are associated with CH, particularly DNA repair genes ATM, CHEK2, and NBN, are also associated with prostate and other non- hematologic cancers 8-11 , suggesting a potential mechanistic link between these two conditions. In this investigation, we evaluated the association of age-related CHIP with overall and aggressive prostate cancer risk in two large whole-exome sequencing studies of European ancestry men. 4.2 Methods Participants and Genetic Sequencing 67 To investigate the association between CHIP variants and risk of overall prostate cancer, we analyzed 69,502 European ancestry men from the UK Biobank with whole exome-sequencing (WES) data, which was generated at the Regeneron Genetics Center 12,13 . On average, 95.6% of the targeted regions were sequenced with at least 20x coverage 12 . Prostate cancer cases were identified through linkage to the NHS Central Register for the first diagnosis of prostate cancer. Quality control criteria applied to this sample of men from the UK Biobank have been previously described 14 . To investigate the association between CHIP variants and risk of aggressive prostate cancer, 5,545 European ancestry men were included from WESP, consisting of 12 large European and US studies 9 . WES was performed at the Center for Inherited Diseases Research with average targeted exon coverage of 56x and 95.7% of targeted regions sequenced with at least 10x coverage. Aggressive prostate cancer was defined as men who either died due to prostate cancer, had metastatic disease, had stage T4 disease, or had stage T3 disease with a Gleason score ≥8 tumor. Non-aggressive prostate cancer was defined as men who had stage T1/T2 disease and a Gleason score ≤6 tumor, with 71.3% also having ≥10 years of follow-up to indicate that they were alive and without recurrence. Details of the study design, sequencing procedures, and quality control were previously described 9 . Informed consents were obtained from all participants, and study protocols were approved by respective institutional review boards. Identification of Clonal Hematopoiesis Variants Somatic CHIP variants were identified based on a list of 74 genes with previously reported mutations in human hematologic cancers 1,15,16 . Somatic short variants (SNVs and Indels) were 68 identified using the GATK toolkit following the GATK's best practices workflows. In short, somatic variant calling was carried out using the GATK Mutect2 17 in tumor-only mode. A panel of normals (PoN) was incorporated to filter out commonly seen sequencing artefacts. A subset of 100 randomly selected UK Biobank individuals under 40 years of age were used as the PoN for UK Biobank, while WESP used the PoN provided in the GATK resource bundle consisting of several hundred normals. Population allele frequencies of common and rare variants from gnomAD were provided in the GATK resource bundle as an external reference of germline variants and were utilized to filter out possible germline variants. Somatic short variants were further filtered through the GATK FilterMutectCalls. Variants included among these genes were those with deleterious (protein truncating or splice altering) functional consequences 18 or those specifically reported in Jaiswal et al. 1 with minor allele frequencies (MAF) <0.1% (in the respective UK Biobank or WESP data) and a variant allelic fraction (VAF) >5% in the UK Biobank and >10% in WESP given differences in sequence coverage, as previously suggested to increase the likelihood of a mutation being somatic 15,19,20 . This lower bound VAF in WESP was selected given the coverage of our exome sequencing, as 30-50x coverage has been reported to be able to robustly call CHIP variants with VAF >10% 15 . Based on previous literature 20,21 , we also conducted sensitivity analyses using VAFs between 10-40%, 10-60%, and 10-40% or 60-90%. These sensitivity analyses all led to similar results and the same conclusions as our primary analysis. VAFs were calculated with bcftools (fill-tags FORMAT/VAF) as the fraction of reads with the alternate allele 22 . We excluded variants with MAF >0.1% in the Genome Aggregation Database (gnomAD) 23 and those in simple tandem repeat regions 24,25 . This led to a total of 1,778 variants in 55 CHIP genes identified in the UK Biobank and 360 variants in 52 CHIP genes identified in WESP (Supplemental Tables 1-2). 69 Identification of DNA Repair Gene Variants A total of 24 previously curated prostate cancer candidate DNA repair genes 9 were identified, as DNA repair genes have been shown to predispose to prostate cancer 9,10,26-28 and clonal hematopoiesis 8 . Within these genes, pathogenic/likely pathogenic/deleterious (P/LP/D) variants were considered and defined as rare variants (MAF<0.01) in the study population that had either a Variant Effect Predictor (VEP) Impact score of “high” 18 and/or a ClinVar classification of Pathogenic or Likely Pathogenic 29 . We excluded the known low/moderate prostate cancer risk variant c.9976A>T (rs11571833) in BRCA2 30 . Polygenic Risk Score Construction In the UK Biobank, where GWAS data was available, a PRS was constructed based on a multi- ancestry prostate cancer genome-wide association study (GWAS) meta-analysis of >230,000 men, where we developed a PRS using 269 variants and corresponding multi-ancestry weights and found that the PRS was highly predictive of prostate cancer risk across populations 11 . Of these 269 variants, 267 were present in the UK Biobank data and had an imputation info score >0.50 (median info score=0.99). The PRS was calculated as a weighted sum of the number of risk alleles among the 267 variants, using the variant-specific multi-ancestry weights we previously reported. Association Testing We evaluated the association between CHIP variants and prostate cancer risk using gene-based analyses, considering carrier status for qualifying CHIP variants within each gene individually. We also evaluated the association between CHIP variants and prostate cancer risk by aggregating 70 across all genes, considering carrier status for any qualifying variants. Men carrying ≥1 allele among the identified CHIP genes (individually or in aggregate, depending on the assessment) were considered carriers. In the UK Biobank, we examined associations between carrier status and overall incident prostate cancer, age at enrollment, and age at diagnosis (for cases only), adjusting for age at enrollment (for overall prostate cancer and age at diagnosis) and the first 10 genetic principal components of ancestry to account for potential population stratification. In WESP, we examined associations between carrier status and aggressive versus non-aggressive prostate cancer, death due to prostate cancer versus non-aggressive prostate cancer, and metastatic versus non- aggressive prostate cancer. Age at diagnosis, study, country, and the first three principal components of ancestry were adjusted for as covariates. Logistic regression models were used for binary prostate cancer outcomes, while linear regression models were used for continuous age outcomes. Cox proportional hazards models were used to evaluate incident prostate cancer status in the UK Biobank with age in years as the time metric, using age at blood draw as the entry time and age at prostate cancer diagnosis as the exit time. We evaluated the association between the prostate cancer PRS and age-related CHIP carrier status across all CHIP genes and for individual CHIP genes in prostate cancer controls from the UK Biobank. This analysis was performed using logistic regression models with CHIP status as the outcome and the continuous PRS as the predictor, adjusting for age at blood draw and the first 10 principal components of ancestry. We also evaluated the association between carrier status for P/LP/D variants in DNA repair genes and age-related CHIP carrier status using logistic regression models with CHIP status as the outcome and DNA repair gene carrier status as the predictor. In the UK Biobank, analyses were adjusted for age at blood draw and the first 10 principal 71 components of ancestry, and in WESP, analyses were adjusted for age at prostate cancer diagnosis, study, country, and the first three principal components of ancestry. Analyses were performed aggregating across all 24 DNA repair genes and separately aggregating across three DNA repair genes: BRCA2, ATM, and PALB2, which we previously reported to be associated with aggressive prostate cancer 9 . A Bonferroni-corrected P-value <0.05 was considered statistically significant (P-values presented in the “Results” section are unadjusted). R 3.6.0 was used for all analyses. 4.3 Results Whole-exome sequence data was analyzed for 2,118 incident prostate cancer cases and 67,384 controls from the UK Biobank 14 and 2,770 aggressive and 2,775 non-aggressive prostate cancer cases from a cross-sectional case-only study from 12 international study sites, referred to here as the Whole-Exome Sequencing Study in Prostate Cancer (WESP) 9 . All men were of European ancestry. Potential CHIP variants were called with Mutect2 17 and defined based on previous curations of variants within 74 established hematologic cancer genes 1,15,16 . Specifically, CHIP variants in these genes were rare with minor allele frequencies (MAF) <0.1% in the study population, excluded variants with MAF >0.1% in the Genome Aggregation Database (gnomAD) 23 to reduce the potential of capturing germline variants, and were either deleterious (protein truncating or splice altering) 18 or specifically reported in Jaiswal et al. 1 . Variant allelic fractions (VAFs) were >5% in the UK Biobank and >10% in WESP due to differences in exome sequencing coverage, thresholds 72 previously suggested to increase the likelihood of a mutation being somatic 15,19-21 . VAFs were calculated with bcftools (fill-tags FORMAT/VAF) as the fraction of reads with the alternate allele 22 . Several VAF thresholds were tested in sensitivity analyses and led to similar results and the same conclusions (see Materials and Methods for details). A total of 1,778 qualifying CHIP variants in 55 genes were identified in the UK Biobank, while 360 qualifying CHIP variants in 52 genes were identified in WESP (VAFs are described in Table 1 and Supplemental Tables 1-2; Materials and Methods). Overall, 2,874 (4.1%) men in the UK Biobank and 392 (7.1%) men in WESP were found to carry a CHIP variant. The most commonly carried CHIP variants were in DNMT3A (33.6% of the 2,874 CHIP carriers in UK Biobank and 26.0% of the 392 CHIP carriers in WESP), TET2 (25.2% of UK Biobank carriers and 22.7% of WESP carriers), and ASXL1 (9.1% of UK Biobank carriers and 8.9% of WESP carriers; Supplemental Figure 1), consistent with previous studies 1,31,32 . Given the overall CHIP carrier frequencies and study sample sizes, the UK Biobank and WESP both had 80% power to detect an OR of 1.21 for the association of CHIP with overall and aggressive prostate cancer risk. In the UK Biobank, the median age at blood draw was 57 years (interquartile range [IQR]=13) and the median time between blood draw and cancer diagnosis among cases was 4.1 years (IQR=3.6; Supplemental Figure 2). As expected, in the UK Biobank, CHIP status was significantly associated with age at blood draw, with the strongest associations observed with DNMT3A (+4.1 years, 95% CI=3.6-4.6, P=4.4x10 -55 ), ASXL1 (+6.2 years, 95% CI=5.2-7.2, P=1.3x10 -35 ), TET2 (+3.7 years, 95% CI=3.2-4.3, P=2.6x10 -35 ), PPM1D (+5.8 years, 95% CI=4.1-7.5, P=3.8x10 -11 ), and SF3B1 (+8.8 years, 95% CI=5.5-12.0, P=1.1x10 -07 ) and across all 55 genes in aggregate (+3.2 years, 95% CI=2.9-3.5, P=2.4x10 -99 ; Table 2, Figure 1, and Supplemental Figure 2-3). However, 73 no significant associations were observed between CHIP carrier status and age at prostate cancer diagnosis, with or without adjustment for age at blood draw, in the UK Biobank (Table 2 and Supplemental Figure 3). In the UK Biobank, we did not observe a significant difference in CHIP carrier frequencies between prostate cancer cases and controls (4.86% and 4.11%, respectively; HR=0.93, 95% CI=0.76-1.13, P=0.46; Table 1; see Materials and Methods for details). In WESP, CHIP carrier frequencies did not significantly differ when comparing cases with aggressive prostate cancer (7.15%; OR=1.14, 95% CI=0.92-1.41, P=0.22), prostate cancer death (6.38%; OR=1.02, 95% CI=0.81-1.30, P=0.84), or metastatic prostate cancer (7.07%; OR=1.22, 95% CI=0.81-1.83, P=0.34) to non-aggressive prostate cancer cases (6.99%; Table 1). Similarly, in gene-based tests, no significant associations were observed between CHIP carrier status and overall prostate cancer risk in the UK Biobank or with disease aggressiveness in WESP (Table 1 and Supplemental Figure 4). Carrier frequencies for the aggregate of CHIP genes PTEN, TP53, HLA-A, and MAP2K1, which have prior evidence of association with prostate cancer risk 11,33-35 , were not significantly associated with overall prostate cancer risk in the UK Biobank (HR=0.93, 95% CI=0.30-2.89, P=0.90) or with disease aggressiveness in WESP (OR=1.50, 95% CI=0.39-5.82, P=0.56). We found weak evidence of association between genetic susceptibility to prostate cancer, measured by a polygenic risk score (PRS) constructed based on a previous publication 11 , and CHIP carrier status across all CHIP genes in prostate cancer controls from the UK Biobank (OR=1.05 for each additional risk allele, 95% CI=1.01-1.10, P=0.01; Materials and Methods). CHIP carrier 74 status for DNMT3A had the strongest association with the PRS in UK Biobank controls (OR=1.11, 95% CI=1.03-1.19, P=5.6x10 -3 ); however, this association was not significant after adjusting for multiple testing of all individual CHIP genes. We also observed a null association between carrier status for pathogenic/likely pathogenic/deleterious variants across 24 previously curated prostate cancer candidate DNA repair genes 9 and age-related CHIP carrier status across all CHIP genes in WESP (OR=0.99, 95% CI=0.72-1.36, P=0.96) and in the UK Biobank (OR=1.00, 95% CI=0.84- 1.20, P=0.96; see Materials and Methods for details). DNA repair gene carrier status was also not significantly associated with age-related CHIP carrier status for individual CHIP genes in WESP or the UK Biobank, with the exception of KDM6A in the UK Biobank (OR=3.33, 95% CI=1.70-6.52, P=4.4x10 -4 ; all other P-values≥0.002). Likewise, we did not observe a significant association between the aggregate of DNA repair genes BRCA2, ATM, and PALB2, which we previously reported to be associated with aggressive prostate cancer 9 , and age-related CHIP carrier status across all CHIP genes in WESP (OR=0.74, 95% CI=0.30-1.85, P=0.53) or the UK Biobank (OR=1.14, 95% CI=0.76-1.73, P=0.52). 4.4 Discussion In two large European ancestry datasets, we found minimal evidence of an association between age-related CHIP and risk of overall or aggressive prostate cancer. These findings are supported by a previous investigation reporting that clonal mosaicism was associated with increased risk of non-hematological cancers, but not with prostate cancer 4 . While this study was based on two large datasets, if CHIP has a weak association with prostate cancer risk, a larger number of CHIP carrier prostate cancer cases would be needed to detect such 75 an association. Further, our investigation was limited to men of European ancestry, and it is possible that the role of CHIP in prostate cancer risk could vary in non-European ancestry populations. For example, a germline sequencing study suggested that rare deleterious variants in TET2 were in aggregate associated with prostate cancer risk in men of African ancestry 36 . It is also possible that our approach to identifying CHIP variants may have introduced some non-differential exposure misclassification, although sensitivity analyses testing various VAF thresholds suggest that our findings are robust. In particular, based on previous estimates, the depth of our sequencing coverage in WESP provided ~50% sensitivity to detect CHIP variants in the VAF range of 5-10% and would need to be ~100x to capture variants in this range with closer to 100% sensitivity 15 . Observed differences in CHIP carrier frequencies between the UK Biobank and WESP could be in part due to differences in age distributions and the PoN panels used to call CHIP variants. Although our findings do not support an association between age-related CHIP and prostate cancer, a previous study found that therapy-related CH was associated with decreased survival in non- hematologic solid tumor cancer patients 31 . As such, it may be relevant to investigate the impact of CHIP on survival in post-therapy prostate cancer patients. Future studies of other types of CH may also provide important insights into prostate cancer risk, such as LOY, which is present in over 40% of men at age 70, is highly heritable, and has been previously associated with increased genetic risk of prostate cancer 6 . 76 4.5 References 1. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371(26):2488-2498. 2. Terao C, Suzuki A, Momozawa Y, et al. Chromosomal alterations among age-related haematopoietic clones in Japan. Nature. 2020;584(7819):130-135. 3. Loh PR, Genovese G, McCarroll SA. Monogenic and polygenic inheritance become instruments for clonal selection. Nature. 2020;584(7819):136-141. 4. Jacobs KB, Yeager M, Zhou W, et al. Detectable clonal mosaicism and its relationship to aging and cancer. Nat Genet. 2012;44(6):651-658. 5. Forsberg LA, Rasi C, Malmqvist N, et al. Mosaic loss of chromosome Y in peripheral blood is associated with shorter survival and higher risk of cancer. Nat Genet. 2014;46(6):624-628. 6. Thompson DJ, Genovese G, Halvardson J, et al. Genetic predisposition to mosaic Y chromosome loss in blood. Nature. 2019;575(7784):652-657. 7. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209-249. 8. Silver AJ, Bick AG, Savona MR. Germline risk of clonal haematopoiesis. Nat Rev Genet. 2021. 9. Darst BF, Dadaev T, Saunders E, et al. Germline Sequencing DNA Repair Genes in 5545 Men With Aggressive and Nonaggressive Prostate Cancer. J Natl Cancer Inst. 2021;113(5):616-625. 10. Matejcic M, Patel Y, Lilyquist J, et al. Pathogenic Variants in Cancer Predisposition Genes and Prostate Cancer Risk in Men of African Ancestry. JCO Precis Oncol. 2020;4:32-43. 11. Conti DV, Darst BF, Moss LC, et al. Trans-ancestry genome-wide association meta- analysis of prostate cancer identifies new susceptibility loci and informs genetic risk prediction. Nat Genet. 2021;53(1):65-75. 12. Szustakowski JD, Balasubramanian S, Kvikstad E, et al. Advancing human genetics research and drug discovery through exome sequencing of the UK Biobank. Nat Genet. 2021;53(7):942-948. 13. Van Hout CV, Tachmazidou I, Backman JD, et al. Exome sequencing and characterization of 49,960 individuals in the UK Biobank. Nature. 2020;586(7831):749- 756. 77 14. Darst BF, Sheng X, Eeles RA, Kote-Jarai Z, Conti DV, Haiman CA. Combined Effect of a Polygenic Risk Score and Rare Genetic Variants on Prostate Cancer Risk. Eur Urol. 2021. 15. Bick AG, Weinstock JS, Nandakumar SK, et al. Inherited causes of clonal haematopoiesis in 97,691 whole genomes. Nature. 2020;586(7831):763-768. 16. Jaiswal S, Natarajan P, Silver AJ, et al. Clonal Hematopoiesis and Risk of Atherosclerotic Cardiovascular Disease. N Engl J Med. 2017;377(2):111-121. 17. Cibulskis K, Lawrence MS, Carter SL, et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnol. 2013;31(3):213- 219. 18. McLaren W, Gil L, Hunt SE, et al. The Ensembl Variant Effect Predictor. Genome Biol. 2016;17(1):122. 19. Zhu B, Xiao Y, Yeager M, et al. Mutations in the HPV16 genome induced by APOBEC3 are associated with viral clearance. Nat Commun. 2020;11(1):886. 20. Shin HT, Choi YL, Yun JW, et al. Prevalence and detection of low-allele-fraction variants in clinical cancer samples. Nat Commun. 2017;8(1):1377. 21. Desai P, Mencia-Trinchant N, Savenkov O, et al. Somatic mutations precede acute myeloid leukemia years before diagnosis. Nat Med. 2018;24(7):1015-1023. 22. Li H, Handsaker B, Wysoker A, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25(16):2078-2079. 23. Karczewski KJ, Francioli LC, Tiao G, et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2020;581(7809):434-443. 24. Kent WJ, Sugnet CW, Furey TS, et al. The human genome browser at UCSC. Genome Res. 2002;12(6):996-1006. 25. Karolchik D, Hinrichs AS, Furey TS, et al. The UCSC Table Browser data retrieval tool. Nucleic Acids Res. 2004;32(Database issue):D493-496. 26. Pritchard CC, Mateo J, Walsh MF, et al. Inherited DNA-Repair Gene Mutations in Men with Metastatic Prostate Cancer. N Engl J Med. 2016;375(5):443-453. 27. Wu Y, Yu H, Li S, et al. Rare Germline Pathogenic Mutations of DNA Repair Genes Are Most Strongly Associated with Grade Group 5 Prostate Cancer. Eur Urol Oncol. 2020;3(2):224-230. 28. Leongamornlert DA, Saunders EJ, Wakerell S, et al. Germline DNA Repair Gene Mutations in Young-onset Prostate Cancer Cases in the UK: Evidence for a More Extensive Genetic Panel. Eur Urol. 2019;76(3):329-337. 78 29. Landrum MJ, Lee JM, Riley GR, et al. ClinVar: public archive of relationships among sequence variation and human phenotype. Nucleic Acids Res. 2014;42(Database issue):D980-985. 30. Meeks HD, Song H, Michailidou K, et al. BRCA2 Polymorphic Stop Codon K3326X and the Risk of Breast, Prostate, and Ovarian Cancers. J Natl Cancer Inst. 2016;108(2). 31. Coombs CC, Zehir A, Devlin SM, et al. Therapy-Related Clonal Hematopoiesis in Patients with Non-hematologic Cancers Is Common and Associated with Adverse Clinical Outcomes. Cell Stem Cell. 2017;21(3):374-382 e374. 32. Genovese G, Kahler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-2487. 33. Cancer Genome Atlas Research N. The Molecular Taxonomy of Primary Prostate Cancer. Cell. 2015;163(4):1011-1025. 34. Wedge DC, Gundem G, Mitchell T, et al. Sequencing of prostate cancers identifies new cancer genes, routes of progression and drug targets. Nat Genet. 2018;50(5):682-692. 35. Mancuso N, Gayther S, Gusev A, et al. Large-scale transcriptome-wide association study identifies new prostate cancer risk regions. Nat Commun. 2018;9(1):4079. 36. Koboldt DC, Kanchi KL, Gui B, et al. Rare Variation in TET2 Is Associated with Clinically Relevant Prostate Carcinoma in African Americans. Cancer Epidemiol Biomarkers Prev. 2016;25(11):1456-1463. 79 Chapter Five: Atopic allergic conditions and prostate cancer risk and survival in the Multiethnic Cohort Study 5.1 Background Prostate cancer is the most common non-skin malignancy among men in the US and worldwide 1 . It is also one of the leading causes of cancer mortality, contributing to over 359,000 deaths per year globally 1 . Established risk factors for prostate cancer include increasing age, family history of prostate cancer, genetic predisposition, and anthropometric factors (weight and height), whereas evidence for other modifiable factors, such as environmental and lifestyle factors, is less well understood 2,3 . Previous research suggests that the persistence of an inflammatory environment may be associated with prostate cancer development and progression 4,5 . Adiposity, a risk factor for aggressive prostate cancer, is involved in pro-inflammatory signaling and promoting oxidative stress 6-9 . Atopic allergic conditions (AACs), such as asthma, hay fever and food allergies, are caused by an enhanced state of immune response, thereby, potentially posing an elevated risk of prostate cancer. A hyperactive immune status, alternatively, may enhance immune surveillance and have antitumor effects. Current epidemiological evidence on the effect of AACs on prostate cancer has been mixed. The Health Professionals Follow-up Study found that men with self-reported hay fever had an increased risk of total prostate cancer, while men with asthma had a reduced risk of lethal and fatal prostate cancer 10 . By contrast, three large population-based cohort studies, the Busselton Health Survey 11 , Melbourne Collaborative Cohort Study 12 and Taiwan National Health Insurance Research Database 13 , reported positive associations of asthma and overall prostate cancer risk. 80 The Busselton Health Survey also found an increased risk of prostate cancer with any atopy 11 . Two studies examined the effect of allergen-specific immunoglobulin (IgE) positivity on prostate cancer risk and reported a positive and null association, respectively 14,15 . Apart from the mixed direction of the results, few of the aforementioned studies investigated the influence of AACs on the aggressiveness of prostate cancer 12,16 , and most studies were conducted in White populations. Recently, the inflammatory component of diet has been suggested to modulate the chronic inflammatory process 17-19 . The Dietary Inflammatory Index (DII ® ) was developed to quantify the inflammatory potential of diet in epidemiological studies. In the last five years, emerging studies have focused on the relationship between DII and the risk of prostate cancer, and most studies have reported a positive association between a pro-inflammatory diet and total prostate cancer risk 20-28 . Given the inflammatory potential of diet, it may modify the process of prostate cancer development or progression under chronic inflammatory conditions such as AACs. To further understand how AACs contribute to the development of prostate cancer, in particular, aggressive and fatal prostate cancer in a diverse population, we prospectively assessed the association between AACs and prostate cancer risk in the Multiethnic Cohort (MEC) study. We also evaluated the association of AACs with prostate cancer risk and survival by race/ethnicity, as well as whether a pro-inflammatory diet, as measured by DII modifies the association. 5.2 Methods Study population 81 The MEC is a prospective cohort study with more than 215,000 participants, primarily from five racial/ethnic groups (African-Americans, Japanese-Americans, Latinos, Native Hawaiians, and Whites), aged 45 to 75 years when enrolled between 1993 and 1996 in California and Hawaii 29 . Potential participants were identified through drivers’ licenses and supplemented with voter’s registration files in Hawaii and the Health Care Financing Administration files in California 29 . Upon enrollment, each participant completed an extensive questionnaire regarding demographic characteristics, anthropometric measurements, personal and family history of medical conditions, medication use, lifestyle (e.g., smoking history, physical activity), and dietary intake. Specifically, a self-administrative quantitative food frequency questionnaire (QFFQ) was used to measure the frequency and quantity of intake of over 180 food items that contribute substantially to the nutrition of each ethnic group in the MEC 30,31 . Participants were actively followed up approximately every five years (1998-2002, 2003-2008, 2008-2012, and 2012-2016) for updates on health conditions, uptake of screening tests for common diseases, and dietary patterns. The study was approved by the institutional review boards of the University of Hawaii and the University of Southern California. In the current study, we included 96,888 men at baseline of the MEC in the analysis. We excluded men who did not report one of the five major racial/ethnic groups (n=5,941), with an implausible dietary caloric intake (n=3,648), with a prior prostate cancer history diagnosed before baseline (n=2,887), and with missing data on AAC status or other key covariates at baseline (n=9,698). Participants were followed from cohort entry date to the date of prostate cancer diagnosis, death, or administrative censoring on December 31, 2017, whichever was the earliest (median follow-up time = 21.5 years). Incident prostate cancer, as well as the stage (localized/regional/metastatic) 82 and Gleason grade (low-grade [≤7]/ high-grade [(≥8)) at the time of diagnosis, were ascertained by linkage to the Surveillance, Epidemiology, and End Results (SEER) statewide cancer registries in Hawaii and California. Aggressive prostate cancer was defined as either regional or metastatic disease, or, localized disease with Gleason score ≥8. Prostate cancer deaths were determined by linkage to Hawaii or California death certificate files, supplemented by the National Death Index. Survival was considered from age at diagnosis to age at death or at administrative censoring on Dec 31 st , 2017. In total, 74,714 men were included in the analysis and 8,697 incident prostate cancer cases occurred during the follow-up, including 6,559 localized, 878 regional, 438 metastatic (822 cases had missingness on stages); 5,943 low-grade, 2,139 high-grade (615 had missingness on grades); 2,710 aggressive cases; and 1,170 prostate cancer deaths. Only 8,491 cases were included in the survival analysis because 206 cases had missing diagnosis dates, or had the same date of diagnosis as death. AACs and covariate assessment AAC status was assessed at baseline based on a self-reported question “whether a doctor have ever told you had an asthma, hay fever, skin allergy, food allergy or any other allergy condition” in the questionnaire. Self-reported previous use of antihistamines (allergy pills or shots) for AACs (was defined as ever taken for two times per week for one month or longer) and duration of use (one year or less, 2-5 years, >5 years) was also assessed at baseline. The use of antihistamines and duration of use among those who reported AACs may reflect severity of the condition, and were considered as an indicator of long-term heightened immune response. 83 We calculated the DII at baseline for each participant in the MEC as a measurement of the inflammatory potential of diet. The calculation of DII in MEC was also reported in previous studies 32,33 . Briefly, only 28 of the 45 food components were covered in the baseline QFFQ (carbohydrate; protein; total fat; saturated, monounsaturated, and polyunsaturated fats; ω-3 and ω-6 FAs; alcohol; fiber; cholesterol; vitamins A, B-6, B-12, C, D, and E; thiamin; riboflavin; niacin; iron; magnesium; zinc; selenium; folate; β-carotene; isoflavones; and caffeine); foods not included in the MEC were dropped from the DII calculation. For each participant, a z-standardized-score was calculated for each food component. The reference mean and standard deviation for the z score was derived from the dietary intake from surveys or studies conducted in 11 countries 34 . Energy-adjusted DII (E- DII) scores were calculated based on the caloric density of each food component (per 1000 kcal) to account for individual variation in overall energy intake 35 . We categorized participants into 4 groups based on the quartile distribution of E-DII in the overall study population: quartile 1 (≤- 2.44), quartile 2 (>-2.44 to ≤-0.88), quartile 3 (>-0.88 to ≤0.65), and quartile 4 (>0.65). A higher DII score represents a more pro-inflammatory diet, while a lower score represents a more anti- inflammatory composition of diet. Potential confounders obtained on the baseline questionnaire included education (≤12, 13–15, ≥16 years), BMI [<18.5 (underweight), 18.5-25 (normal), 25-29.9 (overweight), ≥30 kilogram/meter 2 (obese)], smoking status (never, former, current smokers), history of diabetes (no/yes), family history of prostate cancer (no/yes), and aspirin and/or statin use (ever use at least two times per week for one month or longer). PSA test utilization history (ever/never) was first collected on the second questionnaire in 1999-2002 and 96.5% of men completed this question. 84 Statistical analysis We used Cox proportional hazards regression models of prostate cancer events with age as the time metric to estimate Hazard Ratios (HRs) and 95% Confidence Intervals (CIs) for the association between AAC status and overall prostate cancer incidence, prostate cancer severity (aggressive, low-grade, high-grade, localized, regional, metastatic) and prostate cancer mortality. As a potential indicator of AAC severity, we also examined the association between AAC medication use and the duration of use and prostate cancer risk. The assumption of proportional hazards was met based on the review of cumulative sums of Martingale residuals 36 . In addition, we conducted survival analysis for men with incident prostate cancer diagnosed after baseline and examined the association between AACs and prostate cancer-specific and all-cause mortality. For all analyses, we first fitted a minimally adjusted model that included age at cohort entry and race/ethnicity, and then a fully adjusted model that included all covariates mentioned above to account for potential confounding. We additionally adjusted for stage and Gleason score in the survival analysis. To address potential effect modification, we stratified models by race/ethnicity, E-DII score and PSA screening, and the likelihood ratio test was used to test for homogeneity of the effects across strata. In sensitivity analyses, PSA screening was additionally included in multivariate models to examine the potential impact of PSA screening on the relationship between AACs and prostate cancer risk. All analyses were conducted using STATA 14.0 and R 3.6.0. A p<0.05 based on a two-tailed test was considered statistically significant. 85 5.3 Results Upon enrollment, 15,630 (20.9%) men reported having been diagnosed with AACs, of whom 5,481 (35.1%) reported past use of antihistamines, with 2,257 (14.4%) reporting use for more than 5 years. The proportions of men who reported a history of AACs in White, African American, Native Hawaiian, Japanese American, and Latino subgroups were 24.9%, 18.5%, 21.5%, 21.5%, and 16.7%, respectively. Table 1 shows the baseline characteristic among study participants by AAC status. The mean age at enrollment was 58.57 ± 8.87 years for men with AACs and 60.03 ± 8.76 for men without AACs (p<0.001). Compared to men without AACs, men with AACs were more likely to be White (31.3% vs. 25.0%), never smokers (31.5% vs. 29.7%), have a graduate college degree or higher education level (40.0% vs. 28.3%), have a family history of prostate cancer (8.2% vs. 7.2%), use aspirin and/or statins regularly (48.2% vs. 45.4%), and were less likely to have diabetes (15.5% vs. 17.8%). All the above differences were statistically significant (p<0.001). The distribution of E-DII was similar between men with and without AACs (p=0.26). Among 56,534 men who completed the question regarding history of PSA screening, men with AACs were more likely to report past PSA screening (50.6%) than men without AACs (44.0%, p<0.001). We detected no statistically significant associations between AACs and total prostate cancer incidence in the minimally adjusted model (HR = 1.00, 95%CI: 0.95-1.05) or the fully adjusted model (HR = 0.98, 95%CI: 0.93-1.03). Similarly, no association was observed between AACs and any prostate cancer subtype (Table 2). Men with and without antihistamine medication use had similar null associations with incidence of total prostate cancer or with any subtype 86 (Figure 1a). Among men who took antihistamines, those who reported 2-5 years of use at baseline showed a non-significant increased risk of total prostate cancer compared to those with less than 1 year of use (HR=1.21, 95%CI: 0.97-1.51; Figure 2a). Antihistamine use for 2-5 years was associated with significant or borderline significant increases in low-grade (HR=1.42, 95%CI: 1.09-1.85) and localized prostate cancer (HR=1.28, 95%CI: 0.99-1.64). However, these associations were not observed in men with over 5 years of antihistamine use. E-DII was not associated with prostate cancer incidence (HR per SD=0.98, 95%CI: 0.96-1.00), but was non- significantly associated with an increased risk of metastatic prostate cancer (HR per SD=1.07, 95%CI: 0.97-1.18; supplemental Table 1). A significantly decreased risk of prostate cancer mortality was observed among men who reported a history of AACs in both minimally (HR=0.78, 95%CI:0.67-0.92) and fully adjusted models (HR=0.79, 95%CI:0.67-0.92; Table 2). The inverse association with prostate cancer mortality was observed for men who did and did not report antihistamine use (Figure 1a). Among men diagnosed with prostate cancer, after additional adjustment for prostate cancer stage and grade, AACs were significantly associated with a reduced risk of prostate cancer-specific mortality (HR=0.75, 95%CI: 0.63-0.89), and a smaller reduction in risk of all-cause mortality (HR=0.93, 95%CI: 0.87-1.00; Table 2). Moreover, compared to men without AACs, men with AACs but without antihistamine use showed a statistically significant 23% reduction (HR=0.77, 95%CI:0.62-0.96) in prostate cancer-specific mortality, while men with antihistamine use showed a 31% reduction in prostate cancer-specific mortality (HR=0.69, 95%CI: 0.53-0.91) and a 13% reduction in all-cause mortality (HR=0.87, 95%CI: 0.78-0.97; Figure 1b). No significant 87 associations were observed between duration of AAC medication use and prostate cancer-specific or all-cause mortality (Figure 2b). Moreover, the association between antihistamines use – regardless of AACs status – with prostate cancer incidence or survival was null (result not shown). E-DII was non-significantly associated with higher risk of prostate cancer-specific mortality, and significantly associated with higher risk of all-cause mortality (HR per SD=1.04, 95%CI: 1.01- 1.07; supplemental Table 1). There was no significant evidence of heterogeneity in the association of AACs with prostate cancer risk or mortality by race/ethnicity (Table 3, p for heterogeneity>0.05 for all). An inverse association between AACs and prostate cancer-specific survival was observed in all groups (HRs<1), but only associations among Whites and Latinos were significant (p’s<0.05). Similarly, we did not detect evidence of heterogeneity in the association of AACs with prostate risk or mortality by E-DII (Table 3). Compared to men without AACs, men with AACs were more likely to report past PSA screening (OR=1.27, 95% CI: 1.22-1.33; Supplemental Table 2), with stronger association in men reporting past AAC medication use (OR=1.41, 95%CI: 1.32-1.50) than in men not taking medications (OR=1.20, 95%CI: 1.14-1.26). Additionally adjusting for PSA screening history did not meaningfully change any of the associations between AACs and prostate cancer risk or mortality (Supplemental Table 3). No statistically significant difference in these associations were observed when stratifying by PSA screening history (Table 3), although inverse associations were greater in men reporting no previous screening. 88 5.4 Discussion In this prospective analysis of 74,714 men from five racial/ethnic groups, AACs were not found to be associated with risk of prostate cancer. We did detect an inverse association between AACs with prostate mortality in all racial/ethnic groups that were independent of the positive association of AACs with PSA screening. Atopic diseases generally refer to immunoglobulin E (IgE)-mediated hypersensitivity reactions to common allergens in the environment. It is estimated that approximately 8% of the US adults aged 45 years and older suffer from asthma and 11% suffer from hay fever, representing a slight increase over the past two decades 37 . The role of an allergy-related immune response in the pathogenesis and progression of prostate tumors remains unclear. Evidence suggests that in the inflammatory microenvironment of the prostate, immune cells release reactive oxygen and nitrogen species as well as pro-inflammatory cytokines that induce genetic and epigenetic mutations in normal prostate cells that could lead to abnormal epithelial cell proliferation 38-42 . A hyper-reactive immune system also may have an anti-tumor effect through enhanced immunosurveillance. The cytokine profile of allergic patients is biased toward Th2 rather than Th1 cells 43 . Th2-mediated immunity activates a number of effector immune cells that are capable of recognizing and attacking cancer cells at an early stage, which could result in lower prostate cancer risk and/or better survival among men with disease. It also is important to note that AACs represent Type 1 hypersensitivity involving an immediate response mediated by IgE. So, even though atopic conditions may be chronic, the biologic mechanisms involved are characterized by acute response. Indeed, it has been observed that individuals with low levels of IgE appear to be at elevated risk of cancer through mechanisms related to tumor cell phagocytosis 44 . This is consistent with proposals to use IgE as 89 an anti-cancer therapy 45 . Mast cells, which are widely distributed in vascular tissue, are commonly associated with Type I hypersensitivity. These cells bind IgE with high affinity, producing TNF-  and granulocyte macrophage colony-stimulating factor (GM-CSF) in the tumor microenvironment. This observation is consistent with the idea of using tumor-targeted, IgE- sensitized mast cells as a platform for developing new cancer immunotherapies 46 . Clearly, the biologic mechanisms linking AACs and, by logical extensions, IgE is complex. Reflecting this complexity, our finding of no association of AACs with total prostate cancer risk is in agreement with results from other studies. A meta-analysis of 20 studies (5 case-control, 15 cohort studies) through 2015 reported no significant associations for atopy (RR=1.25, 95%CI: 0.74–2.10), hay fever (RR=1.04, 95%CI: 0.99–1.09) or any allergy (RR=0.96, 95%CI: 0.86–1.06) 47 . Another meta-analysis that combined results from 14 studies (7 case-control, 7 cohort studies) through 2015, also concluded an overall null association between asthma and prostate cancer risk (OR=0.99, 95% CI: 0.84-1.18) 48 . Similarly, a Mendelian randomization study based on the meta- analysis of GWAS summary statistics also noted no evidence of genetic predisposition to AACs and prostate cancer risk (OR=1.00, 95%CI: 0.94–1.05) 49 . The absence of an association may reflect a dynamic balance of immunosurveillance against tumor cells and tumor-promoting immune responses. Prostate cancer mortality and prostate cancer-specific survival were inversely associated with AACs. The relationship between AACs and prostate cancer mortality or severity has been rarely examined in epidemiological studies. Similar to our findings, the Health Professionals Follow-up Study reported significant protective effects of asthma on lethal (RR=0.71, 95% CI: 0.51-1.00) 90 and fatal (RR=0.64, 95% CI: 0.42-0.96) prostate cancer 10 . This is consistent with IgE-related hypotheses noted earlier 45,46 . Other studies examined asthma and/or allergies have shown no significant association with aggressive prostate cancer 12,50 or prostate cancer mortality 51 . The higher survival rate among men with AACs may imply a stronger immune surveillance role in the tumor microenvironment after prostate cancer onset. This is supported by the observation that AACs were not associated with localized or low-grade prostate cancer, whereas they were inversely associated with risk of disease progression (i.e., metastatic disease), and is consistent with the use of IgE in cancer treatment. This suggests that a hyper-allergic state may not prevent disease development at early stages, but may prevent tumor progression and spread to nearby tissues. In comparison with other studies, our study has several strengths. First, this population-based prospective study included a diverse racial and ethnic population, which results in more generalizable findings. Second, we were able to investigate the impact of PSA screening. Men with AACs are more likely to undergo PSA screening than their non-affected counterparts, which may allow for increased detection of prostate cancer, although we did not find a positive association between AACs and disease risk. The positive association between AACs and PSA screening may also result in early detection, downstaging of disease at the time of diagnosis, and treatment leading to a better survival. However, the results remained robust after adjusting for and stratifying by PSA screening history. Third, this is also the first study to examine and report on inflammatory components of diet and the lack of effect modification on the association between AACs and prostate cancer risk. 91 Despite its strengths, several limitations of our study should also be noted. First, we did not have information on the specific types of AACs in our study, and consequently we were only able to study the combined effect of all types of AAC. Second, we did not inquire about the complete list of AAC-related medications, such as glucocorticoids, which have been found to be associated with prostate cancer incidence. 12,50 However, we expect systemic glucocorticoids to play a relatively small role in explaining the inverse relationship between AACs and prostate cancer mortality, as few patients with AACs take glucocorticoids. Third, PSA screening history was only collected at a single time point, detailed information on timing and frequency of PSA screening test was not available. The PSA screening effect might not be fully accounted for in the PSA screening-adjusted and stratified analyses. Finally, as in any large prospective cohort, we are reliant on participant self-report for a wide range of variables, including those related to AAC symptomology and diet. These reports could be subject to a variety of information biases. In summary, we observed an inverse association between AACs and prostate cancer mortality across White, African-American, Native Hawaiian, Japanese-American, and Latino men that was independent of the effect of PSA screening. Further etiological research on the relationship between allergic response and prostate cancer progression, as well as mechanistic research focused on biological mechanisms is warranted. 92 5.5 References 1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer Journal for Clinicians. 2018;68(6):394-424. 2. Sarto AO. Risk factors for prostate cancer. In: Nicholas Vogelzang RL, MEdJerome P Richie,, ed. UptoDate.2020. 3. Heather H Cheng PSN. Genetic risk factors for prostate cancer. In: Jerome P Richie NV, Robert Lee,, ed. UpToDate.2020. 4. De Marzo AM, Platz EA, Sutcliffe S, et al. Inflammation in prostate carcinogenesis. Nat Rev Cancer. 2007;7(4):256-269. 5. Sfanos KS, De Marzo AM. Prostate cancer and inflammation: the evidence. Histopathology. 2012;60(1):199-215. 6. Johnson AR, Milner JJ, Makowski L. The inflammation highway: metabolism accelerates inflammatory traffic in obesity. Immunol Rev. 2012;249(1):218-238. 7. Paschos A, Pandya R, Duivenvoorden WC, Pinthus JH. Oxidative stress in prostate cancer: changing research concepts towards a novel paradigm for prevention and therapeutics. Prostate Cancer Prostatic Dis. 2013;16(3):217-225. 8. Marseglia L, Manti S, D'Angelo G, et al. Oxidative stress in obesity: a critical component in human diseases. Int J Mol Sci. 2014;16(1):378-400. 9. Sattayapiwat O, Wan P, Hernandez BY, Le Marchand L, Wilkens L, Haiman CA. Association of Anthropometric Measures With the Risk of Prostate Cancer in the Multiethnic Cohort. Am J Epidemiol. 2021;190(9):1770-1783. 10. Platz EA, Drake CG, Wilson KM, et al. Asthma and risk of lethal prostate cancer in the Health Professionals Follow-Up Study. Int J Cancer. 2015;137(4):949-958. 11. Talbot-Smith A, Fritschi L, Divitini ML, Mallon DF, Knuiman MW. Allergy, atopy, and cancer: a prospective study of the 1981 Busselton cohort. Am J Epidemiol. 2003;157(7):606-612. 12. Severi G, Baglietto L, Muller DC, et al. Asthma, asthma medications, and prostate cancer risk. Cancer Epidemiol Biomarkers Prev. 2010;19(9):2318-2324. 13. Su YL, Chou CL, Rau KM, Lee CT. Asthma and Risk of Prostate Cancer: A Population- Based Case-Cohort Study in Taiwan. Medicine (Baltimore). 2015;94(36):e1371. 14. Skaaby T, Nystrup Husemoen LL, Roswall N, Thuesen BH, Linneberg A. Atopy and development of cancer: a population-based prospective study. J Allergy Clin Immunol Pract. 2014;2(6):779-785. 93 15. Wang H, Rothenbacher D, Low M, Stegmaier C, Brenner H, Diepgen TL. Atopic diseases, immunoglobulin E and risk of cancer of the prostate, breast, lung and colorectum. Int J Cancer. 2006;119(3):695-701. 16. Weiss D, El-Zein M, Rousseau MC, Richard H, Karakiewicz PI, Parent ME. Asthma, allergy and the risk of prostate cancer: results from the Montreal PROtEuS study. Cancer Epidemiol. 2014;38(6):695-699. 17. Sears B, Ricordi C. Anti-inflammatory nutrition as a pharmacological approach to treat obesity. J Obes. 2011;2011:431985. 18. Ricordi C, Garcia-Contreras M, Farnetti S. Diet and Inflammation: Possible Effects on Immunity, Chronic Diseases, and Life Span. J Am Coll Nutr. 2015;34 Suppl 1:10-13. 19. Galland L. Diet and inflammation. Nutr Clin Pract. 2010;25(6):634-640. 20. Shivappa N, Jackson MD, Bennett F, Hébert JR. Increased Dietary Inflammatory Index (DII) Is Associated With Increased Risk of Prostate Cancer in Jamaican Men. Nutr Cancer. 2015;67(6):941-948. 21. Zucchetto A, Gini A, Shivappa N, et al. Dietary inflammatory index and prostate cancer survival. Int J Cancer. 2016;139(11):2398-2404. 22. Kim JH, Kim J. Index-Based Dietary Patterns and the Risk of Prostate Cancer. Clin Nutr Res. 2017;6(4):229-246. 23. Shivappa N, Niclis C, Coquet JB, Román MD, Hébert JR, Diaz MDP. Increased inflammatory potential of diet is associated with increased odds of prostate cancer in Argentinian men. Cancer Causes Control. 2018;29(9):803-813. 24. Accardi G, Shivappa N, Di Maso M, et al. Dietary inflammatory index and cancer risk in the elderly: A pooled-analysis of Italian case-control studies. Nutrition. 2019;63-64:205- 210. 25. Hoang DV, Shivappa N, Pham NM, Hebert JR, Binns CW, Lee AH. Dietary inflammatory index is associated with increased risk for prostate cancer among Vietnamese men. Nutrition. 2019;62:140-145. 26. McMahon DM, Burch JB, Hébert JR, et al. Diet-related inflammation and risk of prostate cancer in the California Men's Health Study. Ann Epidemiol. 2019;29:30-38. 27. Vidal AC, Oyekunle T, Howard LE, et al. Dietary inflammatory index (DII) and risk of prostate cancer in a case-control study among Black and White US Veteran men. Prostate Cancer Prostatic Dis. 2019;22(4):580-587. 28. Zhu Y, Li Q, Xu X. Dietary inflammatory index and the risk of prostate cancer: a dose- response meta-analysis. Eur J Clin Nutr. 2020;74(7):1001-1008. 94 29. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and Los Angeles: baseline characteristics. Am J Epidemiol. 2000;151(4):346-357. 30. Stram DO, Hankin JH, Wilkens LR, et al. Calibration of the dietary questionnaire for a multiethnic cohort in Hawaii and Los Angeles. American journal of epidemiology. 2000;151(4):358-370. 31. Sharma S, Murphy S, Wilkens L, Au D, Shen L, Kolonel L. Extending a multiethnic food composition table to include standardized food group servings. Journal of Food Composition and Analysis. 2003;16(4):485-495. 32. Harmon BE, Wirth MD, Boushey CJ, et al. The Dietary Inflammatory Index Is Associated with Colorectal Cancer Risk in the Multiethnic Cohort. The Journal of nutrition. 2017;147(3):430-438. 33. Park S-Y, Boushey CJ, Shvetsov YB, et al. Diet Quality and Risk of Lung Cancer in the Multiethnic Cohort Study. Nutrients. 2021;13(5):1614. 34. Shivappa N, Steck SE, Hurley TG, Hussey JR, Hébert JR. Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr. 2014;17(8):1689-1696. 35. Hébert JR, Shivappa N, Wirth MD, Hussey JR, Hurley TG. Perspective: The Dietary Inflammatory Index (DII)-Lessons Learned, Improvements Made, and Future Directions. Adv Nutr. 2019;10(2):185-195. 36. GRAMBSCH PM, THERNEAU TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81(3):515-526. 37. National Center for Health Statistics. Age-adjusted percentages (with standard errors) of selected respiratory diseases among adults aged 18 and over, by selected characteristics: United States, 2018. National Health Interview Survey. 38. De Marzo AM, Marchi VL, Epstein JI, Nelson WG. Proliferative inflammatory atrophy of the prostate: implications for prostatic carcinogenesis. Am J Pathol. 1999;155(6):1985- 1992. 39. Putzi MJ, De Marzo AM. Morphologic transitions between proliferative inflammatory atrophy and high-grade prostatic intraepithelial neoplasia. Urology. 2000;56(5):828-832. 40. Bostwick DG, Brawer MK. Prostatic intra-epithelial neoplasia and early invasion in prostate cancer. Cancer. 1987;59(4):788-794. 41. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell. 2010;140(6):883-899. 42. De Marzo AM, Platz EA, Sutcliffe S, et al. Inflammation in prostate carcinogenesis. Nat Rev Cancer. 2007;7(4):256-269. 95 43. Kay AB. Allergy and allergic diseases. First of two parts. N Engl J Med. 2001;344(1):30- 37. 44. Ferastraoaru D, Jordakieva G, Jensen-Jarolim E. The other side of the coin: IgE deficiency, a susceptibility factor for malignancy occurrence. World Allergy Organ J. 2021;14(1):100505. 45. Chauhan J, McCraw AJ, Nakamura M, et al. IgE Antibodies against Cancer: Efficacy and Safety. Antibodies (Basel). 2020;9(4). 46. Fereydouni M, Motaghed M, Ahani E, et al. Harnessing the Anti-Tumor Mediators in Mast Cells as a New Strategy for Adoptive Cell Transfer for Cancer. Front Oncol. 2022;12:830199. 47. Zhu J, Song J, Liu Z, et al. Association between allergic conditions and risk of prostate cancer: A Prisma-Compliant Systematic Review and Meta-Analysis. Sci Rep. 2016;6:35682. 48. Li W, Mao S, Tu M, et al. Asthma and the risk of prostate cancer. J Cancer Res Ther. 2018;14(Supplement):S571-S575. 49. Jiang X, Dimou NL, Zhu Z, et al. Allergy, asthma, and the risk of breast and prostate cancer: a Mendelian randomization study. Cancer Causes Control. 2020;31(3):273-282. 50. Beckmann K, Russell B, Josephs D, et al. Chronic inflammatory diseases, anti- inflammatory medications and risk of prostate cancer: a population-based case-control study. BMC Cancer. 2019;19(1):612. 51. Turner MC, Chen Y, Krewski D, Ghadirian P, Thun MJ, Calle EE. Cancer Mortality among US Men and Women with Asthma and Hay Fever. American Journal of Epidemiology. 2005;162(3):212-221. 96 Chapter Six: Summary and Future Directions Prostate cancer is a major global public health concern, being one of the most common cancers among men and accounts for a large proportion of deaths among men aged over 50 years. Understanding the risk factors for prostate cancer, including genetic and environmental factors, is crucial for developing effective prevention and treatment strategies. Studying prostate cancer in multiethnic populations is particularly important, as differences in genetic backgrounds and environmental exposures may reveal new insights into the etiology of the disease. This dissertation investigates a comprehensive spectrum of risk factors for prostate cancer in racially and ethnically diverse populations, addressing aspects such as germline variation, somatic mutations, and chronic inflammatory status. The study findings presented in this dissertation advance our understanding of prostate cancer risk factors in several ways. In Chapter 3, a total of 451 genetic risk variants for prostate cancer, including 86 novel variants, were identified through large-scale genome-wide association studies (GWAS) in multi-ancestry populations. A genetic risk score (GRS), incorporating the 451 identified variants, was also constructed and demonstrated improved characterization of prostate cancer risk across populations of different ancestries compared to previous GRS. These results highlight the potential of integrating the GRS with other screening tools for personalized medicine and risk prediction, thereby improving the efficacy of preventive measures and early detection strategies. Indeed, while our current expanded multi-ancestry GWAS has demonstrated the value of incorporating ancestrally diverse populations in identifying risk loci across ancestries and ancestry-specific loci, there are several ways to further improve it. For instance, although the sample size of non-European ancestry population in this study has notably increased compared to 97 previous ones, European ancestry samples still constitutes the majority of the samples included in the current GWAS. It is crucial to continue expanding the representation of non-European populations in future studies to ensure a more balanced and comprehensive understanding of genetic risk factors across diverse ancestries, and to enhance the generalizability of GWAS findings. Furthermore, most variants be examined in the current GWAS are common (MAF > 1% or 0.1%) in populations, highlighting the importance of investigating rare variants with potentially larger effect sizes to gain a more comprehensive understanding of the genetic architecture of prostate cancer. Additionally, as more longitudinal bio-cohorts become available, time-to-event analysis may offer a more robust estimation of the variant-prostate cancer association compared to logistic regression. Given that prostate cancer typically develops later in life, this type of analysis can better account for the potential misclassification of disease status by ensuring a more comparable age metric between cases and controls. Our GWAS results can also be combined with other types of data, such as transcriptomic, epigenomic, and proteomic information, to further elucidate the functional consequences of identified risk loci and provide a more comprehensive understanding of the molecular mechanisms underlying prostate cancer. On the other hand, the GRS constructed with the 451 risk variants for prostate cancer, has shown its effectiveness in differentiating the risk stratification of prostate cancer among multi-ancestry populations. This GRS can be a valuable resource in various research settings, as it represents the genetic predisposition to prostate cancer. The GRS can be used to study gene-environment interactions, examining how lifestyle factors such as diet, physical activity, smoking, and alcohol consumption may modify genetic risk for prostate cancer; the GRS can be integrated with other screening tools, such as prostate-specific antigen (PSA) tests, digital rectal exams (DRE), and imaging techniques, to create a more comprehensive risk prediction model for prostate cancer; the GRS can also be 98 leveraged to explore the underlying biological pathways and mechanisms linking genetic variants to prostate cancer risk. In Chapter 4, the association between clonal hematopoiesis of intermediate potential (CHIP), an age-related expansion of blood cells with somatic mutations, and prostate cancer was investigated in two large samples of European ancestry men. Only minimal evidence was found for the association between the overall CHIP mutation driver genes carrier status and total prostate cancer, with no evidence pertaining to prostate cancer aggressiveness. However, a significant association was observed between the most frequently mutated CHIP gene, DNMT3A, and total prostate cancer incidence. The biological effect of DNMT3A on prostate cancer warrant further investigation. Additionally, considering the varying frequency of CHIP mutations across different populations, investigating this association in other populations could offer a more comprehensive understanding of the effect of CHIP on prostate cancer risk. Furthermore, since CHIP gene mutations are typically rare events, it is worth considering the use of data with deeper sequencing depth to increase the detection of variants with lower variant allele frequencies (VAFs), which are more common and may be missed with many current whole-exome sequencing studies. Although our study focused on prostate cancer incidence, future research could also explore the association between CHIP and prostate cancer survival among men diagnosed with prostate cancer. Previous studies have suggested CHIP to be associated with inflammatory conditions in the blood, which may contribute to prostate cancer progression and may serve as a potential therapeutic target. In Chapter 5, the effect of atopic allergic conditions (AACs) on prostate cancer risk, disease severity, and survival was examined in a multi-ethnic cohort with over 25 years of follow-up. The 99 analysis did not find an association between AACs and overall prostate cancer incidence; however, a statistically significant inverse relationship was observed between AACs and prostate cancer mortality among men who were diagnosed with the disease. This association was consistent across all racial/ethnic groups examined. Although not statistically significant, the study also found that men with AACs were less likely to be diagnosed with high-grade or metastatic diseases, suggesting a role for enhanced immunosurveillance in individuals with AACs, protecting against from tumor progression and potentially contributing to the lower risk of death. These results shed lights on the potential role of immune system in prostate cancer development and progression. Further exploration of the mechanisms underlying the association between atopic allergic conditions and prostate cancer progression, such as from a molecular or cellular level, could provide valuable insights on our understanding of the complex interplay between the immune system and prostate cancer. Such studies could also inform the impact of immunotherapy approaches on the treatment of prostate cancer. Moreover, future studies could explore the potential of modulating the immune system to prevent or treat prostate cancer, possibly through interventions that mimic the effects of atopic allergic conditions. This could involve investigating the use of immune-modulating drugs or vaccines, as well as the potential of lifestyle changes, such as dietary modifications or exercise, in enhancing immune function. In summary, this dissertation advances our understanding of prostate cancer risk factors by examining genetic variants through multi-ancestry GWAS, investigating CHIP mutations, and analyzing atopic allergic conditions in relation to prostate cancer risk and survival across diverse populations. Building upon these findings, future studies should focus on elucidating the complex interplay of genetic, environmental, and immune system factors, with the ultimate goal of the 100 development of more effective and personalized strategies for prevention, early detection, and treatment of prostate cancer. 101 TABLES Chapter Four Table 1. Associations of four common CHIP genes and the aggregate of all identified CHIP genes with overall and aggressive prostate cancer risk. #Carriers (Carrier Frequencies) UK Biobank WESP Variant Allelic Fraction, median (min-max) UK Biobank WESP Case vs controls Aggressive vs non-aggressive Prostate cancer death vs non- aggressive M1 vs. non- aggressive Gene (#Variants in UKB/WESP) UK Biobank WESP Control (N=67,384) Case (N=2,118) Non-Agg (N=2,775) Agg (N=2,770) Death (N=2,052) M1 (N=467) HR (95% CI) P OR (95% CI) P OR (95% CI) P OR (95% CI) P DNMT3A (438/79) 0.10 (0.05- 0.56) 0.18 (0.10- 0.48) 929 (1.38%) 36 (1.70%) 53 (1.91%) 49 (1.77%) 37 (1.80%) 8 (1.71%) 1.04 (0.75- 1.45) 0.80 1.05 (0.70- 1.59) 0.80 1.09 (0.70- 1.70) 0.70 1.10 (0.50- 2.46) 0.81 ASXL1 (147/30) 0.11 (0.05- 0.48) 0.16 (0.10- 0.40) 255 (0.38%) 7 (0.33%) 14 (0.50%) 20 (0.72%) 13 (0.63%) 3 (0.64%) 0.77 (0.37- 1.62) 0.50 1.61 (0.79- 3.28) 0.19 1.61 (0.73- 3.54) 0.24 1.95 (0.52- 7.38) 0.33 TET2 (456/88) 0.14 (0.05- 0.75) 0.21 (0.10- 0.60) 690 (1.02%) 34 (1.61%) 41 (1.48%) 48 (1.73%) 30 (1.46%) 8 (1.71%) 0.97 (0.69- 1.36) 0.86 1.33 (0.86- 2.05) 0.20 1.14 (0.70- 1.87) 0.60 1.69 (0.75- 3.81) 0.21 JAK2 (1/1) 0.17 (0.07- 0.90) 0.13 (0.11- 0.19) 44 (0.07%) 1 (0.05%) 4 (0.14%) 2 (0.07%) 2 (0.10%) 0 (0%) 0.55 (0.08- 3.89) 0.55 0.65 (0.10- 4.24) 0.65 0.85 (0.14- 5.25) 0.86 -- -- All 55 UKB/52 WESP CHIP Genes (1,778/360) 0.11 (0.05- 0.90) 0.19 (0.10-1.0) 2,771 (4.11%) 103 (4.86%) 194 (6.99%) 198 (7.15%) 131 (6.38%) 33 (7.07%) 0.93 (0.76- 1.13) 0.46 1.14 (0.92- 1.41) 0.22 1.02 (0.81- 1.30) 0.84 1.22 (0.81- 1.83) 0.34 -- Estimate could not be reliably calculated due to lack of carriers. CHIP: Clonal hematopoiesis of indeterminate potential; WESP: Whole-Exome Sequencing Study in Prostate Cancer; M1: Metastatic prostate cancer; UKB: UK Biobank 102 Table 2. Associations of four common CHIP genes and the aggregate of all identified CHIP genes with age at blood draw and prostate cancer diagnosis in the UK Biobank. Age at blood draw 1 Age at prostate cancer diagnosis 2 Age at prostate cancer diagnosis adjusted for age at blood draw 2 Gene (#Variants) Age difference in years (95% CI) P Age difference in years (95% CI) P Age difference in years (95% CI) P DNMT3A (438) 4.08 (3.57-4.60) 4.4x10 -55 1.35 (-0.39-3.09) 0.13 -0.67 (-1.39-0.04) 0.07 ASXL1 (147) 6.22 (5.24-7.19) 1.3x10 -35 0.78 (-3.15-4.70) 0.70 -1.32 (-2.94-0.30) 0.11 TET2 (457) 3.74 (3.15-4.33) 2.6x10 -35 1.16 (-0.63-2.95) 0.20 0.52 (-0.22-1.26) 0.17 JAK2 (1) 3.83 (1.47-6.19) 1.5x10 -03 1.91 (-8.44-12.25) 0.72 2.39 (-1.87-6.66) 0.27 All 55 CHIP Genes (1,778) 3.25 (2.95-3.55) 2.4x10 -99 0.81 (-0.23-1.86) 0.13 -0.21 (-0.64-0.22) 0.34 1 Analysis performed in all participants 2 Analysis performed in incident prostate cancer cases CHIP: Clonal hematopoiesis of indeterminate potential 103 Chapter Five Table 1. Baseline characteristics by status of AACs in the 74,714 men in the Multiethnic Cohort study. Total No AACs With AACs P- value a N=74,714 N=59,084 N=15,630 No. (%) No. (%) No. (%) Age at baseline (years) ≤50yr 15245 (20.4%) 11320 (19.2%) 3925 (25.1%) <0.001 50-60yr 23248 (31.1%) 18304 (31.0%) 4944 (31.6%) 60-70yr 26112 (34.9%) 21154 (35.8%) 4958 (31.7%) 70-80yr 10109 (13.5%) 8306 (14.1%) 1803 (11.5%) Education level ≤8th grade 7276 (9.7%) 6307 (10.7%) 969 (6.2%) <0.001 9th-12th grade 22173 (29.7%) 18643 (31.6%) 3530 (22.6%) Vocational school/some college 22283 (29.8%) 17398 (29.4%) 4885 (31.3%) Graduated college or higher 22982 (30.8%) 16736 (28.3%) 6246 (40.0%) Ethnicity White 19662 (26.3%) 14765 (25.0%) 4897 (31.3%) <0.001 African American 9310 (12.5%) 7588 (12.8%) 1722 (11.0%) Native Hawaiian 5355 (7.2%) 4206 (7.1%) 1149 (7.4%) Japanese American 23267 (31.1%) 18261 (30.9%) 5006 (32.0%) Latino 17120 (22.9%) 14264 (24.1%) 2856 (18.3%) Smoking status Never 22492 (30.1%) 17571 (29.7%) 4921 (31.5%) <0.001 Former 38813 (51.9%) 30326 (51.3%) 8487 (54.3%) Current 13409 (17.9%) 11187 (18.9%) 2222 (14.2%) BMI Underweight 482 (0.6%) 390 (0.7%) 92 (0.6%) <0.001 Normal 26615 (35.6%) 20877 (35.3%) 5738 (36.7%) Overweight 34798 (46.6%) 27749 (47.0%) 7049 (45.1%) Obese 12819 (17.2%) 10068 (17.0%) 2751 (17.6%) Diabetes history 12923 (17.3%) 10499 (17.8%) 2424 (15.5%) <0.001 Family history of prostate cancer in father or brothers 5514 (7.4%) 4234 (7.2%) 1280 (8.2%) <0.001 E-DII b Quartile 1 (≤-2.44) 18661 (25.0%) 14678 (24.8%) 3983 (25.5%) 0.26 Quartile 2 (-2.44 ~-0.88) 18670 (25.0%) 14841 (25.1%) 3829 (24.5%) Quartile 3 (-0.88 ~0.65) 18664 (25.0%) 14773 (25.0%) 3891 (24.9%) Quartile 4 (>0.65) 18719 (25.1%) 14792 (25.0%) 3927 (25.1%) Regular use of aspirin or statin 34338 (46.0%) 26799 (45.4%) 7539 (48.2%) <0.001 PSA screening history c 25831 (45.7%) 19553 (44.2%) 6278 (50.9%) <0.001 Regular use of antihistamine 8398 (11.5%) 2917 (5.1%) 5,481 (35.1%) <0.001 a p-value was obtained by chi-square test. b Energy-adjusted Dietary Inflammatory Index. c PSA screening history was limited to men who completed the 2nd questionnaire. 104 Table 2. Hazard ratios of prostate cancer outcomes associated with AACs in the Multiethnic Cohort, 1993-2017 (N=74,714). Events Total Minimally adjusted HR (95% CI) a Fully adjusted HR (95% CI) b Incident prostate cancer Total 8697 74714 1.00 (0.95,1.05) 0.98 (0.93,1.03) Aggressive 2710 73742 0.97 (0.89,1.06) 0.96 (0.88,1.05) Low-grade 5943 74099 1.03 (0.97,1.09) 1.01 (0.95,1.07) High-grade 2139 74099 0.92 (0.83,1.02) 0.92 (0.83,1.02) Localized 6559 73892 1.00 (0.94,1.06) 0.98 (0.93,1.04) Regional 878 73892 1.06 (0.91,1.23) 1.04 (0.90,1.22) Metastatic 438 73892 0.86 (0.68,1.07) 0.88 (0.70,1.11) Prostate cancer mortality 1170 74714 0.78 (0.67,0.92) 0.79 (0.67,0.92) Prostate cancer survival c Prostate cancer-specific mortality 984 8491 0.70 (0.59,0.83) 0.75 (0.63,0.89) All-cause death 4854 8491 0.89 (0.83,0.95) 0.93 (0.87,1.00) a Adjusted for year at cohort entry and ethnicity b Adjusted for year at cohort entry, ethnicity, education, smoking, baseline BMI, diabetes, family history, and aspirin/statin intake. c Additionally adjusted for prostate cancer stage and grade. 105 Table 3. Hazard ratios of prostate cancer associated with aacs status among subgroups in the Multiethnic Cohort, 1993-2017. Subgroup Total Prostate Cancer Incidence Prostate Cancer Mortality Prostate Cancer Survival (Case only) Prostate Cancer All-Cause Event s Total HR (95% CI) P- het Event s Total HR (95% CI) P- het Event s Tota l HR (95% CI) P- het Event s Tota l HR (95% CI) P- het Ethnicity White 1978 1966 2 0.94 (0.85,1.05 ) 0.1 3 313 1966 2 0.79 (0.59,1.04 ) 0.8 8 245 1907 0.71 (0.51,0.99 ) 0.4 0 1091 1907 0.95 (0.82,1.10 ) 0.3 7 African America n 1775 9310 1.01 (0.89,1.14 ) 309 9310 0.90 (0.66,1.21 ) 264 1723 0.94 (0.68,1.31 ) 1172 1723 1.02 (0.87,1.18 ) Native Hawaiia n 452 5355 1.12 (0.89,1.40 ) 60 5355 0.63 (0.31,1.30 ) 57 448 0.55 (0.27,1.15 ) 249 448 0.72 (0.52,1.00 ) Japanese America n 2493 2326 7 0.97 (0.88,1.07 ) 219 2326 7 0.73 (0.50,1.06 ) 193 2462 0.73 (0.49,1.09 ) 1300 2462 0.90 (0.78,1.04 ) Latino 1999 1712 0 1.13 (1.01,1.27 ) 269 1712 0 0.70 (0.49,1.01 ) 225 1951 0.56 (0.37,0.84 ) 1042 1951 0.83 (0.71,0.98 ) E-DII a Quartile 1 2405 1866 1 1.06 (0.96,1.17 ) 0.6 4 339 1866 1 0.83 (0.62,1.10 ) 0.5 0 289 2350 0.78 (0.58,1.06 ) 0.0 7 1460 2350 0.93 (0.82,1.06 ) 0.8 1 Quartile 2 2269 1867 0 1.03 (0.93,1.14 ) 281 1867 0 0.75 (0.54,1.04 ) 238 2221 0.63 (0.43,0.92 ) 1290 2221 0.89 (0.77,1.03 ) Quartile 3 2122 1866 4 0.99 (0.89,1.10 ) 310 1866 4 0.65 (0.47,0.90 ) 258 2066 0.52 (0.36,0.77 ) 1133 2066 0.88 (0.76,1.03 ) Quartile 4 1901 1871 9 0.95 (0.85,1.07 ) 240 1871 9 0.93 (0.67,1.30 ) 199 1854 1.02 (0.72,1.45 ) 971 1854 0.99 (0.84,1.16 ) PSA screening b Never 2523 3070 3 0.90 (0.82,1.00 ) 0.5 5 323 3070 3 0.65 (0.47,0.90 ) 0.4 5 248 2444 0.64 (0.43,0.94 ) 0.4 9 1150 2444 0.91 (0.77,1.08 ) 0.7 9 Ever 2880 2583 1 0.95 (0.87,1.04 ) 361 2583 1 0.81 (0.62,1.06 ) 295 2808 0.80 (0.60,1.07 ) 1399 2808 0.91 (0.80,1.03 ) a Energy-adjusted Dietary Inflammatory Index. b The date of 2nd questionnaire completion was used as the entry time for time-to-event analysis. 106 Figures Chapter Three Figure 1. Manhattan plot of results from the multi-ancestry PCa meta-analysis. Purple and orange circles indicate previously known or novel risk variants, respectively, that were genome-wide significant in multi-ancestry or ancestry-specific meta-analyses. The plot is truncated at -log 10P=600. 107 Figure 2. Comparisons of genome-wide significant variants across populations. (a) Venn diagram of genome-wide significant variants (P < 5x10 -8 ) among European, African, Asian, and Hispanic populations. (b) Venn diagram of nominally significant variants (P<0.05) among European, African, Asian, and Hispanic populations; only common variants across all populations (MAF>1%, N=370) were included. (c) Comparison of ancestry-specific ORs between European and African, Asian, and Hispanic populations, respectively. Only variants with MAF>1% in both populations are compared; the number of variants is denoted in the lower right corner. Genome-wide significant variants among African, Asian, or Hispanic populations are highlighted in orange. The Pearson’s correlation coefficient between effect size and corresponding p-value are denoted in the upper left in each sub-panel. a c b 108 Figure 3. Percentage of cases in the top (quintile 1) and bottom (quintile 5) GRS risk quintile based on GRS 100, GRS 181, GRS 269, and GRS 451 in the multi- ancestry sample. GRS risk quintiles were categorized based on GRS distributions among controls. 40.5% 47.2% 49.3% 51.2% 7.5% 5.7% 5.0% 4.4% 0% 10% 20% 30% 40% 50% 60% GRS100 GRS181 GRS269 GRS451 Quintile 1 Quintile 5 109 Figure 4. Associations between standardized GRS and prostate cancer by populations. Odds ratios and 95% confidence intervals for one SD increase in (a) GRS 100, GRS 181, GRS 269, and GRS 451 and total prostate cancer risk by ancestry in the GWAS discovery studies; (b) GRS 269 and GRS 451 and total prostate cancer risk in the replication studies; (c) GRS 451 and total prostate cancer risk by age; (d) GRS 451 and GRS 400 and prostate cancer aggressiveness among prostate cancer cases in the GWAS discovery studies. Incremental percentage change of ORs were calculated for each comparison. 110 111 Figure 5. Cumulative absolute risk by age. Solid lines are the cumulative absolute risk for individuals in the top 16% GRS for African ancestry and top 20% for European ancestry. These GRS categories represent the percent of individuals in each population with at least a 2-fold increase in risk in comparison to the median GRS (as indicated in the inset distributions for African and European ancestries, respectively). Dashed horizontal lines indicate the lifetime absolute risk achieved at age 85 for the average (50% GRS) in African (11.6%) and European (7.8%) ancestry populations. Solid dots indicate the ages at which lifetime absolute risk levels are achieved for men of African ancestry in the top 16% GRS (age = 66 years) and men of European ancestry in the top 20% GRS (age = 69 years). 50 55 60 65 70 75 80 85 Age 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Absolute Risk African European 2−fold increase in risk within a population Average lifetime risk within a population African 16% at > 2−fold risk European 20% at > 2−fold risk 112 Extended Data Figure 1. Venn diagram of variants common (MAF>1%) among European, African, Asian, and Hispanic populations. Five variants with MAF ≤1% in all populations were included in the population with highest MAF (European population). 113 Extended Data Figure 2. Odds ratios and 95% confidence intervals for one SD increase in GRS451 and total prostate cancer risk in the GWAS discovery and replication sub-studies and meta-analysis by ancestry. 114 Chapter Four Figure 1. Manhattan plots of associations between CHIP genes and age at blood draw in the UK Biobank. CHIP: Clonal hematopoiesis of indeterminate potential. 115 Chapter Five Figure 1. Association between antihistamine use on a) prostate cancer incidence and mortality, and b) prostate cancer survival in the multiethnic cohort. a) b) Outcome Total No AACs (ref) AACs w/o med AACs w/ med Aggressive No AACs (ref) AACs w/o med AACs w/ med Low grade No AACs (ref) AACs w/o med AACs w/ med High grade No AACs (ref) AACs w/o med AACs w/ med Localized No AACs (ref) AACs w/o med AACs w/ med Regional No AACs (ref) AACs w/o med AACs w/ med Metastatic No AACs (ref) AACs w/o med AACs w/ med PCa mortality No AACs (ref) AACs w/o med AACs w/ med Events 6,891 1,135 671 2,168 341 201 4,659 810 474 1,731 254 154 5,178 873 508 680 124 74 365 44 29 982 118 70 Total 59,084 10,149 5,481 58,293 10,040 5,409 58,583 10,078 5,438 58,583 10,078 5,438 58,416 10,055 5,421 58,416 10,055 5,421 58,416 10,055 5,421 59,084 10,149 5,481 HR (95%CI) reference 1.00 (0.94−1.07) 1.04 (0.96−1.12) reference 0.93 (0.83−1.05) 1.00 (0.86−1.16) reference 1.05 (0.98−1.13) 1.07 (0.97−1.18) reference 0.87 (0.76−1.00) 0.98 (0.83−1.15) reference 1.01 (0.94−1.09) 1.04 (0.95−1.15) reference 1.06 (0.87−1.28) 1.10 (0.87−1.40) reference 0.77 (0.56−1.05) 0.90 (0.61−1.31) reference 0.78 (0.64−0.95) 0.77 (0.60−0.98) 0.5 0.75 1 1.25 1.5 Hazard Ratio (95% Confidence Inter val) Outcome PCa−specific death No AACs (ref) AACs w/o med AACs w/ med All cause death No AACs (ref) AACs w/o med AACs w/ med Events 831 95 58 3,934 589 331 Total 6,724 1,108 659 6,724 1,108 659 HR (95%CI) reference 0.77 (0.62−0.96) 0.69 (0.53−0.91) reference 0.97 (0.89−1.06) 0.87 (0.78−0.97) 0.5 0.75 1 1.25 1.5 Hazard Ratio (95% Confidence Inter val) 116 Figure 2. Association between the duration of antihistamine use on a) prostate cancer incidence and mortality, and b) prostate cancer survival in the multiethnic cohort. a) b) Outcome Total One Year or Less 2 to 5 Years >5 Years Aggressive One Year or Less 2 to 5 Years >5 Years Low grade One Year or Less 2 to 5 Years >5 Years High grade One Year or Less 2 to 5 Years >5 Years Localized One Year or Less 2 to 5 Years >5 Years Regional One Year or Less 2 to 5 Years >5 Years Metastatic One Year or Less 2 to 5 Years >5 Years PCa mortality One Year or Less 2 to 5 Years >5 Years Events 137 205 260 46 64 71 86 150 183 40 42 62 101 159 195 17 22 25 4 11 12 18 18 27 Total 1,160 1,525 2,257 1,143 1,507 2,227 1,149 1,512 2,242 1,149 1,512 2,242 1,145 1,512 2,229 1,145 1,512 2,229 1,145 1,512 2,229 1,160 1,525 2,257 HR (95%CI) reference 1.21 (0.97−1.51) 1.02 (0.83−1.26) reference 1.10 (0.75−1.61) 0.80 (0.55−1.17) reference 1.41 (1.08−1.84) 1.16 (0.89−1.50) reference 0.83 (0.54−1.29) 0.79 (0.52−1.18) reference 1.27 (0.99−1.63) 1.03 (0.81−1.32) reference 1.05 (0.56−2.00) 0.81 (0.43−1.51) reference 2.20 (0.69−6.97) 1.58 (0.50−4.96) reference 0.77 (0.40−1.50) 0.79 (0.43−1.45) 0.5 1 2 3 4 4.5 Hazard Ratio (95% Confidence Inter val) Outcome PCa−specific death One Year or Less 2 to 5 Years >5 Years All cause death One Year or Less 2 to 5 Years >5 Years Events 14 15 22 72 96 120 Total 133 202 255 133 202 255 HR (95%CI) reference 0.76 (0.36−1.61) 0.77 (0.38−1.53) reference 1.03 (0.75−1.41) 0.91 (0.67−1.24) 0.5 1 2 3 4 4.5 Hazard Ratio (95% Confidence Inter val) 117 Appendix: Supplementary Materials Chapter Three Supplementary Table 1. Inflation statistics () by population/study a  1000: Inflation for a study with 1000 cases and 1000 controls Study/Consortium No. SNPs and indels l l 1000 a African Ancestry Studies AAPC GWAS 17,406,370 1.031 1.007 Ghana Prostate Study 16,809,398 1.014 1.022 ProHealth Kaiser GWAS 17,119,655 1.009 1.011 ELLIPSE OncoArray 17,013,228 1.041 1.010 CA UG 21,035,217 1.017 1.014 IPM BioMe 22,287,350 1.088 1.302 BioVu 16,616,307 1.025 1.058 eMERGE 13,582,678 1.011 1.029 NCI-MD 17,401,209 1.024 1.062 MVP 22,829,524 1.045 1.004 Meta-Analysis 27,753,840 1.123 1.004 European Ancestry Studies 8 study meta analysis (ELLIPSE OncoArray/BPC3/CAPS1/CAPS2/PR ACTICAL iCOGS/Pegasus/UK GWAS1/UK GWAS2) 20,364,338 1.141 1.002 ProHealth Kaiser GWAS 10,097,661 1.058 1.005 UK Biobank 17,181,834 1.018 1.001 FinnGen 9,004,211 1.040 1.003 IPM BioMe 14,002,071 1.084 1.254 BioVu 9,715,097 1.023 1.008 eMERGE 7,756,364 1.047 1.009 PLCO 8,567,332 1.003 1.001 MVP 14,618,785 1.098 1.004 MDAnderson 9,059,886 1.060 1.043 Meta-Analysis 27,244,980 1.169 1.001 Hispanic Ancestry Studies ELLIPSE OncoArray 10,595,258 1.025 1.022 LAPC GWAS 10,441,344 1.005 1.005 ProHealth Kaiser GWAS 10,748,753 1.017 1.020 IPM BioMe 21,704,778 1.117 1.448 MVP 21,618,733 1.092 1.045 Meta-Analysis 26,567,548 1.116 1.017 Asian Ancestry Studies JAPC GWAS 8,559,772 1.018 1.018 ELLIPSE OncoArray 9,129,812 1.018 1.025 ProHealth Kaiser GWAS 9,270,529 1.015 1.029 PLCO 3,880,579 0.981 0.950 BBJ 9,878,439 1.058 1.004 Meta-Analysis 11,465,277 1.053 1.003 Multiethnic Meta-Analysis 42,428,922 1.158 1.001 118 Supplementary Table 2. Prostate cancer association results for all known and novel prostate cancer risk variants by race/ethnicity Multiethnic Analysis Marginal association Conditional association* rsID Conti et al., 269 SNPs Wang et al., 451 SNPs Known/N ovel 800 kb from known risk region Lead Variant Updated OR 95%CI P-value OR 95%CI P-value rs7542260 x Known drop 1.00 (0.99 - 1.02) 0.5693 rs2847344 x Known replaced 1.04 (1.03 - 1.05) 4.16E-12 rs616402 x Known new lead SNP 1.04 (1.03 - 1.05) 5.13E-13 rs10803412 x Known replaced 1.04 (1.03 - 1.05) 2.53E-09 rs34623941 x Known new lead SNP 1.04 (1.02 - 1.05) 1.36E-09 rs509526 x Novel - 1.03 (1.02 - 1.04) 2.89E-10 rs3813795 x Novel - 1.03 (1.02 - 1.04) 3.17E-08 rs12118566 x Novel - 1.03 (1.02 - 1.05) 4.70E-08 rs544780844 x Known replaced 1.05 (1.04 - 1.07) 5.92E-12 rs4420029 x Known new lead SNP 1.05 (1.04 - 1.06) 4.21E-17 rs6656241 x Novel - 1.04 (1.03 - 1.05) 3.55E-12 rs11206341 x Novel - 1.04 (1.03 - 1.05) 1.24E-09 rs365222 x Novel - 1.03 (1.02 - 1.04) 1.15E-11 rs3116433 x Novel - 1.03 (1.02 - 1.04) 1.07E-08 rs56391074 x Known replaced 1.04 (1.03 - 1.05) 4.38E-16 rs12131120 x Known new lead SNP 1.05 (1.04 - 1.06) 4.87E-19 1.05 (1.03 - 1.06) 2.13E-17 rs1413528 x Novel x - 1.03 (1.02 - 1.04) 1.70E-10 1.03 (1.02 - 1.04) 1.22E-08 rs79239530 x Novel - 1.26 (1.19 - 1.34) 5.78E-14 1.24 (1.16 - 1.33) 1.21E-09 rs141153466 x Novel - 1.15 (1.10 - 1.21) 2.03E-08 1.16 (1.1 - 1.22) 7.83E-08 rs2797527 x Novel - 1.03 (1.02 - 1.04) 1.80E-09 rs34379037 x Novel x - 1.10 (1.07 - 1.12) 2.65E-12 1.09 (1.06 - 1.12) 7.93E-10 rs116173394 x Novel x - 1.08 (1.05 - 1.10) 1.01E-09 1.08 (1.05 - 1.11) 1.17E-09 rs1811698 x x Known not replaced 1.07 (1.05 - 1.08) 2.94E-20 1.06 (1.04 - 1.07) 1.84E-13 rs607518 x x Known not replaced 1.07 (1.06 - 1.08) 3.64E-27 1.06 (1.05 - 1.08) 2.31E-21 rs10127983 x Known replaced 1.07 (1.06 - 1.08) 7.71E-36 rs11264734 x Known new lead SNP 1.07 (1.06 - 1.08) 1.93E-38 119 rs56103503 x x Known not replaced 1.07 (1.06 - 1.08) 2.58E-33 1.06 (1.05 - 1.07) 1.27E-27 rs147847496 x Known replaced 1.16 (1.13 - 1.20) 4.73E-19 rs146564277 x Known new lead SNP 1.19 (1.15 - 1.23) 6.91E-23 1.16 (1.12 - 1.2) 1.84E-16 rs184104770 x Known drop 1.16 (1.11 - 1.21) 3.41E-10 rs2485662 x Novel x - 1.03 (1.02 - 1.04) 7.95E-10 1.03 (1.02 - 1.04) 3.33E-09 rs80237341 x x Known not replaced 1.12 (1.09 - 1.16) 1.46E-14 1.12 (1.08 - 1.16) 6.83E-10 rs36042807 x Novel x - 1.05 (1.04 - 1.07) 5.54E-11 1.05 (1.04 - 1.07) 2.09E-10 rs6660538 x Known replaced 1.03 (1.02 - 1.04) 8.42E-11 rs11388275 x Known new lead SNP 1.04 (1.03 - 1.05) 1.36E-11 rs4075646 x x Known not replaced 1.07 (1.05 - 1.08) 2.53E-12 1.06 (1.04 - 1.08) 1.02E-10 rs72691600 x Novel x - 1.04 (1.02 - 1.05) 1.43E-09 1.03 (1.02 - 1.05) 5.09E-09 rs507603 x Known replaced 1.04 (1.03 - 1.05) 1.18E-09 rs555526 x Known new lead SNP 1.04 (1.03 - 1.05) 8.44E-10 rs56677963 x Known replaced 1.04 (1.03 - 1.05) 1.14E-15 rs3835740 x Known new lead SNP 1.04 (1.03 - 1.05) 9.80E-16 rs997343 x Novel - 1.03 (1.02 - 1.04) 5.79E-10 rs41333244 x Novel - 1.07 (1.05 - 1.10) 2.98E-09 rs138638958 x Known replaced 1.05 (1.03 - 1.06) 1.46E-16 rs2886543 x Known new lead SNP 1.04 (1.03 - 1.05) 8.89E-19 rs4245739 x Known replaced 1.09 (1.08 - 1.11) 7.91E-56 rs2926533 x Known new lead SNP 1.09 (1.08 - 1.11) 2.93E-56 rs708723 x x Known not replaced 1.04 (1.03 - 1.05) 3.74E-15 rs1294247 x Novel - 1.03 (1.02 - 1.04) 1.27E-11 rs11686272 x Known replaced 1.04 (1.03 - 1.05) 7.48E-13 rs10188360 x Known new lead SNP 1.04 (1.03 - 1.05) 6.94E-15 rs73913932 x x Known not replaced 1.08 (1.06 - 1.10) 7.64E-20 1.08 (1.06 - 1.1) 1.10E-18 rs1990613 x x Known not replaced 1.06 (1.05 - 1.07) 8.55E-33 1.06 (1.05 - 1.07) 1.16E-32 rs7602028 x Known replaced 1.05 (1.04 - 1.06) 8.26E-18 rs4668475 x Known new lead SNP 1.05 (1.04 - 1.06) 1.98E-18 rs9306894 x Known replaced 1.08 (1.07 - 1.09) 5.59E-53 rs9306895 x Known new lead SNP 1.08 (1.07 - 1.09) 4.55E-53 120 rs6722589 x Novel - 1.03 (1.02 - 1.04) 1.39E-09 rs6738169 x Known replaced 1.05 (1.04 - 1.06) 1.28E-18 rs4953671 x Known new lead SNP 1.05 (1.04 - 1.06) 4.16E-20 1.05 (1.04 - 1.06) 3.01E-18 rs7591218 x x Known not replaced 1.08 (1.07 - 1.09) 4.82E-56 1.08 (1.07 - 1.1) 3.24E-56 rs73923570 x Known not replaced 1.12 (1.08 - 1.17) 8.79E-09 1.15 (1.1 - 1.21) 4.99E-10 rs28514770 x Known drop 1.06 (1.05 - 1.07) 3.75E-25 rs12999901 x Novel - 1.03 (1.02 - 1.04) 2.80E-08 rs11125927 x x Known not replaced 1.13 (1.11 - 1.14) 3.46E-67 1.09 (1.07 - 1.1) 3.26E-33 rs58235267 x x Known not replaced 1.12 (1.11 - 1.14) 1.36E- 126 1.12 (1.1 - 1.13) 1.16E- 105 rs139283528 x x Known not replaced 1.21 (1.16 - 1.27) 2.78E-16 1.25 (1.19 - 1.32) 7.41E-18 rs74702681 x x Known not replaced 1.14 (1.10 - 1.18) 2.88E-13 rs2902065 x Novel - 1.03 (1.02 - 1.04) 2.60E-08 rs2028900 x Known replaced 1.08 (1.07 - 1.09) 3.04E-56 rs1078004 x Known new lead SNP 1.08 (1.07 - 1.09) 1.33E-56 rs2165108 x Known replaced 1.08 (1.06 - 1.10) 1.81E-17 rs4848397 x Known new lead SNP 1.08 (1.07 - 1.10) 1.49E-28 1.09 (1.07 - 1.1) 4.23E-31 rs11691517 x Known replaced 1.05 (1.04 - 1.07) 2.90E-20 rs3761705 x Known new lead SNP 1.04 (1.03 - 1.05) 1.24E-11 1.04 (1.03 - 1.05) 1.69E-14 rs4849090 x Novel - 1.04 (1.03 - 1.05) 1.70E-10 rs111595856 x Known replaced 1.08 (1.06 - 1.10) 6.41E-20 rs11900747 x Known new lead SNP 1.08 (1.06 - 1.10) 6.04E-22 1.08 (1.07 - 1.1) 2.07E-21 rs4848599 x Novel x - 1.04 (1.03 - 1.06) 2.93E-09 1.04 (1.03 - 1.06) 4.04E-09 rs10206072 x Known replaced 1.05 (1.04 - 1.07) 2.08E-13 rs6541731 x Known new lead SNP 1.05 (1.04 - 1.07) 1.14E-13 1.05 (1.04 - 1.07) 4.10E-13 rs1037137 x Novel - 1.04 (1.02 - 1.05) 2.24E-11 rs79120240 x Novel - 1.05 (1.03 - 1.06) 8.97E-10 1.04 (1.03 - 1.06) 5.03E-08 rs11315085 x Novel - 1.04 (1.03 - 1.05) 3.41E-09 1.04 (1.02 - 1.05) 1.96E-08 rs79256286 x Novel - 1.03 (1.02 - 1.05) 1.09E-09 rs16854905 x Known replaced 1.04 (1.03 - 1.06) 2.28E-09 rs11892008 x Known new lead SNP 1.05 (1.03 - 1.06) 2.00E-10 121 rs77167534 x Known replaced 1.23 (1.20 - 1.25) 2.69E-90 rs1574259 x Known new lead SNP 1.22 (1.20 - 1.25) 1.57E-90 1.23 (1.21 - 1.26) 2.55E-92 rs34925593 x x Known not replaced 1.05 (1.04 - 1.06) 5.24E-21 1.05 (1.04 - 1.06) 2.23E-22 rs7578920 x Novel x - 1.05 (1.03 - 1.07) 3.19E-09 1.05 (1.03 - 1.07) 3.41E-09 rs67228975 x Novel - 1.04 (1.03 - 1.05) 1.12E-10 rs1861270 x x Known not replaced 1.05 (1.03 - 1.06) 1.36E-16 rs12621900 x Known replaced 1.03 (1.02 - 1.05) 6.56E-09 rs57314858 x Known new lead SNP 1.03 (1.02 - 1.04) 6.83E-11 rs6733869 x Novel - 1.03 (1.02 - 1.04) 5.12E-10 rs13395911 x Novel - 1.03 (1.02 - 1.04) 4.54E-11 rs74001374 x x Known not replaced 1.15 (1.12 - 1.18) 9.80E-28 1.13 (1.1 - 1.17) 6.93E-16 rs2292884 x Known replaced 1.06 (1.04 - 1.07) 7.48E-24 rs58057291 x Known new lead SNP 1.07 (1.05 - 1.08) 1.53E-28 1.06 (1.05 - 1.07) 4.63E-22 rs2074840 x Known replaced 1.03 (1.02 - 1.04) 3.30E-08 rs2097976 x Known new lead SNP 1.06 (1.04 - 1.07) 1.10E-16 1.05 (1.03 - 1.06) 4.00E-12 rs77559646 x x Known not replaced 1.25 (1.20 - 1.29) 1.40E-34 1.27 (1.23 - 1.32) 9.18E-39 rs77482050 x x Known not replaced 1.44 (1.36 - 1.52) 1.06E-38 1.42 (1.34 - 1.5) 2.70E-36 rs76832527 x x Known not replaced 1.10 (1.08 - 1.11) 7.86E-43 1.10 (1.08 - 1.11) 5.74E-44 rs60985508 x Known not replaced 1.11 (1.08 - 1.14) 1.15E-13 1.09 (1.04 - 1.14) 1.15E-13 rs6550597 x x Known not replaced 1.03 (1.02 - 1.04) 4.35E-10 1.03 (1.02 - 1.04) 4.64E-09 rs4688916 x Novel x - 1.03 (1.02 - 1.04) 1.24E-08 1.03 (1.02 - 1.04) 1.24E-08 rs7618603 x Known replaced 1.06 (1.05 - 1.07) 3.31E-23 rs9815122 x Known new lead SNP 1.06 (1.05 - 1.07) 1.92E-23 rs34680713 x Known replaced 1.05 (1.04 - 1.07) 9.26E-12 rs140813356 x Known new lead SNP 1.05 (1.04 - 1.07) 6.57E-15 rs73121401 x Novel - 1.07 (1.05 - 1.10) 9.88E-09 rs13091518 x x Known not replaced 1.04 (1.03 - 1.05) 5.24E-16 1.04 (1.03 - 1.05) 1.14E-15 rs6147879 x Novel x - 1.05 (1.04 - 1.07) 7.19E-09 1.05 (1.03 - 1.07) 2.19E-08 rs77973813 x Novel x - 1.05 (1.04 - 1.06) 2.02E-23 1.03 (1.02 - 1.04) 7.40E-10 rs143745027 x Known replaced 1.20 (1.17 - 1.22) 7.85E-86 rs78267036 x Known new lead SNP 1.21 (1.18 - 1.23) 2.19E-95 1.16 (1.14 - 1.18) 3.77E-55 122 rs7628934 x Known replaced 1.11 (1.10 - 1.12) 9.32E-99 rs4527381 x Known new lead SNP 1.11 (1.10 - 1.13) 6.73E- 108 1.08 (1.07 - 1.09) 2.18E-47 rs114810266 x Known replaced 1.15 (1.12 - 1.19) 4.74E-17 rs77464751 x Known new lead SNP 1.24 (1.20 - 1.28) 6.39E-39 1.18 (1.14 - 1.22) 3.59E-22 rs1283104 x Known replaced 1.05 (1.04 - 1.06) 3.44E-24 rs1283108 x Known new lead SNP 1.05 (1.04 - 1.06) 9.93E-25 1.05 (1.04 - 1.06) 5.98E-22 rs151038334 x x Known not replaced 1.07 (1.05 - 1.09) 9.29E-13 1.06 (1.04 - 1.08) 1.36E-10 rs2271494 x x Known not replaced 1.08 (1.07 - 1.09) 9.92E-54 rs72960383 Known replaced 1.05 (1.03 - 1.08) 1.20E-06 rs58418703 x Known new lead SNP 1.09 (1.06 - 1.12) 7.07E-11 rs184923198 x Novel - 1.53 (1.33 - 1.77) 4.54E-09 1.44 (0.834 - 2.49) 4.54E-09 rs2811476 x x Known not replaced 1.10 (1.08 - 1.11) 5.04E-65 1.10 (1.09 - 1.11) 2.45E-66 rs35006112 x Known replaced 1.07 (1.06 - 1.09) 1.80E-28 rs4857898 x Known new lead SNP 1.07 (1.06 - 1.09) 1.37E-28 1.07 (1.06 - 1.08) 1.24E-26 rs61792779 x Novel - 1.04 (1.03 - 1.05) 3.07E-11 rs835648 x Novel - 1.03 (1.02 - 1.04) 7.19E-09 rs1457063 x x Known not replaced 1.04 (1.03 - 1.05) 1.37E-13 rs7650602 x Known replaced 1.05 (1.04 - 1.06) 2.36E-20 rs1344674 x Known new lead SNP 1.05 (1.04 - 1.06) 1.97E-20 rs1879422 x Novel - 1.05 (1.04 - 1.07) 5.16E-10 rs145053401 x Known replaced 1.06 (1.04 - 1.08) 1.98E-10 rs1426607 x Known new lead SNP 1.07 (1.05 - 1.09) 1.41E-13 rs55681402 x Novel - 1.05 (1.03 - 1.06) 5.42E-09 rs1317365 x Novel - 1.03 (1.02 - 1.04) 2.98E-08 rs2293607 x x Known not replaced 1.04 (1.03 - 1.06) 7.47E-14 1.06 (1.04 - 1.07) 8.98E-21 rs78416326 x x Known not replaced 1.18 (1.17 - 1.20) 1.20E- 144 1.20 (1.19 - 1.22) 3.30E- 169 rs66512269 x Known replaced 1.05 (1.03 - 1.07) 5.42E-09 rs6784631 x Known new lead SNP 1.06 (1.04 - 1.07) 6.13E-11 1.10 (1.08 - 1.11) 1.22E-27 rs144600367 x Known replaced 1.04 (1.03 - 1.05) 3.70E-13 rs9836594 x Known new lead SNP 1.04 (1.03 - 1.05) 1.26E-13 1.04 (1.03 - 1.05) 7.06E-14 123 rs3136601 x Novel x - 1.03 (1.02 - 1.04) 2.05E-12 1.03 (1.02 - 1.04) 2.63E-12 rs263023 x Novel - 1.05 (1.04 - 1.07) 2.29E-11 rs17804499 x x Known not replaced 1.15 (1.12 - 1.18) 8.57E-29 1.19 (1.16 - 1.22) 3.89E-39 rs72649118 x Novel x - 1.13 (1.08 - 1.17) 1.30E-08 1.17 (1.12 - 1.22) 5.93E-13 rs13142786 x Known replaced 1.04 (1.04 - 1.05) 7.09E-20 rs13131954 x Known new lead SNP 1.05 (1.04 - 1.06) 4.18E-20 1.07 (1.06 - 1.08) 9.39E-37 rs9993810 x Novel - 1.04 (1.03 - 1.05) 1.05E-13 rs6853490 x x Known not replaced 1.08 (1.07 - 1.09) 1.38E-54 rs13103835 x Novel - 1.03 (1.02 - 1.04) 2.15E-08 rs7679673 x Known replaced 1.12 (1.11 - 1.13) 1.85E- 116 rs10007915 x Known new lead SNP 1.12 (1.11 - 1.13) 9.28E- 118 1.10 (1.09 - 1.11) 7.79E-64 rs17035310 x Known replaced 1.13 (1.11 - 1.15) 2.51E-60 rs34316731 x Known new lead SNP 1.14 (1.12 - 1.15) 1.30E-66 1.07 (1.05 - 1.09) 5.65E-17 rs7689017 x Novel - 1.03 (1.02 - 1.05) 2.87E-08 1.04 (1.03 - 1.05) 1.04E-10 rs10025078 x Novel - 1.03 (1.02 - 1.04) 3.64E-08 1.03 (1.02 - 1.04) 8.54E-11 rs77821238 x Known replaced 1.05 (1.04 - 1.07) 3.45E-14 rs1402672 x Known new lead SNP 1.04 (1.03 - 1.05) 8.78E-16 rs72725734 x x Known not replaced 1.06 (1.04 - 1.07) 2.91E-18 rs147762399 x Known replaced 1.06 (1.04 - 1.08) 4.64E-09 rs77063077 x Known new lead SNP 1.06 (1.04 - 1.08) 3.88E-09 rs1521717 x Novel - 1.03 (1.02 - 1.04) 2.16E-08 rs10021065 x Novel - 1.04 (1.03 - 1.05) 5.56E-12 rs2242652 x x Known not replaced 1.15 (1.14 - 1.17) 6.44E- 111 1.14 (1.12 - 1.15) 1.74E-86 rs71595003 x Known replaced 1.15 (1.12 - 1.19) 4.55E-20 rs13190087 x Known new lead SNP 1.14 (1.11 - 1.17) 2.79E-20 1.13 (1.1 - 1.16) 1.38E-16 rs2736098 x x Known not replaced 1.09 (1.07 - 1.10) 9.34E-49 1.07 (1.06 - 1.08) 1.94E-32 rs4975758 x Known replaced 1.10 (1.09 - 1.11) 1.54E-89 rs199577062 x Known new lead SNP 1.13 (1.12 - 1.15) 2.09E- 115 1.13 (1.12 - 1.14) 2.73E- 109 rs71599622 x Known replaced 1.03 (1.02 - 1.04) 6.59E-07 124 rs61169618 x Known new lead SNP 1.03 (1.02 - 1.04) 3.47E-08 rs10941370 x Known drop 1.03 (1.02 - 1.04) 2.06E-07 rs35211243 x Novel - 1.04 (1.03 - 1.05) 2.48E-13 rs1482675 x Known replaced 1.03 (1.02 - 1.04) 6.71E-10 rs1482680 x Known new lead SNP 1.04 (1.03 - 1.05) 4.29E-12 rs6895682 x Novel - 1.03 (1.02 - 1.04) 9.69E-10 rs62357764 x Novel - 1.04 (1.03 - 1.06) 4.25E-09 rs9292122 x Known replaced 1.03 (1.02 - 1.04) 2.25E-08 rs33329 x Known new lead SNP 1.04 (1.03 - 1.05) 7.48E-11 rs6884284 x Novel - 1.03 (1.02 - 1.04) 2.05E-09 1.03 (1.02 - 1.04) 3.29E-09 rs139921834 x Novel - 1.03 (1.02 - 1.05) 9.10E-09 1.03 (1.02 - 1.05) 3.68E-08 rs4704108 x Novel - 1.04 (1.02 - 1.05) 1.67E-10 rs10793821 x Known replaced 1.06 (1.05 - 1.07) 3.02E-33 rs11242218 x Known new lead SNP 1.07 (1.05 - 1.08) 6.34E-35 rs4509063 x Novel - 1.03 (1.02 - 1.04) 1.08E-11 rs76551843 x Known replaced 1.27 (1.18 - 1.37) 3.38E-10 rs116483731 x Known new lead SNP 1.31 (1.23 - 1.39) 2.14E-17 rs72848121 x Novel - 1.06 (1.04 - 1.07) 2.16E-09 rs9686557 x x Known not replaced 1.04 (1.03 - 1.05) 1.01E-13 rs144442485 x Novel - 1.06 (1.04 - 1.08) 8.06E-09 rs61739424 x x Known not replaced 1.09 (1.07 - 1.12) 1.24E-13 1.09 (1.07 - 1.12) 7.13E-13 rs2672843 x Known replaced 1.04 (1.03 - 1.05) 2.67E-16 rs10043416 x Known new lead SNP 1.05 (1.04 - 1.06) 1.34E-17 1.04 (1.03 - 1.05) 1.33E-15 rs566243 x Novel x - 1.03 (1.02 - 1.05) 1.26E-10 1.03 (1.02 - 1.04) 4.41E-09 rs2814811 x x Known not replaced 1.04 (1.03 - 1.05) 4.31E-15 rs2018336 x Known replaced 1.05 (1.04 - 1.07) 3.60E-16 rs4713266 x Known new lead SNP 1.04 (1.03 - 1.05) 1.65E-16 1.04 (1.03 - 1.05) 7.06E-18 rs4713343 x Novel x - 1.03 (1.02 - 1.05) 4.32E-10 1.04 (1.02 - 1.05) 1.13E-10 rs6927369 x Known replaced 1.07 (1.06 - 1.08) 3.95E-30 rs9348465 x Known new lead SNP 1.07 (1.06 - 1.08) 1.65E-30 1.07 (1.05 - 1.08) 4.99E-28 rs4269363 x Known replaced 1.04 (1.03 - 1.05) 1.64E-12 125 rs9466096 x Known new lead SNP 1.04 (1.03 - 1.05) 7.47E-17 1.04 (1.03 - 1.05) 1.02E-14 rs12665509 x Known replaced 1.04 (1.03 - 1.05) 7.59E-17 rs2182997 x Known new lead SNP 1.04 (1.03 - 1.05) 1.17E-17 1.04 (1.03 - 1.05) 1.54E-18 rs9393282 x Novel x - 1.04 (1.03 - 1.05) 2.96E-10 1.04 (1.03 - 1.06) 1.23E-11 rs35159226 x Known drop 1.04 (1.02 - 1.05) 1.42E-08 rs375611687 x Known replaced 1.06 (1.05 - 1.07) 4.35E-22 rs16896742 x Known new lead SNP 1.05 (1.04 - 1.06) 6.11E-23 1.04 (1.03 - 1.05) 8.17E-13 rs2844626 x Novel x - 1.06 (1.05 - 1.07) 1.14E-27 1.05 (1.04 - 1.06) 3.44E-18 rs9275160 x x Known not replaced 1.04 (1.03 - 1.05) 6.37E-17 1.04 (1.03 - 1.05) 1.95E-12 rs9469899 x Known replaced 1.04 (1.03 - 1.05) 1.30E-15 rs2744961 x Known new lead SNP 1.04 (1.03 - 1.05) 1.23E-15 1.04 (1.03 - 1.05) 4.00E-17 rs262936 x Novel - 1.04 (1.03 - 1.05) 2.46E-12 1.04 (1.03 - 1.05) 1.81E-14 rs4714485 x Known replaced 1.11 (1.09 - 1.12) 2.24E-86 rs4714486 x Known new lead SNP 1.11 (1.09 - 1.12) 1.56E-86 rs4711748 x Novel x - 1.04 (1.03 - 1.05) 1.22E-13 1.03 (1.02 - 1.05) 1.93E-09 rs9472120 x Known replaced 1.04 (1.03 - 1.05) 7.93E-17 rs9472121 x Known new lead SNP 1.04 (1.03 - 1.05) 7.07E-17 1.04 (1.03 - 1.05) 1.19E-12 rs594260 x Novel - 1.03 (1.02 - 1.04) 8.96E-12 rs77054685 x Novel - 1.04 (1.02 - 1.05) 2.30E-08 rs16881098 x Novel - 1.05 (1.03 - 1.07) 8.97E-09 rs9443189 x Known replaced 1.05 (1.04 - 1.07) 4.02E-16 rs6906615 x Known new lead SNP 1.05 (1.04 - 1.07) 1.97E-16 rs146575246 x Novel - 1.03 (1.02 - 1.04) 3.66E-08 rs62430067 x Novel - 1.05 (1.03 - 1.06) 9.79E-12 rs2038542 x Known replaced 1.06 (1.05 - 1.08) 8.48E-21 rs78550764 x Known new lead SNP 1.08 (1.07 - 1.10) 5.54E-26 rs339351 x Known replaced 1.11 (1.10 - 1.13) 1.55E-95 rs339328 x Known new lead SNP 1.11 (1.10 - 1.13) 4.44E-96 rs7740107 x Novel - 1.05 (1.03 - 1.06) 3.45E-14 rs71681861 x Known replaced 1.04 (1.03 - 1.06) 1.30E-12 rs228867 x Known new lead SNP 1.04 (1.03 - 1.06) 3.20E-17 1.05 (1.03 - 1.06) 3.17E-17 126 rs10499188 x Novel x - 1.09 (1.06 - 1.12) 4.29E-09 1.08 (1.05 - 1.12) 4.29E-09 rs34379299 x Novel - 1.04 (1.02 - 1.05) 2.02E-09 rs13215045 x x Known not replaced 1.07 (1.06 - 1.08) 9.41E-40 rs963800 x Known replaced 1.06 (1.04 - 1.07) 5.61E-22 rs57022984 x Known new lead SNP 1.06 (1.05 - 1.08) 5.34E-18 1.07 (1.05 - 1.08) 3.65E-19 rs4646284 x x Known not replaced 1.17 (1.16 - 1.18) 3.21E- 184 1.17 (1.16 - 1.18) 2.13E- 176 rs4513875 x Known replaced 1.04 (1.03 - 1.05) 3.23E-12 rs12671113 x Known new lead SNP 1.04 (1.03 - 1.05) 3.62E-13 rs11452686 x Known replaced 1.03 (1.02 - 1.04) 1.72E-06 rs6973059 x Known new lead SNP 1.04 (1.03 - 1.06) 1.30E-13 1.04 (1.03 - 1.05) 1.60E-08 rs9655205 x Known replaced 1.09 (1.08 - 1.10) 1.48E-46 rs10713532 x Known new lead SNP 1.09 (1.08 - 1.10) 1.43E-46 1.09 (1.08 - 1.1) 3.09E-42 rs35389879 x x Known not replaced 1.04 (1.03 - 1.05) 6.64E-11 1.03 (1.02 - 1.04) 6.81E-09 rs5882865 x Novel - 1.03 (1.02 - 1.05) 7.25E-09 1.03 (1.02 - 1.04) 7.25E-09 rs6956484 x Known replaced 1.08 (1.07 - 1.09) 3.36E-55 rs940526 x Known new lead SNP 1.07 (1.06 - 1.08) 4.54E-41 1.04 (1.03 - 1.05) 1.97E-16 rs10486567 x x Known not replaced 1.15 (1.14 - 1.16) 5.42E- 131 1.14 (1.12 - 1.15) 2.10E- 106 rs6964685 x Novel - 1.05 (1.03 - 1.06) 4.93E-10 rs12701838 x Known replaced 1.06 (1.05 - 1.07) 7.26E-24 rs12701832 x Known new lead SNP 1.06 (1.05 - 1.07) 2.05E-25 rs700755 x Novel x - 1.04 (1.03 - 1.05) 1.31E-19 1.04 (1.03 - 1.05) 1.08E-14 rs834608 x Known replaced 1.04 (1.03 - 1.05) 9.15E-19 rs11770955 x Known new lead SNP 1.05 (1.04 - 1.06) 1.09E-19 1.04 (1.03 - 1.05) 4.95E-15 rs5884169 x Novel - 1.03 (1.02 - 1.04) 6.85E-09 rs1019617 x Novel - 1.03 (1.02 - 1.04) 4.19E-09 rs6955627 x Known replaced 1.04 (1.03 - 1.05) 3.87E-08 rs141155748 x Known new lead SNP 1.04 (1.03 - 1.06) 1.77E-08 rs4727386 x Known replaced 1.10 (1.09 - 1.11) 1.83E-67 rs6465657 x Known new lead SNP 1.10 (1.09 - 1.11) 2.71E-77 rs870167 x Known replaced 1.06 (1.04 - 1.07) 2.44E-11 127 rs62494897 x Known new lead SNP 1.06 (1.05 - 1.08) 1.63E-12 rs2572375 x Known replaced 1.04 (1.02 - 1.05) 2.48E-08 rs200170336 x Known new lead SNP 1.03 (1.02 - 1.04) 1.86E-08 rs4871844 x Novel x - 1.04 (1.03 - 1.05) 3.94E-13 1.04 (1.03 - 1.05) 1.36E-12 rs6557704 x Known replaced 1.04 (1.03 - 1.05) 1.68E-12 rs992530 x Known new lead SNP 1.04 (1.03 - 1.05) 1.08E-12 1.07 (1.05 - 1.08) 3.76E-28 rs1160267 x x Known not replaced 1.16 (1.15 - 1.17) 1.38E- 207 1.17 (1.15 - 1.18) 5.44E- 215 rs141811748 x x Known not replaced 1.06 (1.05 - 1.08) 4.31E-17 1.07 (1.05 - 1.08) 1.12E-17 rs12677206 x x Known not replaced 1.04 (1.03 - 1.05) 3.62E-16 1.04 (1.03 - 1.05) 8.02E-16 rs11467 x x Known not replaced 1.03 (1.02 - 1.04) 8.09E-10 rs876957 x Novel - 1.03 (1.02 - 1.05) 8.57E-11 rs75110624 x Novel - 1.06 (1.04 - 1.09) 8.82E-11 rs34368548 x Novel - 1.04 (1.03 - 1.05) 3.57E-13 rs11786992 x Novel - 1.03 (1.02 - 1.04) 4.53E-08 rs77244444 x Novel - 1.10 (1.06 - 1.14) 2.59E-08 rs73700335 x Known replaced 1.05 (1.03 - 1.06) 1.78E-11 rs4735020 x Known new lead SNP 1.04 (1.03 - 1.06) 9.68E-15 1.04 (1.03 - 1.05) 7.02E-12 rs13253117 x Novel x - 1.04 (1.03 - 1.06) 6.22E-13 1.04 (1.03 - 1.05) 5.11E-11 rs57588856 x Known drop 1.05 (1.04 - 1.06) 1.14E-18 rs902126 x Known new lead SNP 1.06 (1.05 - 1.07) 8.05E-23 1.05 (1.03 - 1.06) 5.77E-15 rs6984837 x Known replaced 1.08 (1.07 - 1.09) 2.98E-44 rs35358635 x Known new lead SNP 1.08 (1.07 - 1.09) 5.86E-52 1.06 (1.05 - 1.07) 2.45E-38 rs7011138 x Known drop 1.05 (1.04 - 1.07) 4.07E-15 rs7463326 x Known replaced 1.17 (1.16 - 1.19) 2.24E- 173 rs201495744 x Known new lead SNP 1.15 (1.12 - 1.17) 3.99E-34 1.13 (1.11 - 1.16) 2.65E-28 rs72725817 x Known new lead SNP 1.27 (1.18 - 1.36) 8.30E-12 1.22 (1.14 - 1.31) 2.07E-08 rs180753027 x Known new lead SNP 1.34 (1.25 - 1.43) 3.98E-17 1.27 (1.17 - 1.38) 3.08E-09 rs72725854 x x Known not replaced 2.38 (2.27 - 2.49) 2.09E- 296 2.02 (1.92 - 2.11) 3.10E- 185 rs77541621 x x Known not replaced 1.84 (1.79 - 1.90) 8.24e-338 1.69 (1.63 - 1.75) 1.94E- 174 128 rs72725879 x x Known not replaced 1.26 (1.25 - 1.27) 1.00e-359 1.24 (1.23 - 1.26) 3.26E- 295 rs183373024 x x Known not replaced 3.18 (3.01 - 3.37) 3.29e-359 2.63 (2.47 - 2.8) 2.22E- 204 rs78809737 x Known drop 1.24 (1.20 - 1.27) 1.81E-55 rs60681470 x Known new lead SNP 1.47 (1.44 - 1.49) 1.55e-429 1.16 (1.14 - 1.19) 1.05E-51 rs182395051 x Known new lead SNP 1.14 (1.09 - 1.20) 2.38E-08 1.19 (1.13 - 1.25) 4.40E-12 rs201057014 x Known drop 1.15 (1.14 - 1.17) 3.64E-84 rs17464492 x x Known not replaced 1.15 (1.14 - 1.16) 2.02E- 134 1.12 (1.11 - 1.14) 2.04E-95 rs6983267 x x Known not replaced 1.23 (1.22 - 1.25) 3.27e-387 1.19 (1.17 - 1.2) 2.85E- 250 rs10090154 x Known replaced 1.43 (1.41 - 1.45) 9.81e-551 rs11986220 x Known new lead SNP 1.46 (1.44 - 1.48) 9.10e-568 1.37 (1.35 - 1.39) 1E- 378.2077 25470649 rs34265760 x Known drop 1.14 (1.11 - 1.18) 4.66E-17 rs12549761 x x Known not replaced 1.24 (1.22 - 1.26) 1.41E- 170 1.18 (1.17 - 1.2) 3.18E- 101 rs4285449 x Known new lead SNP 1.04 (1.03 - 1.05) 9.73E-17 1.05 (1.04 - 1.06) 4.17E-25 rs72697078 x Novel - 1.06 (1.05 - 1.08) 3.62E-11 rs7045411 x Novel - 1.04 (1.02 - 1.05) 1.12E-08 rs34540271 x Known replaced 1.04 (1.03 - 1.05) 5.57E-12 rs12004058 x Known new lead SNP 1.04 (1.03 - 1.05) 3.45E-13 1.04 (1.03 - 1.05) 1.58E-12 rs10122990 x Known replaced 1.04 (1.03 - 1.06) 1.70E-18 rs2171954 x Known new lead SNP 1.05 (1.04 - 1.06) 1.79E-20 1.05 (1.04 - 1.06) 3.63E-20 rs17694493 x x Known not replaced 1.07 (1.06 - 1.09) 6.39E-21 rs10122495 x Known replaced 1.04 (1.03 - 1.05) 2.47E-13 rs307645 x Known new lead SNP 1.04 (1.03 - 1.05) 1.32E-15 rs142727307 x x Known not replaced 1.08 (1.06 - 1.09) 1.73E-24 rs4451364 x x Known not replaced 1.05 (1.04 - 1.06) 6.40E-17 1.05 (1.04 - 1.06) 4.03E-17 rs817872 x Known replaced 1.06 (1.05 - 1.07) 4.78E-31 rs12156453 x Known new lead SNP 1.06 (1.05 - 1.08) 3.38E-33 1.07 (1.06 - 1.08) 1.67E-37 rs143655302 x x Known not replaced 1.28 (1.22 - 1.33) 5.34E-27 1.31 (1.25 - 1.38) 2.89E-26 rs7034326 x Novel x - 1.05 (1.03 - 1.06) 2.11E-10 1.04 (1.03 - 1.06) 1.55E-08 129 rs201081105 x Novel - 1.15 (1.11 - 1.20) 3.55E-11 rs7861040 x Novel - 1.03 (1.02 - 1.04) 1.48E-10 rs2241167 x Known replaced 1.04 (1.03 - 1.05) 1.31E-12 rs6478788 x Known new lead SNP 1.04 (1.03 - 1.05) 1.29E-13 rs12634 x x Known not replaced 1.05 (1.04 - 1.06) 3.39E-19 rs12781100 x Known replaced 1.06 (1.04 - 1.08) 1.68E-14 rs34487581 x Known new lead SNP 1.06 (1.04 - 1.08) 4.25E-15 rs11259192 x Novel - 1.07 (1.05 - 1.09) 2.04E-17 1.08 (1.06 - 1.1) 2.85E-22 rs72772400 x Novel - 1.04 (1.03 - 1.05) 1.13E-09 1.05 (1.04 - 1.06) 9.50E-15 rs56137930 x Novel - 1.03 (1.02 - 1.05) 4.94E-09 rs7075427 x Known drop 1.12 (1.10 - 1.15) 4.96E-28 rs11599847 x Known drop 1.12 (1.09 - 1.15) 2.49E-14 rs10993994 x x Known not replaced 1.21 (1.20 - 1.22) 1.13e-325 rs11817544 x Known replaced 1.08 (1.06 - 1.10) 1.31E-19 rs59483558 x Known new lead SNP 1.08 (1.06 - 1.10) 1.04E-19 1.08 (1.06 - 1.1) 2.00E-17 rs12412705 x x Known not replaced 1.07 (1.06 - 1.09) 9.90E-21 1.07 (1.05 - 1.08) 3.91E-17 rs1935581 x Known replaced 1.04 (1.03 - 1.05) 5.85E-12 rs792207 x Known new lead SNP 1.03 (1.02 - 1.04) 3.68E-12 rs12262998 x Known replaced 1.07 (1.06 - 1.08) 7.06E-40 rs4522108 x Known new lead SNP 1.07 (1.06 - 1.08) 1.29E-41 rs11813268 x Novel - 1.05 (1.03 - 1.06) 1.12E-13 rs10885396 x Known replaced 1.04 (1.03 - 1.05) 1.80E-15 rs2104598 x Known new lead SNP 1.04 (1.03 - 1.05) 1.37E-16 rs528488 x Novel x - 1.04 (1.03 - 1.05) 1.60E-09 1.04 (1.03 - 1.05) 1.53E-09 rs677062 x Novel - 1.04 (1.03 - 1.05) 2.35E-09 1.04 (1.03 - 1.05) 3.49E-09 rs10886810 x Novel x - 1.05 (1.04 - 1.07) 1.21E-11 1.05 (1.03 - 1.06) 1.10E-08 rs4558107 x x Known not replaced 1.06 (1.05 - 1.07) 5.08E-30 1.06 (1.05 - 1.07) 3.23E-30 rs140783917 x Known replaced 1.18 (1.13 - 1.23) 4.56E-14 rs77612511 x Known new lead SNP 1.20 (1.15 - 1.25) 8.57E-17 1.19 (1.1 - 1.28) 8.57E-17 rs10788167 x Known replaced 1.05 (1.04 - 1.06) 8.32E-17 rs10886900 x Known new lead SNP 1.05 (1.04 - 1.06) 9.68E-17 1.05 (1.04 - 1.06) 9.07E-18 130 rs10749415 x x Known not replaced 1.12 (1.10 - 1.15) 9.26E-35 1.12 (1.1 - 1.15) 4.72E-34 rs34032268 x Novel x - 1.03 (1.02 - 1.04) 3.40E-09 1.03 (1.02 - 1.04) 4.02E-10 rs12769682 x Known replaced 1.05 (1.04 - 1.07) 1.89E-20 rs11371876 x Known new lead SNP 1.06 (1.05 - 1.07) 1.35E-21 rs1881502 x x Known not replaced 1.04 (1.03 - 1.05) 6.79E-09 1.04 (1.03 - 1.05) 5.33E-09 rs11043143 x Known replaced 1.15 (1.13 - 1.16) 4.27E- 118 rs7123299 x Known new lead SNP 1.15 (1.13 - 1.16) 2.22E- 118 1.15 (1.14 - 1.16) 5.24E- 122 rs57372985 x Novel - 1.05 (1.03 - 1.07) 1.17E-10 rs61890184 x x Known not replaced 1.07 (1.06 - 1.09) 1.96E-22 rs117833876 x Novel - 1.10 (1.07 - 1.14) 1.23E-09 rs11032076 x Novel - 1.05 (1.03 - 1.06) 1.83E-09 rs74897609 x Novel - 1.10 (1.06 - 1.13) 1.57E-08 rs68010938 x Known replaced 1.05 (1.04 - 1.06) 1.66E-14 rs10838703 x Known new lead SNP 1.04 (1.03 - 1.06) 2.42E-15 rs1048374 x Known replaced 1.11 (1.08 - 1.14) 1.41E-13 rs17153415 x Known new lead SNP 1.05 (1.04 - 1.06) 3.63E-16 1.05 (1.04 - 1.06) 2.52E-15 rs533676902 x Novel x - 1.43 (1.33 - 1.54) 2.09E-22 1.44 (1.34 - 1.56) 2.74E-22 rs2277283 x x Known not replaced 1.05 (1.04 - 1.06) 3.37E-20 rs11227678 x Novel x - 1.05 (1.03 - 1.07) 2.98E-10 1.05 (1.03 - 1.06) 1.21E-08 rs12785905 x Known replaced 1.09 (1.06 - 1.12) 9.17E-12 rs35153256 x Known new lead SNP 1.09 (1.06 - 1.11) 1.16E-13 1.08 (1.06 - 1.11) 3.22E-12 rs3018690 x x Known not replaced 1.10 (1.09 - 1.11) 1.85E-83 1.05 (1.04 - 1.06) 6.43E-18 rs11825796 x x Known not replaced 1.05 (1.04 - 1.06) 5.47E-17 1.09 (1.07 - 1.1) 6.95E-46 rs11228580 x x Known not replaced 1.26 (1.25 - 1.28) 2.95E- 275 1.27 (1.25 - 1.29) 4.22E- 233 rs3918298 x Known replaced 1.14 (1.11 - 1.16) 1.70E-28 rs36225067 x Known new lead SNP 1.18 (1.15 - 1.22) 8.52E-34 1.17 (1.14 - 1.2) 2.78E-28 rs56159348 x Known replaced 1.05 (1.04 - 1.06) 2.84E-18 rs7928373 x Known new lead SNP 1.05 (1.04 - 1.06) 3.83E-20 rs12796422 x Novel - 1.03 (1.02 - 1.04) 4.92E-09 rs11568818 x x Known not replaced 1.07 (1.06 - 1.08) 1.52E-38 1.07 (1.06 - 1.08) 7.93E-40 131 rs7945830 x Novel x - 1.03 (1.02 - 1.04) 4.96E-11 1.03 (1.02 - 1.04) 3.15E-11 rs10749918 x Novel x - 1.04 (1.03 - 1.05) 3.35E-14 1.03 (1.02 - 1.04) 4.47E-10 rs9666351 x Novel x - 1.05 (1.04 - 1.07) 3.88E-11 1.05 (1.03 - 1.07) 6.85E-10 rs74911261 x x Known not replaced 1.17 (1.13 - 1.21) 1.35E-18 1.14 (1.1 - 1.18) 3.48E-12 rs4936662 x Novel - 1.03 (1.02 - 1.04) 2.46E-08 rs5794883 x x Known not replaced 1.05 (1.04 - 1.07) 9.28E-21 rs138466039 x x Known not replaced 1.32 (1.25 - 1.39) 3.05E-23 rs79186742 x Novel - 1.06 (1.04 - 1.08) 2.94E-09 rs878987 x Known replaced 1.05 (1.04 - 1.06) 2.45E-12 rs28378882 x Known new lead SNP 1.05 (1.04 - 1.07) 3.44E-15 rs2066827 x x Known not replaced 1.04 (1.02 - 1.05) 1.49E-09 1.04 (1.03 - 1.06) 3.38E-12 rs77216612 x Known replaced 1.03 (1.02 - 1.05) 3.32E-09 rs7956514 x Known new lead SNP 1.03 (1.02 - 1.05) 2.05E-09 1.04 (1.03 - 1.05) 2.56E-11 rs10845938 x Known replaced 1.05 (1.04 - 1.06) 8.60E-26 rs11055891 x Known new lead SNP 1.07 (1.06 - 1.09) 2.29E-26 1.07 (1.05 - 1.08) 5.01E-23 rs4764122 x Novel x - 1.04 (1.03 - 1.06) 2.51E-14 1.03 (1.02 - 1.05) 6.55E-09 rs7307567 x Novel x - 1.05 (1.04 - 1.07) 1.29E-13 1.04 (1.03 - 1.06) 2.35E-08 rs80130819 x x Known not replaced 1.08 (1.06 - 1.11) 5.66E-17 rs56222401 x x Known not replaced 1.08 (1.07 - 1.10) 1.25E-52 rs113925811 x Known replaced 1.18 (1.16 - 1.20) 1.16E- 114 rs73110464 x Known new lead SNP 1.18 (1.17 - 1.20) 7.73E- 115 1.18 (1.16 - 1.2) 7.62E-82 rs187809440 x Known replaced 1.47 (1.35 - 1.59) 6.69E-20 rs142025399 x Known new lead SNP 1.65 (1.51 - 1.81) 2.39E-27 1.60 (1.36 - 1.9) 3.31E-08 rs7968403 x x Known not replaced 1.06 (1.05 - 1.07) 9.55E-26 rs975917 x Novel - 1.05 (1.03 - 1.06) 2.34E-09 rs4842687 x Known replaced 1.06 (1.05 - 1.08) 8.20E-31 rs2037376 x Known new lead SNP 1.06 (1.05 - 1.07) 9.49E-32 rs77121786 x Known replaced 1.04 (1.03 - 1.06) 1.34E-12 rs76825149 x Known new lead SNP 1.04 (1.03 - 1.05) 2.74E-13 rs9593 x Novel - 1.03 (1.02 - 1.04) 3.54E-10 132 rs1270884 x Known replaced 1.08 (1.06 - 1.09) 3.62E-48 rs2555009 x Known new lead SNP 1.08 (1.07 - 1.09) 1.88E-50 1.08 (1.07 - 1.09) 5.65E-51 rs11067228 x Novel x - 1.03 (1.02 - 1.04) 1.35E-11 1.03 (1.02 - 1.04) 9.93E-12 rs7295014 x x Known not replaced 1.06 (1.05 - 1.07) 7.57E-27 rs275960 x Novel - 1.04 (1.02 - 1.05) 1.26E-08 rs1327653 x Known replaced 1.04 (1.03 - 1.05) 4.46E-13 rs35679553 x Known new lead SNP 1.04 (1.03 - 1.05) 6.27E-11 1.03 (1.02 - 1.04) 3.12E-08 rs12583030 x Novel x - 1.06 (1.05 - 1.08) 1.42E-17 1.06 (1.04 - 1.07) 9.61E-15 rs9563009 x Novel x - 1.05 (1.03 - 1.06) 4.23E-10 1.05 (1.03 - 1.06) 2.87E-10 rs7489409 x x Known not replaced 1.09 (1.07 - 1.10) 1.01E-44 1.09 (1.07 - 1.1) 3.39E-42 rs7336001 x x Known not replaced 1.12 (1.10 - 1.14) 3.54E-28 1.12 (1.1 - 1.15) 3.72E-29 rs7995740 x Novel - 1.03 (1.02 - 1.04) 1.34E-08 rs9523312 x Novel - 1.03 (1.02 - 1.04) 5.56E-09 rs9583309 x Novel - 1.03 (1.02 - 1.04) 2.49E-11 1.03 (1.02 - 1.04) 4.84E-11 rs12873308 x Novel x - 1.04 (1.03 - 1.05) 2.61E-13 1.04 (1.03 - 1.05) 4.41E-13 rs75823044 x x Known not replaced 1.39 (1.29 - 1.51) 9.79E-16 1.37 (1.15 - 1.63) 9.79E-16 rs1004030 x x Known not replaced 1.04 (1.03 - 1.05) 2.35E-12 1.04 (1.03 - 1.05) 1.08E-12 rs11333310 x Novel x - 1.03 (1.02 - 1.04) 1.14E-08 1.03 (1.02 - 1.04) 3.02E-08 rs117531891 x Novel - 1.18 (1.13 - 1.25) 3.24E-11 rs6571758 x Known replaced 1.05 (1.04 - 1.06) 8.41E-20 rs12885529 x Known new lead SNP 1.05 (1.04 - 1.06) 1.88E-20 1.05 (1.03 - 1.06) 1.53E-18 rs145024784 x Novel x - 1.14 (1.09 - 1.18) 2.33E-09 1.13 (1.08 - 1.18) 2.26E-08 rs146403404 x Novel x - 1.21 (1.15 - 1.28) 3.25E-11 1.21 (1.14 - 1.28) 8.59E-10 rs11849126 x Known drop 1.03 (1.02 - 1.04) 3.60E-07 rs4901313 x Known replaced 1.08 (1.06 - 1.09) 2.29E-31 rs62003517 x Known new lead SNP 1.09 (1.07 - 1.10) 2.81E-32 rs8005621 x Known replaced 1.07 (1.06 - 1.09) 1.10E-23 rs2351179 x Known new lead SNP 1.06 (1.05 - 1.07) 3.36E-28 1.06 (1.05 - 1.07) 1.19E-27 rs11158335 x Novel x - 1.04 (1.03 - 1.05) 3.09E-12 1.04 (1.03 - 1.05) 8.87E-14 rs114053368 x Known not replaced 1.04 (1.02 - 1.06) 0.000114 8 133 rs79133931 x x Known not replaced 1.08 (1.03 - 1.12) 0.00109 rs2093202 x Known replaced 1.03 (1.02 - 1.04) 3.20E-11 rs1023529 x Known new lead SNP 1.04 (1.03 - 1.05) 2.07E-12 1.04 (1.03 - 1.05) 4.44E-14 rs767127 x x Known not replaced 1.05 (1.04 - 1.06) 2.18E-19 1.05 (1.04 - 1.06) 4.74E-21 rs17565772 x Known replaced 1.04 (1.03 - 1.05) 2.32E-12 rs1044527 x Known new lead SNP 1.04 (1.03 - 1.05) 5.15E-13 rs3213716 x Novel - 1.03 (1.02 - 1.05) 2.63E-11 rs17802308 x Novel - 1.04 (1.03 - 1.06) 4.12E-09 rs12586382 x Novel - 1.04 (1.03 - 1.06) 4.00E-08 rs28929474 x Novel - 1.14 (1.10 - 1.19) 1.38E-10 1.14 (1.09 - 1.19) 6.95E-10 rs72695000 x Novel - 1.04 (1.03 - 1.06) 6.35E-09 1.04 (1.02 - 1.06) 6.35E-09 rs201750321 x Novel - 1.08 (1.05 - 1.11) 4.03E-08 rs8004408 x Novel - 1.03 (1.02 - 1.04) 4.39E-10 rs4983387 x Novel - 1.04 (1.03 - 1.06) 3.22E-10 rs757438 x Novel - 1.04 (1.03 - 1.05) 1.00E-12 rs922511 x Novel - 1.04 (1.03 - 1.05) 4.38E-14 rs11561564 x Known replaced 1.06 (1.04 - 1.07) 5.21E-17 rs8041443 x Known new lead SNP 1.06 (1.05 - 1.07) 1.64E-18 rs12912118 x Novel x - 1.03 (1.02 - 1.04) 7.40E-09 rs33984059 x x Known not replaced 1.16 (1.11 - 1.21) 1.89E-13 rs6151589 x Novel - 1.03 (1.02 - 1.04) 2.41E-10 rs74634457 x Known replaced 1.05 (1.04 - 1.07) 5.45E-20 rs112914414 x Known new lead SNP 1.06 (1.04 - 1.07) 2.78E-20 1.05 (1.04 - 1.07) 1.29E-19 rs8023793 x x Known not replaced 1.03 (1.02 - 1.04) 2.10E-12 1.03 (1.02 - 1.04) 3.33E-10 rs12913603 x Known replaced 1.04 (1.03 - 1.05) 3.88E-12 rs34627158 x Known new lead SNP 1.04 (1.03 - 1.05) 6.85E-14 rs79548680 x Novel - 1.05 (1.04 - 1.06) 3.51E-13 rs28680698 x Novel - 1.03 (1.02 - 1.04) 6.27E-10 rs2412005 x Novel - 1.03 (1.02 - 1.04) 2.11E-08 rs75854896 x Novel - 1.04 (1.02 - 1.05) 2.25E-08 rs111832163 x Novel - 1.04 (1.03 - 1.05) 1.80E-13 1.04 (1.03 - 1.05) 1.91E-13 134 rs35193388 x Novel - 1.03 (1.02 - 1.05) 1.26E-08 1.03 (1.02 - 1.05) 2.56E-08 rs760118 x Novel - 1.04 (1.03 - 1.05) 5.82E-16 1.04 (1.03 - 1.05) 1.17E-15 rs3751723 x Novel x - 1.03 (1.02 - 1.04) 1.94E-09 1.03 (1.02 - 1.04) 1.19E-08 rs7188897 x Known replaced 1.04 (1.03 - 1.05) 1.22E-12 rs1420285 x Known new lead SNP 1.04 (1.03 - 1.05) 1.18E-12 1.04 (1.03 - 1.05) 5.90E-13 rs13380763 x Known replaced 1.05 (1.04 - 1.07) 1.47E-15 rs3893264 x Known new lead SNP 1.05 (1.04 - 1.07) 1.61E-16 1.06 (1.04 - 1.07) 7.68E-17 rs11863709 x x Known not replaced 1.11 (1.08 - 1.14) 5.21E-13 1.11 (1.08 - 1.14) 3.58E-13 rs11866053 x Novel x - 1.05 (1.04 - 1.06) 2.75E-15 1.04 (1.03 - 1.05) 2.69E-11 rs8045257 x Novel x - 1.04 (1.03 - 1.05) 1.15E-12 1.03 (1.02 - 1.04) 6.41E-09 rs9930024 x Novel - 1.03 (1.02 - 1.04) 5.26E-10 rs1858800 x Novel - 1.03 (1.02 - 1.04) 2.56E-09 rs28709974 x x Known not replaced 1.10 (1.08 - 1.12) 1.91E-19 1.09 (1.07 - 1.12) 2.64E-17 rs12445706 x Novel x - 1.07 (1.05 - 1.09) 1.24E-10 1.06 (1.04 - 1.09) 1.94E-09 rs8052913 x Known replaced 1.04 (1.03 - 1.05) 2.42E-14 rs7405197 x Known new lead SNP 1.04 (1.03 - 1.05) 4.46E-15 rs72801582 x Novel - 1.05 (1.04 - 1.06) 1.14E-12 rs684232 x x Known not replaced 1.08 (1.07 - 1.09) 1.77E-52 rs5030755 x Novel - 1.05 (1.03 - 1.06) 2.26E-08 rs78378222 x x Known not replaced 1.25 (1.19 - 1.31) 8.19E-21 1.25 (1.19 - 1.31) 8.15E-20 rs78044421 x Novel x - 1.13 (1.09 - 1.17) 1.39E-10 1.12 (1.07 - 1.16) 1.87E-08 rs9895704 x Known not replaced 1.13 (1.08 - 1.18) 9.36E-09 1.12 (1.04 - 1.21) 9.36E-09 rs28441558 x Known replaced 1.16 (1.13 - 1.19) 1.04E-39 rs60424695 x Known new lead SNP 1.16 (1.13 - 1.19) 5.04E-40 1.16 (1.13 - 1.18) 1.25E-38 rs72811270 x Known drop 1.04 (1.02 - 1.05) 9.38E-07 rs5819638 x Novel - 1.04 (1.03 - 1.05) 2.14E-11 rs73991216 x Known not replaced 1.19 (1.14 - 1.24) 2.42E-14 1.19 (1.13 - 1.25) 5.08E-12 rs4795646 x Known replaced 1.06 (1.04 - 1.07) 3.22E-19 rs34441237 x Known new lead SNP 1.06 (1.04 - 1.07) 2.72E-19 1.06 (1.03 - 1.08) 2.72E-19 rs3110641 x x Known not replaced 1.07 (1.05 - 1.08) 4.66E-31 1.06 (1.04 - 1.07) 2.38E-22 rs11649743 x x Known not replaced 1.13 (1.11 - 1.14) 4.39E-78 1.10 (1.08 - 1.11) 2.56E-46 135 rs11263763 x x Known not replaced 1.22 (1.21 - 1.24) 1.18e-367 1.21 (1.2 - 1.22) 1E- 326.3079 57499221 rs67633759 x Novel - 1.03 (1.02 - 1.04) 1.72E-08 rs138213197 x x Known not replaced 4.32 (3.89 - 4.80) 1.53E- 165 4.23 (3.81 - 4.7) 1.41E- 160 rs2960158 x Known replaced 1.04 (1.03 - 1.05) 8.02E-11 rs7212344 x Known new lead SNP 1.05 (1.04 - 1.06) 8.80E-16 1.03 (1.02 - 1.05) 1.04E-08 rs565189650 x Known replaced 1.13 (1.10 - 1.15) 7.32E-31 rs73324308 x Known new lead SNP 1.13 (1.11 - 1.15) 9.26E-42 1.12 (1.1 - 1.14) 1.82E-32 rs12938538 x Known replaced 1.03 (1.02 - 1.04) 6.16E-10 rs2526370 x Known new lead SNP 1.03 (1.02 - 1.04) 8.81E-11 rs72843168 x Novel - 1.04 (1.02 - 1.05) 1.31E-08 rs199554208 x Novel - 1.04 (1.02 - 1.05) 1.18E-08 rs148511027 x Known replaced 1.14 (1.13 - 1.15) 1.03E- 147 rs984434 x Known new lead SNP 1.14 (1.13 - 1.15) 6.11E- 153 rs12607091 x Novel - 1.03 (1.02 - 1.04) 2.39E-08 rs140293195 x Novel - 1.04 (1.03 - 1.05) 4.55E-13 1.03 (1.02 - 1.05) 1.38E-11 rs11660291 x Novel - 1.03 (1.02 - 1.04) 2.92E-08 1.02 (1.01 - 1.03) 1.26E-06 rs8089411 x Known replaced 1.03 (1.02 - 1.04) 8.60E-10 rs12955457 x Known new lead SNP 1.04 (1.03 - 1.05) 1.21E-12 rs35283980 x Known replaced 1.03 (1.02 - 1.04) 1.01E-07 rs12956892 x Known new lead SNP 1.03 (1.02 - 1.04) 1.99E-08 rs11381388 x Known replaced 1.04 (1.03 - 1.05) 7.61E-10 rs62098660 x Known new lead SNP 1.04 (1.03 - 1.06) 5.03E-17 rs11876000 x Known replaced 1.02 (1.01 - 1.03) 1.89E-06 rs6566073 x Known new lead SNP 1.03 (1.02 - 1.04) 3.14E-08 rs11663079 x Novel - 1.03 (1.02 - 1.04) 1.08E-10 rs9959454 x Known replaced 1.06 (1.05 - 1.08) 1.27E-31 rs7236307 x Known new lead SNP 1.06 (1.05 - 1.07) 1.20E-32 1.06 (1.05 - 1.07) 4.80E-26 rs8096321 x Novel x - 1.06 (1.05 - 1.07) 3.88E-21 1.05 (1.03 - 1.06) 2.25E-13 rs10412482 x Known replaced 1.05 (1.04 - 1.07) 2.38E-22 136 rs10421727 x Known new lead SNP 1.05 (1.04 - 1.06) 1.59E-23 rs7245983 x Novel - 1.03 (1.02 - 1.04) 7.35E-09 rs17501397 x x Known not replaced 1.06 (1.04 - 1.08) 1.44E-09 rs353411 x Novel x - 1.05 (1.04 - 1.06) 9.39E-17 1.04 (1.03 - 1.05) 2.44E-13 rs59710626 x Known drop 1.05 (1.03 - 1.06) 1.88E-09 rs4802297 x Known replaced 1.10 (1.09 - 1.11) 6.46E-79 rs12976534 x Known new lead SNP 1.10 (1.09 - 1.11) 4.99E-79 1.09 (1.08 - 1.1) 3.93E-77 rs11673591 x x Known not replaced 1.07 (1.06 - 1.08) 2.42E-35 rs67546213 x Novel - 1.05 (1.03 - 1.06) 2.54E-11 rs2659051 x Known replaced 1.14 (1.13 - 1.16) 4.21E- 106 rs11665748 x Known new lead SNP 1.13 (1.12 - 1.14) 3.51E- 124 1.09 (1.08 - 1.1) 2.74E-52 rs61752561 x x Known not replaced 1.11 (1.08 - 1.14) 2.49E-14 1.16 (1.13 - 1.2) 3.60E-25 rs76765083 x Known replaced 1.36 (1.33 - 1.39) 1.57E- 178 rs62113212 x Known new lead SNP 1.36 (1.33 - 1.39) 8.65E- 179 1.29 (1.27 - 1.32) 5.10E- 112 rs6039055 x Known drop 1.02 (1.01 - 1.03) 3.93E-05 rs11698745 x Novel - 1.04 (1.03 - 1.05) 7.12E-11 rs6058431 x Novel - 1.03 (1.02 - 1.04) 2.80E-09 rs11480453 x Known replaced 1.04 (1.02 - 1.05) 7.25E-10 rs2424905 x Known new lead SNP 1.03 (1.02 - 1.05) 6.36E-11 rs34633531 x Novel - 1.05 (1.03 - 1.07) 2.14E-10 rs6141551 x Known replaced 1.03 (1.02 - 1.05) 7.69E-12 rs143383 x Known new lead SNP 1.04 (1.03 - 1.05) 1.08E-12 rs73909841 x Known replaced 1.10 (1.08 - 1.12) 9.64E-26 rs76685994 x Known new lead SNP 1.10 (1.08 - 1.12) 5.66E-26 rs61744628 x Novel x - 1.15 (1.11 - 1.18) 2.64E-16 1.13 (1.1 - 1.17) 2.11E-13 rs6126986 x Known replaced 1.06 (1.05 - 1.07) 6.31E-28 rs4809941 x Known new lead SNP 1.06 (1.05 - 1.07) 1.96E-28 1.06 (1.05 - 1.07) 2.16E-27 rs1276400 x Novel - 1.03 (1.02 - 1.04) 2.31E-09 rs11467114 x Known replaced 1.04 (1.03 - 1.05) 2.09E-12 137 rs35418657 x Known new lead SNP 1.04 (1.03 - 1.05) 1.43E-13 rs381331 x x Known not replaced 1.06 (1.05 - 1.07) 2.26E-26 1.04 (1.03 - 1.05) 5.34E-13 rs74179816 x Novel x - 1.30 (1.23 - 1.36) 2.43E-26 1.18 (1.12 - 1.25) 2.25E-10 rs3787099 x Known replaced 1.11 (1.09 - 1.13) 1.06E-25 rs77552606 x Known new lead SNP 1.11 (1.09 - 1.13) 7.29E-26 1.13 (1.11 - 1.15) 2.15E-33 rs1056990 x Novel x - 1.09 (1.07 - 1.12) 6.29E-20 1.06 (1.04 - 1.08) 2.01E-09 rs1058319 x x Known not replaced 1.11 (1.09 - 1.13) 5.23E-47 1.10 (1.08 - 1.11) 1.61E-32 rs150947563 x Known not replaced 1.47 (1.31 - 1.66) 3.02E-10 1.47 (1.29 - 1.68) 1.47E-08 rs2824039 x Novel - 1.03 (1.02 - 1.04) 3.03E-08 rs10154043 x Novel - 1.03 (1.02 - 1.04) 8.96E-09 1.03 (1.02 - 1.04) 6.78E-09 rs7275340 x Novel - 1.03 (1.02 - 1.04) 3.01E-08 1.03 (1.02 - 1.04) 1.48E-08 rs2242798 x Novel - 1.04 (1.03 - 1.05) 8.42E-12 1.04 (1.03 - 1.05) 3.65E-11 rs116988298 x Novel - 1.10 (1.07 - 1.13) 3.73E-12 1.10 (1.07 - 1.13) 2.04E-12 rs11701433 x Known replaced 1.03 (1.02 - 1.04) 6.78E-10 rs2836750 x Known new lead SNP 1.04 (1.03 - 1.05) 1.02E-13 rs61735792 x x Known not replaced 1.23 (1.18 - 1.29) 1.26E-20 1.22 (1.16 - 1.28) 1.07E-16 rs9978557 x Known replaced 1.10 (1.08 - 1.12) 2.77E-22 rs9984523 x Known new lead SNP 1.09 (1.07 - 1.11) 7.51E-23 1.09 (1.07 - 1.11) 4.34E-22 rs1978060 x x Known not replaced 1.05 (1.04 - 1.06) 1.27E-23 rs9625483 x Known replaced 1.11 (1.07 - 1.15) 3.38E-09 rs12165913 x Known new lead SNP 1.06 (1.04 - 1.08) 9.94E-10 1.07 (1.04 - 1.09) 2.51E-08 rs17886163 x x Known not replaced 1.58 (1.41 - 1.78) 1.13E-14 1.53 (1.24 - 1.88) 1.13E-14 rs555607708 x x Known not replaced 1.64 (1.45 - 1.85) 1.17E-15 1.63 (1.44 - 1.84) 2.59E-15 rs193478 x Novel - 1.04 (1.02 - 1.05) 3.73E-11 1.03 (1.02 - 1.05) 1.27E-09 rs138708 x x Known not replaced 1.15 (1.11 - 1.18) 3.95E-20 rs34584683 x x Known not replaced 1.06 (1.05 - 1.07) 3.58E-20 rs6003062 x x Known not replaced 1.24 (1.22 - 1.27) 1.18E-88 1.17 (1.15 - 1.2) 2.40E-45 rs5759167 x x Known not replaced 1.13 (1.12 - 1.14) 9.71E- 147 1.11 (1.1 - 1.12) 1.15E-97 rs9615099 x Known replaced 1.03 (1.02 - 1.04) 3.71E-08 rs6007594 x Known new lead SNP 1.03 (1.02 - 1.05) 1.33E-10 138 rs960417 x x Known not replaced 1.05 (1.04 - 1.06) 8.42E-29 1.10 (1.08 - 1.12) 5.10E-29 rs17321482 x x Known not replaced 1.07 (1.06 - 1.08) 9.13E-23 1.14 (1.11 - 1.17) 7.74E-21 rs61018477 x Novel x - 1.04 (1.03 - 1.05) 1.25E-09 1.10 (1.06 - 1.13) 6.68E-10 rs5924562 x Novel - 1.03 (1.02 - 1.03) 1.27E-11 rs5972255 x x Known not replaced 1.04 (1.03 - 1.05) 2.53E-20 rs4824810 x Novel - 1.04 (1.03 - 1.05) 7.98E-12 rs11338635 x x Known not replaced 1.10 (1.09 - 1.11) 5.62E- 133 1.21 (1.19 - 1.23) 3.96E- 106 rs5943724 x Known drop 1.04 (1.03 - 1.05) 3.13E-19 rs4826594 x Known drop 1.00 (0.99 - 1.01) 0.9864 rs5919393 x Known replaced 1.05 (1.04 - 1.07) 1.11E-22 rs7888856 x Known new lead SNP 1.05 (1.04 - 1.07) 8.33E-23 1.10 (1.08 - 1.13) 3.98E-20 rs5965467 x Novel x - 1.09 (1.06 - 1.11) 2.45E-13 1.14 (1.09 - 1.2) 5.90E-08 rs371707439 x Known replaced 1.06 (1.05 - 1.07) 3.01E-25 rs13441059 x Known new lead SNP 1.04 (1.04 - 1.05) 3.41E-26 1.09 (1.07 - 1.12) 1.10E-12 rs747181 x Novel x - 1.03 (1.02 - 1.04) 8.70E-14 1.06 (1.05 - 1.08) 6.81E-14 rs5931727 x Novel - 1.03 (1.02 - 1.04) 1.33E-09 139 Supplementary Table 3. The net reclassification improvement (NRI) of GRS models in GWAS discovery samples GRS100 vs. age only* GRS181 vs. age only GRS269 vs. age only GRS451 vs. age only GRS451 vs. GRS269 Individu als Reclassi fied (%Up / %Down ) NRI (95% CI) Individu als Reclassi fied (%Up / %Down ) NRI (95% CI) Individu als Reclassi fied (%Up / %Down ) NRI (95% CI) Individu als Reclassi fied (%Up / %Down ) NRI (95% CI) Individu als Reclassi fied (%Up / %Down ) NRI (95% CI) P value ** Euro pean Over all -- 49.5% (48.5% - 50.6%) -- 61.9% (60.9% - 62.9%) -- 64.2% (62.9% - 65.6%) -- 69.9% (68.7% - 71.2%) -- 17.8% (15.7% - 19.8%) 2.02 E-65 Case s 62.2% / 37.8% 24.4% (23.8% - 25.0%) 64.9% / 35.1% 29.7% (29.1% - 30.3%) 65.8% / 34.2% 31.6% (30.6% - 32.5%) 67.3% / 32.7% 34.6% (33.7% - 35.4%) 54.1% / 45.9% 8.3% (7.2% - 9.4%) Cont rols 37.5% / 62.5% 24.9% (24.3% - 25.5%) 34.1% / 65.9% 31.9% (31.3% - 32.5%) 33.7% / 66.3% 32.6% (32.0% - 33.2%) 32.4% / 67.6% 35.2% (34.6% - 35.8%) 45.3% / 54.7% 9.5% (8.4% - 10.6%) Afric an Over all -- 30.2% (27.5% - 32.9%) -- 45.4% (42.7% - 48.0%) -- 55.8% (53.2% - 58.5%) -- 58.5% (55.9% - 61.1%) -- 13.1% (8.3% - 17.9%) 9.07 E-08 Case s 57.3% / 42.7% 14.6% (13.0% - 16.2%) 60.2% / 39.8% 20.4% (18.9% - 22.0%) 63.3% / 36.7% 26.6% (25.0% - 28.1%) 64.1% / 35.9% 28.3% (26.8% - 29.8%) 53.3% / 46.7% 6.6% (4.0% - 9.1%) Cont rols 42.2% / 57.8% 15.6% (14.0% - 17.2%) 37.5% / 62.5% 24.9% (23.4% - 26.4%) 35.4% / 64.6% 29.3% (27.7% - 30.8%) 34.9% / 65.1% 30.2% (28.7% - 31.8%) 46.7% / 53.3% 6.6% (4.0% - 9.1%) Asian Over all -- 44.8% (37.8% - 51.8%) -- 46.1% (39.2% - 52.9%) -- 59.5% (53.0% - 66.1%) -- 60.4% (53.9% - 66.9%) -- 3.3% (-9.4% - 16.1%) 6.07 E-01 Case s 60.2% / 39.8% 20.5% (16.3% - 24.6%) 62.1% / 37.9% 24.2% (20.1% - 28.3%) 66.3% / 33.7% 32.5% (28.7% - 36.3%) 65.9% / 34.1% 31.9% (28.0% - 35.7%) 50.8% / 49.2% 1.6% (-4.7% - 7.9%) Cont rols 37.8% / 62.2% 24.3% (20.4% - 28.3%) 39.1% / 60.9% 21.8% (17.9% - 25.8%) 36.4% / 63.6% 27.1% (23.2% - 31.0%) 35.7% / 64.3% 28.7% (24.9% - 32.4%) 49.1% / 50.9% 1.8% (-5.3% - 8.9%) Hispa nic Over all -- 40.3% (34.5% - 46.1%) -- 49.5% (43.3% - 55.7%) -- 58.7% (52.6% - 64.8%) -- 61.9% (55.8% - 68.0%) -- 21.7% (10.7% - 32.6%) 1.03 E-04 Case s 59.3% / 40.7% 18.6% (15.1% - 22.0%) 61.2% / 38.8% 22.4% (18.7% - 26.1%) 64.5% / 35.5% 28.9% (25.4% - 32.4%) 65.0% / 35.0% 30.1% (26.5% - 33.6%) 54.9% / 45.1% 9.8% (4.0% - 15.5%) Cont rols 39.2% / 60.8% 21.7% (18.2% - 25.1%) 36.5% / 63.5% 27.1% (23.5% - 30.6%) 35.1% / 64.9% 29.8% (26.2% - 33.4%) 34.1% / 65.9% 31.8% (28.3% - 35.4%) 44.0% / 56.0% 11.9% (6.0% - 17.8%) * Age-only models included age, sub-study, and top 10 principal components, and GRS models included GRS, age, sub-study, and top 10 principal components; ** Test for the statistical significance of the NRI of model with GRS451 relative to model with GRS269. 140 Supplementary Table 4. The association of GRS and prostate cancer risk in GWAS discovery and replication samples GWAS Discovery Samples GRS GRS Catego ry European African Asian Hispanic Cont rols Ca ses O R (95% CI) P value Cont rols Ca ses O R (95% CI) P value Cont rols Ca ses O R (95% CI) P valu e Cont rols Ca ses O R (95% CI) P value GRS 269 [0- 10%] 4535 3 17 76 0. 27 (0.25- 0.28) 0.00E +00 6417 50 3 0. 34 (0.31- 0.38) 2.23E -88 181 44 0. 40 (0.27- 0.58) 1.28 E-06 2272 86 0. 34 (0.27- 0.45) 8.25E -16 (10- 20%] 4535 1 29 51 0. 44 (0.42- 0.47) 1.74E- 233 6413 78 3 0. 53 (0.48- 0.58) 2.83E -42 180 58 0. 52 (0.37- 0.73) 1.78 E-04 2270 13 3 0. 53 (0.43- 0.67) 3.26E -08 (20- 30%] 4535 0 39 75 0. 59 (0.57- 0.62) 2.83E- 114 6414 94 6 0. 65 (0.59- 0.7) 3.98E -23 181 84 0. 78 (0.57- 1.06) 1.09 E-01 2269 16 6 0. 66 (0.54- 0.82) 1.03E -04 (30- 40%] 4535 2 50 01 0. 76 (0.73- 0.79) 5.27E- 37 6413 11 00 0. 73 (0.67- 0.8) 3.81E -13 179 98 0. 87 (0.64- 1.17) 3.58 E-01 2270 21 0 0. 86 (0.71- 1.04) 1.19E -01 (40- 60%] 9070 1 13 33 5 1. 00 ref - 1282 7 29 32 1. 00 ref - 361 22 5 1. 00 ref - 4539 48 3 1. 00 ref - (60- 70%] 4535 1 90 18 1. 35 (1.3- 1.4) 1.17E- 55 6413 19 17 1. 30 (1.21- 1.4) 1.29E -12 179 16 1 1. 45 (1.1- 1.9) 8.00 E-03 2269 33 9 1. 33 (1.12- 1.59) 1.33E -03 (70- 80%] 4535 1 11 17 8 1. 68 (1.63- 1.75) 5.55E- 180 6414 22 70 1. 53 (1.42- 1.64) 5.62E -32 181 19 5 1. 74 (1.34- 2.27) 3.98 E-05 2270 44 5 1. 83 (1.56- 2.15) 2.32E -13 (80- 90%] 4535 0 15 49 2 2. 32 (2.25- 2.4) <2.23 E-308 6413 29 68 1. 98 (1.85- 2.12) 2.15E -87 180 28 7 2. 61 (2.02- 3.36) 1.26 E-13 2269 54 0 2. 13 (1.81- 2.49) 1.35E -20 (90- 100%] 4535 2 29 39 2 4. 36 (4.22- 4.5) <2.23 E-308 6417 46 73 3. 12 (2.93- 3.33) 3.84E -270 181 50 0 4. 56 (3.59- 5.81) 4.90 E-35 2271 90 6 3. 62 (3.13- 4.19) 2.65E -66 (99- 100%] 4538 60 86 8. 62 (8.1- 9.16) <2.23 E-308 644 78 6 5. 24 (4.61- 5.96) 2.82E -139 19 11 2 9. 43 (5.6- 15.88) 3.22 E-17 228 16 4 6. 83 (5.19- 8.98) 4.68E -43 Per SD GRS 4535 11 92 11 8 2. 17 (2.15- 2.19) <2.23 E-308 6414 1 18 09 2 1. 86 (1.83- 1.9) 1.38E -99 1803 16 52 2. 11 (1.95- 2.27) 4.38 E-82 2269 9 33 08 1. 96 (1.87- 2.05) 1.50E -184 GRS 451 [0- 10%] 4535 3 15 43 0. 24 (0.22- 0.25) 0.00E +00 6417 40 3 0. 26 (0.23- 0.29) 5.39E -113 181 33 0. 29 (0.19- 0.44) 5.70 E-09 2272 55 0. 22 (0.16- 0.3) 3.74E -22 (10- 20%] 4535 1 25 81 0. 40 (0.38- 0.42) 8.36E- 275 6413 65 6 0. 43 (0.39- 0.48) 6.07E -62 180 67 0. 58 (0.42- 0.81) 1.22 E-03 2270 11 5 0. 46 (0.36- 0.58) 5.81E -11 (20- 30%] 4535 0 34 86 0. 54 (0.52- 0.57) 1.22E- 144 6414 89 2 0. 61 (0.56- 0.67) 2.60E -27 181 83 0. 71 (0.52- 0.97) 2.90 E-02 2269 17 1 0. 68 (0.55- 0.84) 2.84E -04 141 (30- 40%] 4535 2 45 99 0. 71 (0.68- 0.74) 7.35E- 54 6413 10 77 0. 73 (0.67- 0.8) 5.64E -13 179 87 0. 77 (0.56- 1.05) 9.74 E-02 2270 20 8 0. 83 (0.68- 1) 5.54E -02 (40- 60%] 9070 1 12 84 5 1. 00 ref - 1282 7 28 69 1. 00 ref - 361 23 0 1. 00 ref - 4539 49 3 1. 00 ref - (60- 70%] 4535 1 89 42 1. 39 (1.34- 1.44) 6.97E- 67 6413 17 92 1. 26 (1.17- 1.36) 1.57E -09 179 15 5 1. 37 (1.04- 1.8) 2.49 E-02 2269 31 2 1. 29 (1.09- 1.54) 3.64E -03 (70- 80%] 4535 1 11 23 7 1. 76 (1.69- 1.82) 2.45E- 207 6414 23 74 1. 66 (1.55- 1.78) 5.46E -45 181 21 9 1. 92 (1.48- 2.49) 1.01 E-06 2270 44 2 1. 79 (1.52- 2.11) 3.59E -12 (80- 90%] 4535 0 15 72 8 2. 45 (2.37- 2.53) <2.23 E-308 6413 29 77 2. 12 (1.99- 2.27) 2.57E -105 180 28 3 2. 50 (1.94- 3.22) 1.23 E-12 2269 55 1 2. 31 (1.97- 2.7) 6.70E -26 (90- 100%] 4535 2 31 15 7 4. 90 (4.74- 5.05) <2.23 E-308 6417 50 52 3. 59 (3.37- 3.82) 0.00E +00 181 49 5 4. 40 (3.46- 5.59) 1.60 E-33 2271 96 1 3. 81 (3.3- 4.41) 1.28E -72 (99- 100%] 4538 66 75 9. 99 (9.41- 10.61) <2.23 E-308 644 90 5 6. 39 (5.62- 7.25) 1.52E -180 19 11 1 9. 28 (5.53- 15.59) 3.73 E-17 228 18 7 7. 10 (5.4- 9.35) 1.98E -44 Per SD GRS 4535 11 92 11 8 2. 34 (2.31- 2.36) <2.23 E-308 6414 1 18 09 2 2. 06 (2.02- 2.11) 0.00E +00 1803 16 52 2. 14 (1.98- 2.31) 7.58 E-85 2269 9 33 08 2. 11 (2.01- 2.21) 4.94E -217 Replication Samples GRS GRS Categor y MGB (European) MGI (European) EstBB (European) Contro ls Case s OR (95%CI) P value Contro ls Case s OR (95%CI) P value Contro ls Case s OR (95%CI) P value GRS 269 [0-10%] 1193 55 0.34 (0.25- 0.46) - 1054 100 0.3 8 (0.3-0.48) 8.53E- 17 2855 72 0.3 2 (0.25- 0.42) <1E-8 (10- 20%] 1168 79 0.51 (0.39- 0.67) - 1054 151 0.5 7 (0.47- 0.69) 1.96E- 08 2856 92 0.4 2 (0.33- 0.53) <1E-8 (20- 30%] 1149 98 0.66 (0.52- 0.85) - 1053 167 0.6 3 (0.52- 0.76) 2.07E- 06 2854 142 0.6 4 (0.53- 0.78) 1.24E- 05 (30- 40%] 1148 100 0.68 (0.53- 0.87) - 1054 188 0.7 2 (0.6-0.87) 4.72E- 04 2854 159 0.7 3 (0.6-0.88) 1.05E- 03 (40- 60%] 2210 285 1.00 ref - 2107 526 1.0 0 ref - 5709 421 1.0 0 ref - (60- 70%] 1080 166 1.21 (0.98- 1.50) - 1054 354 1.3 5 (1.16- 1.58) 1.36E- 04 2854 257 1.2 5 (1.06- 1.47) 8.43E- 03 (70- 80%] 1064 183 1.42 (1.15- 1.74) - 1053 371 1.4 3 (1.22- 1.66) 5.38E- 06 2855 273 1.3 0 (1.11- 1.53) 1.25E- 03 142 (80- 90%] 1024 223 1.87 (1.54- 2.29) - 1054 521 2.0 0 (1.74- 2.31) 1.98E- 21 2854 358 1.7 6 (1.51- 2.05) <1E-8 (90- 100%] 883 365 3.89 (3.24- 4.68) - 1054 866 3.3 9 (2.97- 3.87) 9.34E- 73 2855 578 3.0 1 (2.62- 3.45) <1E-8 (99- 100%] - - - - - 106 173 6.7 4 (5.19- 8.77) 4.03E- 46 286 112 6.3 0 (4.92- 8.06) <1E-8 Per SD GRS 10980 1868 2.04 (1.91- 2.18) 1.5e-93 10537 3244 1.9 7 (1.88- 2.06) 1.19E- 198 28546 2352 2.1 4 (2.03- 2.25) 8.05E- 176 GRS 451 [0-10%] 1098 51 0.34 (0.25- 0.46) 2.33E-11 1054 75 0.3 2 (0.24- 0.41) 9.33E- 19 2855 57 0.2 7 (0.21- 0.36) <1E-8 (10- 20%] 1098 65 0.45 (0.33- 0.59) 3.26E-08 1054 130 0.5 5 (0.44- 0.67) 1.26E- 08 2855 88 0.4 3 (0.34- 0.54) <1E-8 (20- 30%] 1098 97 0.68 (0.53- 0.87) 2.34E-03 1053 164 0.7 0 (0.57- 0.84) 2.35E- 04 2854 130 0.6 4 (0.52- 0.78) 1.83E- 05 (30- 40%] 1098 120 0.84 (0.67- 1.06) 1.53E-01 1054 220 0.9 4 (0.78- 1.12) 4.72E- 01 2855 164 0.8 2 (0.68- 0.99) 4.17E- 02 (40- 60%] 2197 276 1.00 ref 2107 475 1.0 0 ref - 5709 404 1.0 0 ref - (60- 70%] 1098 187 1.39 (1.13- 1.71) 6.00E-03 1054 305 1.2 9 (1.1-1.52) 1.92E- 03 2854 257 1.3 2 (1.12- 1.55) 1.13E- 03 (70- 80%] 1097 276 2.06 (1.70- 2.50) 7.71E-13 1053 446 1.9 1 (1.64- 2.22) 3.26E- 17 2855 266 1.3 7 (1.16- 1.61) 1.92E- 04 (80- 90%] 1098 293 2.36 (1.95- 2.85) 4.92E-18 1054 519 2.2 3 (1.93- 2.58) 5.20E- 27 2854 369 1.9 3 (1.66- 2.25) <1E-8 (90- 100%] 1098 503 4.33 (3.64- 5.15) 4.35E-60 1054 910 3.9 5 (3.45- 4.52) 6.52E- 89 2855 617 3.5 1 (3.06- 4.03) <1E-8 (99- 100%] 110 101 10.1 3 (7.32- 14.03) 3.42E-43 106 196 8.6 6 (6.69- 11.22) 5.79E- 60 286 127 8.1 3 (6.39- 10.34) <1E-8 Per SD GRS 10980 1868 2.21 (2.08- 2.34) 1.797E- 157 10537 3244 2.0 9 (2.00- 2.18) 2.06E- 226 28546 2352 2.3 2 (2.20- 2.45) 9.90E- 211 GRS GRS Categor y MADCaP (African) MGB (African) MGI (African) Control s Case s OR (95%CI) P value Control s Case s OR (95%CI) P value Control s Case s OR (95%CI) P value GRS26 9 [0-10%] 216 102 0.4 6 (0.35- 0.61) 4.25E- 08 52 1 0.1 5 (0.01-0.92) - 45 10 0.6 1 (0.27-1.37) 2.28E- 01 (10- 20%] 216 139 0.6 1 (0.48- 0.79) 1.78E- 04 46 6 1.2 1 (0.37-3.70) - 45 9 0.5 0 (0.22-1.17) 1.12E- 01 143 (20- 30%] 216 153 0.7 (0.54- 0.89) 4.65E- 03 48 5 0.7 3 (0.21-2.32) - 45 9 0.5 4 (0.23-1.27) 1.59E- 01 (30- 40%] 216 149 0.6 6 (0.51- 0.85) 1.20E- 03 47 5 0.5 9 (0.16-1.93) - 45 11 0.6 0 (0.27-1.31) 2.00E- 01 (40- 60%] 432 457 1 ref - 94 10 1.0 0 ref - 90 38 1.0 0 ref - (60- 70%] 216 219 0.9 6 (0.76- 1.22) 7.58E- 01 48 5 0.7 7 (0.21-2.49) - 45 20 1.0 9 (0.55-2.15) 8.09E- 01 (70- 80%] 216 306 1.4 1 (1.13- 1.76) 2.48E- 03 45 7 1.3 1 (0.42-3.90) - 45 30 1.8 4 (0.99-3.45) 5.54E- 02 (80- 90%] 216 439 2.0 3 (1.64- 2.51) 8.76E- 11 39 13 2.2 7 (0.85-6.16) - 45 15 0.7 6 (0.37-1.56) 4.49E- 01 (90- 100%] 216 541 2.5 5 (2.07- 3.14) 2.22E- 18 38 15 3.8 1 (1.48- 10.19) - 45 47 2.9 2 (1.61-5.27) 4.00E- 04 (99- 100%] 22 78 3.6 4 (2.21- 5.98) 3.82E- 07 - - - - - 5 5 2.5 6 (0.65- 10.08) 1.79E- 01 Per SD GRS 2160 2505 1.6 6 (1.56- 1.76) 2.00E- 59 - - - - - 450 189 1.6 2 (1.34-1.95) 5.79E- 07 GRS 451 [0-10%] 216 100 0.5 2 (0.4-0.69) 5.61E- 06 48 - - - - 45 6 0.5 3 (0.2-1.43) 2.10E- 01 (10- 20%] 216 123 0.6 6 (0.5-0.86) 2.06E- 03 47 5 0.9 2 (0.25-2.98) 8.89E- 01 45 10 0.9 9 (0.42-2.32) 9.81E- 01 (20- 30%] 216 137 0.7 1 (0.55- 0.92) 9.14E- 03 47 11 1.5 5 (0.54-4.41) 1.59E+0 0 45 16 1.4 8 (0.69-3.2) 3.17E- 01 (30- 40%] 216 151 0.8 0 (0.62- 1.03) 8.65E- 02 47 3 0.3 8 (0.08-1.40) 1.77E- 01 45 11 1.0 6 (0.46-2.44) 8.99E- 01 (40- 60%] 432 388 1.0 0 ref - 94 12 1.0 0 ref - 90 23 1.0 0 ref - (60- 70%] 216 280 1.4 9 (1.19- 1.88) 6.00E- 04 47 8 1.4 0 (0.46-4.12) 6.65E- 01 45 27 2.3 4 (1.16-4.7) 1.70E- 02 (70- 80%] 216 302 1.6 0 (1.28- 2.01) 4.69E- 05 47 8 1.4 7 (0.49-4.29) 6.22E- 01 45 26 2.3 9 (1.18-4.84) 1.58E- 02 (80- 90%] 216 389 2.1 4 (1.71- 2.67) 1.44E- 11 47 17 3.1 4 (1.23-8.27) 4.85E- 02 45 32 3.2 4 (1.63-6.42) 7.63E- 04 (90- 100%] 216 635 3.5 4 (2.87- 4.37) 6.08E- 32 47 21 4.9 3 (2.00- 12.74) 2.52E- 03 45 38 3.5 4 (1.82-6.86) 1.83E- 04 (99- 100%] 22 76 4.2 5 (2.57- 7.01) 1.50E- 08 5 1 2.5 6 (0.10- 28.48) 6.29E- 01 5 4 2.9 0 (0.67- 12.66) 1.56E- 01 Per SD GRS 2160 2505 2 (1.67- 1.90) 1.29E- 73 471 85 2.0 8 (1.54-2.86) 1.53E- 05 450 189 1.7 6 (1.45-2.13) 9.87E- 09 144 Supplementary Table 5. The association of GRS and prostate cancer risk stratified by age GWAS Discovery Samples Age ≤ 55 years Age > 55 years P-heterogeneity GRS GRS Category* Controls Cases OR (95%CI) P value Controls Cases OR (95%CI) P value European GRS451 [0-10%] 11210 96 0.20 (0.16-0.25) 7.70E-45 34145 1419 0.24 (0.23-0.26) <2.23E-308 9.05E-02 (10-20%] 11208 160 0.32 (0.27-0.39) 7.47E-33 34143 2375 0.40 (0.38-0.43) 3.43E-237 1.95E-02 (20-30%] 11208 234 0.46 (0.39-0.54) 1.09E-20 34141 3261 0.56 (0.53-0.59) 4.28E-119 1.89E-02 (30-40%] 11208 293 0.58 (0.49-0.67) 3.79E-12 34144 4229 0.72 (0.69-0.75) 1.34E-44 6.75E-03 (40-60%] 22416 995 1.00 ref - 68284 11763 1.00 ref - - (60-70%] 11207 632 1.28 (1.12-1.45) 1.52E-04 34143 8196 1.39 (1.34-1.45) 5.69E-60 1.99E-01 (70-80%] 11209 898 1.76 (1.56-1.98) 9.71E-21 34142 10297 1.75 (1.68-1.82) 5.37E-184 9.45E-01 (80-90%] 11207 1427 2.85 (2.55-3.17) 6.13E-79 34142 14310 2.38 (2.3-2.47) <2.23E-308 2.38E-03 (90-100%] 11210 3523 7.17 (6.49-7.92) <2.23E-308 34144 28010 4.60 (4.45-4.75) <2.23E-308 1.07E-16 (99-100%] 1123 875 17.62 (14.78-21) 3.37E-225 3417 5927 9.00 (8.44-9.59) <2.23E-308 1.75E-12 Per SD 112083 8258 2.86 (2.76-2.97) 0.00E+00 341428 83860 2.27 (2.25-2.3) 0.00E+00 8.73E-33 African GRS451 [0-10%] 1340 54 0.24 (0.18-0.34) 3.64E-18 5079 340 0.26 (0.23-0.3) 7.45E-95 7.07E-01 (10-20%] 1338 84 0.41 (0.31-0.54) 7.97E-10 5075 570 0.44 (0.4-0.49) 1.23E-50 5.97E-01 (20-30%] 1335 120 0.58 (0.46-0.74) 8.72E-06 5076 732 0.59 (0.54-0.65) 1.76E-25 8.92E-01 (30-40%] 1340 150 0.69 (0.55-0.87) 1.45E-03 5076 903 0.73 (0.66-0.8) 9.30E-12 7.13E-01 (40-60%] 2673 416 1.00 ref - 10151 2463 1.00 ref - - (60-70%] 1340 282 1.38 (1.14-1.67) 1.13E-03 5076 1519 1.23 (1.14-1.34) 5.01E-07 3.02E-01 (70-80%] 1335 365 1.84 (1.54-2.2) 3.29E-11 5076 2023 1.63 (1.51-1.76) 2.32E-35 2.23E-01 (80-90%] 1338 623 2.91 (2.46-3.44) 4.69E-36 5075 2428 1.99 (1.85-2.15) 4.07E-73 4.85E-05 (90-100%] 1340 974 4.77 (4.08-5.58) 1.22E-85 5078 4046 3.23 (3.01-3.46) 2.30E-237 7.30E-06 (99-100%] 138 238 10.96 (8.29-14.48) 1.01E-63 510 705 5.49 (4.76-6.33) 2.68E-122 1.47E-05 Per SD 21786 12728 2.41 (2.29-2.54) 2.37E-247 53134 17103 1.98 (1.94-2.03) 0.00E+00 7.97E-12 Asian GRS451 [0-10%] 9 1 0.00 (0-0.3) 1.84E-02 173 29 0.27 (0.17-0.41) 2.30E-09 5.59E-02 (10-20%] 8 1 2.10 (0.04-108.03) 7.11E-01 172 68 0.62 (0.44-0.86) 4.20E-03 5.42E-01 (20-30%] 8 1 0.03 (0-18.56) 2.76E-01 172 83 0.74 (0.54-1.02) 6.53E-02 3.17E-01 (30-40%] 8 3 2.13 (0.08-54.53) 6.47E-01 172 83 0.76 (0.55-1.05) 9.27E-02 5.36E-01 (40-60%] 16 9 1.00 ref - 344 219 1.00 ref - - (60-70%] 8 5 0.28 (0.02-3.73) 3.35E-01 172 139 1.27 (0.95-1.68) 1.04E-01 2.56E-01 (70-80%] 8 8 2.75 (0.24-32.06) 4.20E-01 172 223 2.04 (1.57-2.67) 1.29E-07 8.15E-01 (80-90%] 8 6 6.18 (0.26-147.76) 2.61E-01 172 266 2.44 (1.88-3.16) 1.78E-11 5.66E-01 (90-100%] 8 12 1.41 (0.05-37.34) 8.36E-01 173 496 4.52 (3.53-5.77) 2.06E-33 4.88E-01 (99-100%] 2 4 11.46 (0.03-4328.53) 4.20E-01 18 106 9.33 (5.47-15.9) 2.31E-16 9.46E-01 Per SD 1803 1652 2.35 (1.06-5.19) 3.47E-02 1803 1652 2.14 (1.98-2.32) 3.88E-82 8.20E-01 Hispanic 145 GRS451 [0-10%] 424 6 0.23 (0.07-0.77) 1.70E-02 1849 50 0.22 (0.16-0.3) 8.04E-21 9.15E-01 (10-20%] 422 4 0.23 (0.06-0.81) 2.18E-02 1847 111 0.48 (0.38-0.61) 1.76E-09 2.58E-01 (20-30%] 422 13 0.98 (0.46-2.07) 9.53E-01 1847 158 0.68 (0.55-0.85) 5.57E-04 3.65E-01 (30-40%] 422 21 1.04 (0.49-2.2) 9.16E-01 1847 185 0.81 (0.66-0.99) 3.98E-02 5.23E-01 (40-60%] 845 34 1.00 ref - 3694 452 1.00 ref - - (60-70%] 422 33 2.13 (1.16-3.91) 1.47E-02 1847 284 1.28 (1.07-1.54) 7.85E-03 1.16E-01 (70-80%] 422 36 2.85 (1.54-5.28) 8.56E-04 1847 402 1.76 (1.48-2.09) 1.66E-10 1.37E-01 (80-90%] 422 53 3.35 (1.85-6.06) 6.53E-05 1847 512 2.32 (1.97-2.74) 5.11E-24 2.44E-01 (90-100%] 424 86 6.04 (3.48-10.48) 1.63E-10 1849 868 3.70 (3.18-4.31) 3.16E-63 9.32E-02 (99-100%] 44 13 13.68 (5.52-33.93) 1.65E-08 187 159 6.00 (4.47-8.07) 1.54E-32 9.11E-02 Per SD 6161 2327 2.37 (2.01-2.79) 9.65E-25 18636 3207 2.09 (1.99-2.19) 5.03E-192 1.48E-01 Replication Sample Age ≤ 55 years Age > 55 years P-heterogeneity GRS GRS Category* Controls Cases OR (95%CI) P value Controls Cases OR (95%CI) P value MGI (European) GRS451 Per SD 1622 332 2.69 (2.32-3.13) 2.97E-38 8915 2912 2.03 (1.93-2.12) 2.11E-187 2.09E-16 EstBB (European) GRS451 [0-10%] 1190 5 0.25 (0.10-0.63) 3.28E-03 1665 50 0.28 (0.21-0.39) 0.00E+00 7.81E-01 (10-20%] 1190 8 0.38 (0.18-0.82) 1.39E-02 1665 75 0.43 (0.33-0.55) 0.00E+00 7.89E-01 (20-30%] 1190 16 0.84 (0.47-1.50) 5.51E-01 1664 114 0.66 (0.53-0.82) 1.84E-04 4.53E-01 (30-40%] 1190 21 1.05 (0.61-1.79) 8.65E-01 1665 142 0.82 (0.67-1.01) 5.83E-02 3.85E-01 (40-60%] 2380 40 1.00 ref - 3329 347 1.00 ref - - (60-70%] 1189 28 1.42 (0.87-2.31) 1.64E-01 1665 233 1.36 (1.14-1.62) 6.30E-04 8.80E-01 (70-80%] 1190 33 1.68 (1.05-2.68) 3.01E-02 1664 240 1.39 (1.17-1.66) 2.25E-04 4.59E-01 (80-90%] 1190 46 2.36 (1.53-3.64) 9.56E-05 1665 316 1.84 (1.56-2.17) 0.00E+00 2.86E-01 (90-100%] 1190 101 5.24 (3.60-7.63) 0.00E+00 1665 537 3.22 (2.78-3.74) 0.00E+00 1.75E-02 (99-100%] 119 20 11.25 (6.29-20.11) 0.00E+00 167 119 7.46 (5.74-9.70) 0.00E+00 2.10E-01 Per SD 11899 298 2.63 (2.29-3.02) 2.43E-43 16647 2054 2.24 (2.12-2.38) 6.70E-163 3.56E-02 MADCaP (African) GRS451 [0-10%] 23 4 0.44 (0.12-1.54) 1.99E-01 191 76 0.48 (0.36-0.66) 4.42E-06 8.81E-01 (10-20%] 23 4 0.44 (0.12-1.55) 1.98E-01 190 101 0.63 (0.47-0.84) 1.81E-03 5.74E-01 (20-30%] 23 9 1.05 (0.38-2.95) 9.21E-01 191 109 0.67 (0.51-0.89) 5.94E-03 4.08E-01 (30-40%] 22 5 0.49 (0.15-1.62) 2.40E-01 190 118 0.73 (0.55-0.97) 2.84E-02 5.16E-01 (40-60%] 46 17 1.00 ref - 381 315 1.00 ref - - (60-70%] 22 18 2.15 (0.87-5.28) 9.55E-02 190 213 1.38 (1.07-1.77) 1.18E-02 3.53E-01 (70-80%] 23 14 1.76 (0.69-4.46) 2.36E-01 191 235 1.47 (1.15-1.88) 2.24E-03 7.17E-01 (80-90%] 23 22 3.12 (1.3-7.53) 1.12E-02 190 276 1.83 (1.44-2.34) 1.02E-06 2.52E-01 (90-100%] 23 41 5.84 (2.59-13.16) 2.03E-05 191 501 3.31 (2.64-4.16) 7.79E-25 1.87E-01 (99-100%] 3 7 6.30 (1.39-28.59) 1.71E-02 20 71 4.45 (2.63-7.52) 2.56E-08 6.71E-01 Per SD 255 561 2.24 (1.73-2.90) 1.16E-09 1932 2371 1.76 (1.65-1.89) 3.72E-60 7.96E-02 146 Supplementary Table 6. The association of GRS and aggressive prostate cancer risk GWAS Discovery Samples Aggressive vs. Non-Aggressive Cases GRS GRS Category* Non- Aggressive Cases Aggressive Cases OR (95%CI) P value European GRS451 [0-10%] 810 454 1.22 (1.06-1.39) 3.85E-03 (10-20%] 1361 689 1.09 (0.97-1.21) 1.33E-01 (20-30%] 1929 907 1.03 (0.94-1.14) 5.19E-01 (30-40%] 2550 1187 1.02 (0.94-1.11) 6.34E-01 (40-60%] 7052 3271 1.00 ref - (60-70%] 4889 2268 0.98 (0.92-1.06) 6.46E-01 (70-80%] 6182 2937 0.99 (0.93-1.06) 8.60E-01 (80-90%] 8578 4055 1.00 (0.94-1.06) 9.21E-01 (90-100%] 17013 8401 1.00 (0.95-1.06) 8.60E-01 (99-100%] 3624 1865 1.04 (0.96-1.12) 3.63E-01 Per SD GRS 50364 24169 0.99 (0.97-1.01) 2.73E-01 African GRS451 [0-10%] 263 105 1.26 (0.97-1.63) 8.74E-02 (10-20%] 438 171 1.16 (0.93-1.43) 1.83E-01 (20-30%] 578 253 1.22 (1.01-1.47) 4.12E-02 (30-40%] 725 275 1.04 (0.87-1.24) 7.00E-01 (40-60%] 1920 732 1.00 ref - (60-70%] 1175 501 1.16 (1-1.35) 4.69E-02 (70-80%] 1523 683 1.21 (1.06-1.39) 5.80E-03 (80-90%] 1891 887 1.24 (1.09-1.42) 1.07E-03 (90-100%] 3198 1512 1.31 (1.16-1.47) 9.00E-06 (99-100%] 553 282 1.45 (1.2-1.76) 1.18E-04 Per SD GRS 11711 5119 1.08 (1.04-1.12) 1.10E-04 Asian GRS451 [0-10%] 14 13 0.66 (0.28-1.56) 3.46E-01 (10-20%] 32 30 0.93 (0.5-1.72) 8.09E-01 (20-30%] 34 37 0.94 (0.53-1.69) 8.39E-01 (30-40%] 33 41 0.95 (0.53-1.69) 8.60E-01 (40-60%] 96 110 1.00 ref - (60-70%] 57 74 1.23 (0.76-2) 4.02E-01 (70-80%] 96 97 0.96 (0.63-1.46) 8.45E-01 (80-90%] 125 128 0.89 (0.6-1.31) 5.53E-01 (90-100%] 220 223 0.96 (0.67-1.37) 8.05E-01 (99-100%] 50 49 0.93 (0.55-1.55) 7.76E-01 Per SD GRS 707 753 1.01 (0.9-1.12) 8.98E-01 Hispanic GRS451 [0-10%] 34 19 2.05 (1.06-3.98) 3.36E-02 (10-20%] 70 37 1.54 (0.95-2.5) 8.15E-02 (20-30%] 101 53 1.40 (0.91-2.16) 1.26E-01 (30-40%] 133 54 1.14 (0.77-1.69) 5.21E-01 (40-60%] 351 126 1.00 ref - (60-70%] 219 79 0.99 (0.7-1.4) 9.70E-01 (70-80%] 302 125 1.07 (0.78-1.47) 6.62E-01 (80-90%] 373 159 1.18 (0.88-1.59) 2.72E-01 (90-100%] 619 285 1.15 (0.88-1.5) 3.06E-01 (99-100%] 109 62 1.51 (0.99-2.28) 5.37E-02 Per SD GRS 2202 937 1.00 (0.92-1.08) 9.67E-01 Replication Sample Aggressive vs. Non-Aggressive Cases 147 GRS GRS Category* Non- Aggressive Cases Aggressive Cases OR (95%CI) P value MADCaP (African) GRS451 [0-10%] 25 65 0.85 (0.49-1.46) 5.49E-01 (10-20%] 26 90 1.25 (0.74-2.1) 4.03E-01 (20-30%] 28 98 1.18 (0.71-1.94) 5.24E-01 (30-40%] 29 113 1.15 (0.7-1.9) 5.75E-01 (40-60%] 94 272 1.00 ref - (60-70%] 60 199 1.11 (0.75-1.64) 6.17E-01 (70-80%] 63 223 1.21 (0.83-1.78) 3.24E-01 (80-90%] 89 270 1.02 (0.72-1.45) 9.13E-01 (90-100%] 123 476 1.40 (1.01-1.94) 4.08E-02 (99-100%] 9 65 2.78 (1.3-5.94) 8.27E-03 Per SD GRS 1806 537 1.12 (1.01-1.23) 2.83E-02 148 Supplementary Table 7. The association of per SD GRS between aggressive prostate cancer in GWAS discovery samples Population Controls Cases GRS OR (95%CI) P value European 50364 24169 GRS451 0.99 (0.97-1.01) 2.73E-01 GRS400* 1.04 (1.03-1.06) 3.21E-07 GRS48** 0.96 (0.95-0.98) 5.36E-06 African 11711 5119 GRS451 1.08 (1.04-1.12) 1.10E-04 GRS400 1.10 (1.06-1.14) 6.95E-07 GRS48 1.00 (0.96-1.04) 9.16E-01 Asian 707 753 GRS451 1.01 (0.90-1.12) 8.98E-01 GRS400 1.01 (0.91-1.13) 8.09E-01 GRS48 1.00 (0.90-1.11) 9.76E-01 Hispanic 2202 937 GRS451 1.00 (0.92-1.08) 9.67E-01 GRS400 1.05 (0.97-1.14) 2.12E-01 GRS48 0.93 (0.85-1.01) 8.33E-02 Multi-ancestry META 64984 30978 GRS451 1.01 (0.99-1.02) 4.61E-01 GRS400 1.05 (1.03-1.06) 1.62E-11 GRS48 0.97 (0.95-0.98) 1.01E-05 * GRS400 was calculated with the conditional weighted dosage of the 400 prostate cancer risk SNPs excluding 3 KLK3 SNPs and 48 SNPs overlapping with the PSA GWAS SNPs; ** GRS48 was calculated with the conditional weighted dosage of the 48 PSA overlapping risk SNPs. 149 Supplementary Figure 1. Comparison of ancestry-specific ORs between European and African, Asian, and Hispanic populations, respectively. Variants present in both populations are compared; the number of variants is denoted in the lower right corner. Genome-wide significant variants among African, Asian, or Hispanic populations are highlighted in orange. The Pearson’s correlation coefficient between effect sizes and corresponding p-value are denoted in the upper left in each sub-panel. 150 Supplementary Figure 2. Sankey diagram of GRS risk categorization based on GRS 100, GRS 181, GRS 269, and GRS 451 in the multi-ancestry sample. (a) GRS quantiles in all controls; (b) GRS quantiles in all cases. Percentage of individuals in each GRS quantile are labelled in corresponding boxes. Percentage of controls that remain in the lowest quintile [0%, 20%] and highest quintile (80%, 100%] from a previous to a more current GRS are indicated on corresponding flows in (a). In (b), the highest GRS quintile contains 51.2% of the cases. (a) (b) 151 Chapter Four Supplemental Table 1. Variant allelic fraction (VAF) and minor allele frequency (MAF) distributions for each CHIP variant in the UK Biobank. VAF (%) Gene Chr Pos Ref Alt Median (Min-Max) MAF # Carriers Consequence GNB1 1 1815788 C G 16.93 16.22-17.65 1.44E-05 2 nonsynonymous SNV GNB1 1 1815788 C A 13.13 8.62-17.65 1.44E-05 2 nonsynonymous SNV GNB1 1 1815790 T C 16.67 7.32-44.83 0.00017266 24 nonsynonymous SNV MPL 1 43352635 T G 5.97 5.97-5.97 7.19E-06 1 nonsynonymous SNV NRAS 1 114716126 C T 15.15 11.11-26.67 2.16E-05 3 nonsynonymous SNV NRAS 1 114716127 C A 12.5 12.5-12.5 7.19E-06 1 nonsynonymous SNV SETDB1 1 150927784 G T 68.29 68.29-68.29 7.19E-06 1 stopgain SETDB1 1 150927922 C T 48.57 48.57-48.57 7.19E-06 1 stopgain SETDB1 1 150927973 - T 6.12 6.12-6.12 7.19E-06 1 frameshift insertion … Note: Due to space limitation, the full content of this table can be found online (https://doi.org/10.1093/hmg/ddac214). 152 Supplemental Table 2. Variant allelic fraction (VAF) and minor allele frequency (MAF) distributions for each CHIP variant in WESP. VAF (%) Gene Chr Pos Ref Alt Median (Min-Max) MAF # Carriers Consequence DNMT3A 2 25463287 G A 18.21 15-21.43 0.00018034 2 nonsynonymous DNMT3A 2 25470620 T C 21.43 21.43-21.43 9.02E-05 1 . DNMT3A 2 25463596 G A 38.46 38.46-38.46 9.0171E-05 1 stopgain DNMT3A 2 25470516 G A 25.22 20-30.43 0.00018034 2 stopgain DNMT3A 2 25470904 A C 38.89 38.89-38.89 9.02E-05 1 . DNMT3A 2 25457243 G A 15.35 13.04-27.42 7.21E-04 8 nonsynonymous SNV DNMT3A 2 25457272 ACTGGGA - 17.95 17.95-17.95 9.02E-05 1 frameshift deletion DNMT3A 2 25463184 G A 17.11 10.94-23.29 1.80E-04 2 nonsynonymous SNV DNMT3A 2 25467158 G A 10.08 10.08-10.08 9.02E-05 1 stopgain DNMT3A 2 25467478 T C 10 10.00-10.00 9.02E-05 1 nonsynonymous SNV DNMT3A 2 25468202 C T 10.71 10.71-10.71 9.02E-05 1 . … Note: Due to space limitation, the full content of this table can be found online (https://doi.org/10.1093/hmg/ddac214). 153 Supplemental Figure 1. Fifteen most common CHIP genes in the A) UK Biobank and B) WESP. The X-axis indicates the number of carriers for the indicated gene. CHIP: Clonal hematopoiesis of indeterminate potential; WESP: WESP: Whole-Exome Sequencing Study in Prostate Cancer. A. B. 154 Supplemental Figure 2. Age at blood draw in the UK Biobank. A) Distribution of age at blood draw by case and control status. B) CHIP carrier frequency by age at blood draw. CHIP: Clonal hematopoiesis of indeterminate potential. A. B. 155 Supplemental Figure 3. Manhattan plots of associations between CHIP genes and A) age at prostate cancer diagnosis and B) age at prostate cancer diagnosis adjusting for age at blood draw. CHIP: Clonal hematopoiesis of indeterminate potential. 156 Supplemental Figure 4. Manhattan plots of associations between CHIP genes and prostate cancer in A) aggressive versus non-aggressive prostate cancer in WESP, B) death due to prostate cancer versus non-aggressive prostate cancer in WESP, C) metastatic versus non-aggressive prostate cancer in WESP, and D) overall prostate cancer versus prostate cancer-free controls in the UK Biobank. CHIP: Clonal hematopoiesis of indeterminate potential; WESP: Whole-Exome Sequencing Study in Prostate Cancer. 157 Chapter Five Supplemental Table 1. Hazard Ratios of Prostate Cancer Outcomes Associated with e-DII in the Multiethnic Cohort, 1993-2017 (N=74,714) Quartiles of e-DII P value for Trend b Per SD increase in e-DII Q1 Q2 Q3 Q4 HR (95%CI) P-value Incident prostate cancer a Total, No. 18661 18670 18664 18719 74714 Total Events, No. 2405 2269 2122 1901 8697 HR (95%CI) 1 [Reference] 0.98 (0.93,1.04) 0.98 (0.92,1.04) 0.96 (0.90,1.02) 0.18 0.98 (0.96,1.00) 0.12 Aggressive Events, No. 774 683 653 600 2710 HR (95%CI) 1 [Reference] 0.93 (0.84,1.03) 0.94 (0.85,1.05) 0.94 (0.84,1.05) 0.26 0.97 (0.94,1.01) 0.20 Low grade Events, No. 1616 1570 1465 1292 5943 HR (95%CI) 1 [Reference] 0.99 (0.92,1.06) 0.97 (0.90,1.04) 0.92 (0.85,0.99) 0.03 0.97 (0.94,1.00) 0.03 High grade Events, No. 630 536 506 467 2139 HR (95%CI) 1 [Reference] 0.93 (0.83,1.04) 0.95 (0.84,1.07) 0.97 (0.85,1.10) 0.62 0.98 (0.94,1.02) 0.35 Localized Events, No. 1866 1709 1585 1399 6559 HR (95%CI) 1 [Reference] 0.97 (0.90,1.03) 0.96 (0.90,1.03) 0.93 (0.86,1.00) 0.05 0.97 (0.94,0.99) 0.01 Regional Events, No. 222 213 229 214 878 HR (95%CI) 1 [Reference] 0.91 (0.75,1.10) 0.96 (0.79,1.16) 0.89 (0.73,1.09) 0.36 1.00 (0.93,1.07) 0.91 Metastatic Events, No. 112 114 101 111 438 HR (95%CI) 1 [Reference] 1.10 (0.84,1.43) 1.07 (0.81,1.40) 1.30 (0.99,1.71) 0.09 1.07 (0.97,1.18) 0.17 Prostate cancer mortality a Total, No. 18661 18670 18664 18719 74714 Events, No. 339 281 310 240 1170 HR (95%CI) 1 [Reference] 0.91 (0.78,1.07) 1.14 (0.97,1.33) 1.03 (0.87,1.23) 0.27 1.04 (0.98,1.11) 0.19 Prostate cancer survival a Total, No. 2350 2221 2066 1854 8491 Prostate cancer- specific mortality Events, No. 289 238 258 199 984 HR (95%CI) 1 [Reference] 0.93 (0.78,1.10) 1.12 (0.95,1.34) 1.05 (0.87,1.27) 0.29 1.05 (0.98,1.12) 0.15 All-cause mortality Events, No. 1460 1290 1133 971 4854 HR (95%CI) 1 [Reference] 1.02 (0.95,1.10) 1.02 (0.95,1.11) 1.13 (1.03,1.23) 0.01 1.04 (1.01,1.07) 0.01 a Adjusted for year at cohort entry, ethnicity, education, smoking, baseline BMI, diabetes, family history, and aspirin/statin intake, survival model (restricted to cases) additionally adjusted for prostate cancer grade and stage. b P value for trend was calculated by modeling the median of each quartile as a continuous term. 158 Supplemental Table 2. Odds Ratios of Association between AACs status and PSA Screening History before the 2 nd Questionnaire (N=56,534) a AACs status PSA screening OR (95% CI) Never Ever No AACs 24655 19553 1 [reference] AACs 6048 6278 1.27 (1.22,1.33) No AACs 24655 19553 1 [reference] AACs w/o med 4073 3875 1.20 (1.14,1.26) AACs w/ med 1975 2403 1.41 (1.32,1.50) Less than 1 year 426 468 1 [reference] 2-5 Years 551 657 1.00 (1.00,1.00) >5 Years 806 1077 1.04 (0.87,1.24) a Adjusted for age at entry, race/ethnicity, education, baseline BMI, diabetes, family history, aspirin/statin intake. 159 Supplemental Table 3. Hazard Ratios of Prostate Cancer Outcomes Associated with AACs Among Men with PSA Screening Information in the Multiethnic Cohort, 1998-2017 (N=56,534) a Events Total Minimally adjusted HR (95% CI) b Fully adjusted HR (95% CI) c PSA screening adjusted HR (95% CI) d Incident prostate cancer Total 5403 56534 0.95 (0.89,1.01) 0.94 (0.88,1.00) 0.93 (0.87,0.99) Aggressive 1697 55904 0.94 (0.84,1.05) 0.93 (0.83,1.04) 0.93 (0.83,1.03) Low-grade 3645 56118 0.96 (0.89,1.03) 0.94 (0.88,1.02) 0.93 (0.86,1.00) High-grade 1342 56118 0.90 (0.79,1.02) 0.90 (0.79,1.02) 0.89 (0.79,1.01) Localized 4036 55988 0.93 (0.87,1.00) 0.92 (0.86,0.99) 0.91 (0.85,0.98) Regional 548 55988 1.02 (0.84,1.23) 0.99 (0.82,1.20) 0.99 (0.82,1.19) Metastatic 273 55988 0.82 (0.62,1.10) 0.85 (0.63,1.13) 0.85 (0.64,1.14) Prostate cancer mortality 684 56534 0.74 (0.61,0.91) 0.75 (0.61,0.92) 0.75 (0.61,0.92) Prostate cancer survival e Prostate cancer- specific mortality 543 5252 0.71 (0.57,0.90) 0.72 (0.57,0.91) - All-cause death 2549 5252 0.87 (0.78,0.96) 0.88 (0.80,0.98) - a The date of 2nd questionnaire completion was used as the entry time for time-to-event analysis. b Adjusted for year at cohort entry and ethnicity. c Adjusted for year at cohort entry, ethnicity, education, smoking, baseline BMI, diabetes, family history, and aspirin/statin intake. d Adjusted for year at cohort entry, ethnicity, education, smoking, baseline BMI, diabetes, family history, aspirin/statin intake, and PSA screening history. e Additionally adjusted for prostate cancer stage and grade.

Abstract (if available)

Abstract Prostate cancer is a significant public health issue and one of the most common cancers among men worldwide. The exact causes of prostate cancer are not yet fully understood, but several risk factors have been identified, including age, family history, genetic risk variants, and environmental factors. This dissertation aims to investigate the genetic and environmental risk factors for prostate cancer and their impact on the development and progression of the disease. Specifically, Chapter 3 focuses on identifying novel genetic risk variants for prostate cancer in multi-ancestry populations through large-scale genome-wide association studies (GWAS). Chapter 3 involves conducting ancestry-specific and multi-ancestry GWAS meta-analyses to identify novel genetic risk variants and develop a genetic risk score to characterize prostate cancer risk in multi-ancestry populations. Chapter 4 investigates the association between clonal hematopoiesis, an aging-related expansion of blood cells with somatic mutations, and prostate cancer. Chapter 5 examines the effect of atopic allergic conditions, a hypertensive immune state, on the risk of prostate cancer, disease severity, and survival in multi-ethnic populations.

The findings from this thesis study have the potential to contribute to a better understanding of prostate cancer etiology and the development of new strategies for prevention, early detection, and treatment. Ultimately, this research may help improve the quality of life and survival rates of those affected by prostate cancer.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Genetic studies of cancer in populations of African ancestry and Latinos

PDF

Using genetic ancestry to improve between-population transferability of a prostate cancer polygenic risk score

PDF

twas_sim, a Python-based tool for simulation and power analysis of transcriptome-wide association analysis

PDF

The role of heritability and genetic variation in cancer and cancer survival

PDF

Genetic risk factors in multiple myeloma

PDF

Utility of polygenic risk score with biomarkers and lifestyle factors in the multiethnic cohort study

PDF

Identification and fine-mapping of genetic susceptibility loci for prostate cancer and statistical methodology for multiethnic fine-mapping

PDF

Dietary and supplementary folate intake and prostate cancer risk

PDF

Identifying genetic, environmental, and lifestyle determinants of ethnic variation in risk of pancreatic cancer

PDF

Examining the relationship between common genetic variation, type 2 diabetes and prostate cancer risk in the multiethnic cohort

PDF

Genes and environment in prostate cancer risk and prognosis

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Association of comorbidity with prostate cancer tumor characteristics in African American men

PDF

Understanding prostate cancer genetic susceptibility and chromatin regulation

PDF

Prostate cancer disparities among Californian Latinos by country of origin: clinical characteristics, incidence, treatment received and survival

PDF

Genomic risk factors associated with Ewing Sarcoma susceptibility

PDF

An analysis of disease-free survival and overall survival in inflammatory breast cancer

PDF

Genes and hormonal factors involved in the development or recurrence of breast cancer

PDF

Diet quality and pancreatic cancer incidence in the multiethnic cohort

PDF

Ancestral/Ethnic variation in the epidemiology and genetic predisposition of early-onset hematologic cancers

Asset Metadata

Creator Wang, Anqi (author)

Core Title Prostate cancer: genetic susceptibility and lifestyle risk factors

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Degree Conferral Date 2023-05

Publication Date 05/01/2024

Defense Date 03/20/2023

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag atopic allergic condition,clonal hematopoiesis,GWAS,OAI-PMH Harvest,prostate cancer

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Haiman, Christopher (committee chair), Conti, David (committee member), Goldkorn, Amir (committee member), Mancuso, Nicholas (committee member), Wu, Anna (committee member)

Creator Email angelwang0629@gmail.com,wanganqi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113089497

Unique identifier UC113089497

Identifier etd-WangAnqi-11748.pdf (filename)

Legacy Identifier etd-WangAnqi-11748

Document Type Dissertation

Format theses (aat)

Rights Wang, Anqi

Internet Media Type application/pdf

Type texts

Source 20230501-usctheses-batch-1034 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu