University of Southern California Dissertations and Theses

Molecular anatomy of Noonan syndrome mutations in the testes of unaffected men

(USC Thesis Other)

Molecular anatomy of Noonan syndrome mutations in the testes of unaffected men

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Molecular Anatomy of Noonan Syndrome Mutations in the Testes of Unaffected Men
By Jordan Eboreime
FACULTY OF THE USC GRADUATE SCHOOL
MOLECULAR AND COMPUTATIONAL BIOLOGY
DOCTOR OF PHILOSOPHY
MOLECULAR BIOLOGY
UNIVERSITY OF SOUTHERN CALIFORNIA
AUGUST 2016
Acknowledgements
Wow! I cannot believe that I have made it to the end of my doctoral education journey. I
would not have made it through without the constant support of my friends, family, and colleagues.
First, I would like to thank my advisor and mentor, Norm Arnheim, for always listening to my
“outside the box” ideas and for his guidance and support. He has taught me how to be an excellent
researcher and scientist and how to question everything! I would also like to thank, Peter Calabrese,
for his dedication and support and for always being patient with my computational questions. Thank
you to Ian Ehrenreich for his listening ear and to the Ehrenreich lab for always making me laugh.
Thank you Matt Dean and Daryl Shibata for being excellent committee members. I want to thank
Soo and Song for being my generous lab family and introducing me to so many delicious Korean
dishes. Thank you to the rest of my USC family including everyone in MCB, MCB administration,
and the USC Career Center.
To Melina Butuči, Ana Carolina Dantas Machado, and Sandra Villwock, thank you for
always being there for me when I wanted to talk, cry, sing, dance, and laugh. I would have lost all of
my sanity if it weren’t for you girls keeping me grounded throughout this entire process. Also, thank
you to Catherine Douglas and Alexandria Hernandez for continuing to support me as your sister and
for being dazzling lights of beauty in my life. I am so blessed to have such amazing friends.
Finally, I would like to pour out so much gratitude and love to my whole family! Thank you
Mom, Grandma, Darren, and Mark for encouraging me to do my best and break down barriers.
Thank you to my in-laws for your constant support, chats, and amazing dinners. You all inspire me
every day! Thank you to the most wonderful, beautiful, caring partner, Stephen Williams. You’re my
biggest fan and I love you! I am the luckiest girl in the world to marry you!
2
Table of Contents
Acknowledgements …………………………………………....………………………………….2
List of Tables ...…………………………………………...…….…………………………...…….5
List of Figures ……………………………………………...…...………………………..……….6
Abstract ………………………….…...………………...…...…………………………………….7
1. INTRODUCTION ……………………………………………………………………………8
1.1. SPONTANEOUS MUTATIONS……….……...……………………………….....….….8
1.2. GAMATEOGENESIS …….……..……………………………………………...….…...9
1.2.1. Spermatogenesis……………….….…………………………...….…..…………...9
1.2.2. Oogenesis………………...……………………………………...….…..………...12
1.3. MALE BIAS AND THE PATERNAL AGE EFFECT…...…………………..……....14
1.4. PAST RESEARCH AND RAMP DISEASES……………………...…..………..…...16
1.4.1. Noonan Syndrome……...…………………..………………..……..…………...18
1.4.2. Past Research on RAMP mutations using the PAP techniques……..……...…...21
1.5. NEXT GENERATION SEQUENCING……...………………………………..….…...23
1.6. SAFE SEQUENCE SYSTEM OVERVIEW...………………….……………..….…...24
2. METHODS……..………………...………………...………………………………..………26
2.1. SOURCE OF TESTIS DNA.…………………...……….…………...………..….……26
2.2. TESTIS DISSECTION/ DNA PURIFICATION AND DNA QUANTITATION...........26
2.3. MODIFIED SAFE SEQUENCE SYSTEM…...………..…..….………...………….…27
2.4. BARCODES………..…...…………………....……………..…………………………30
2.5. QUANTITATIVE METHODS………………....……………..….……..……..………30
2.6. SIMULATIONS……..……....…………………………….…..….……………………31
3
3. RESULTS……..……………...…………………...………………..………..………….……32
3.1. MODIFIED SAFE SEQUENCE SYSTEM……...……..……..………………………32
3.1.1. Sequencing Error…………………...……………..……....……………….……34
3.1.2. Supermutants and Error Rate….…...……….……..……....……………….……35
3.1.3. Germline Mutation Frequency Comparison of PAP and SSS…....…….....…..…36
3.1.4. Mutation Type Frequency ….…...……………………......……..…….....….…..36
3.1.5. Optimal UID Family Size and Utility Ratios…….……......…...…….....…...….38
3.2. MOLECULAR ANATOMY OF NOONAN MUTATIONS IN THE TESTES…..…..40
3.2.1. Noonan Syndrome Mutation Distribution…………...…..…..…………….….…42
3.2.2. Increased Mutation Frequency with C>T and G>T Mutation Types.…..…….....48
3.2.3. Control Sites…...……………………………………………...….…...……....…49
3.2.4. Models and Statistical Analysis……...………………………...…......…….....…51
3.3. CANCER-ASSOCIATED PTPN11 SITES……….………………….…...…......……56
4. DISCUSSION……………...……………………..…………………………….…...……….61
4.1. ANATOMICAL CONCLUSIONS………..……..…....…………..…………….…..…61
4.2. SPERMATOGONIAL SELECTION…………………..…...………...…………….…62
4.3. CANCER-ASSOCIATED PTPN11 MUTATIONS…..…...………………..……….…64
4.3.1. Biochemical Studies…………………..…..……..…………………...…….……65
4.4. ADV ANTAGES AND DRAWBACKS TO THE SSS PROTOCOL…..……..………..68
4.5. ALTERNATIVE APPROACHES………...………..………………...…..…...………..69
5. REFERENCES…………………………………..…………………………...…….….….….71
6. APPENDIX……………………………………...…………………………...…….….….….79
6.1. Noonan-associated mutations in exon 3 of PTPN11……...……………...….…….…..79
6.2. Cancer-associated mutations in exon 3 of PTPN11…….…...…………..….…….…...80
6.3. Control sites in exon 3 of PTPN11 .……………………………………...…….….….81
4
List of Tables
Table 1. Sequencing error compared to Illumina-derived quality score....……………......….….35
Table 2. Mutation frequencies for each mutation type................……………...………….….….38
Table 3. Germline mutations in exon 3 of PTPN11……………………………….….….…..….40
Table 4. Mutation Frequency for each Mutation Type..…. ……………………………....…......48
Table 5. Cancer-associated mutations in exon 3 of PTPN11………………………………........57
5
List of Figures
Figure 1. Testis Anatomy.…………………………………..……….…………..……..….….….10
Figure 2. Diagram of cell divisions in male and female germline development...…....…..…......13
Figure 3. Hypothetical distribution of mutated SSCs based on the hotspot model or the
selection model...………………... ………........………….............…………………......17
Figure 4. Diseases associated with the Ras-MAPK pathway....……………………..…....…......19
Figure 5. Structure of SHP2 and activation of PTP after phosphotyrosyl-peptide binding..........20
Figure 6. Distribution of the most common Noonan Syndrome mutation (PTPN11
c.922 A>G) in the testes of 15 unaffected donors of various ages......…….....…….….22
Figure 7. Safe Sequence System schematic....………………..….………..…………...….….….29
Figure 8. Schematic of modified Safe Sequence System Strategy with true mutation
versus PCR error.....……….…………..…………..….……………………..........….33
Figure 9. PTPN11 mutation frequencies color-coded by reference base..……..…..……….........37
Figure 10. Distribution of Noonan Syndrome mutation c.181 G>C in a 68 year
old man (testis #60891)....…....…...………………. ………..………….………............43
Figure 11. Distribution of twelve Noonan Syndrome sites in testis #60891...……...…...….....…45
Figure 12. Distribution of twelve Noonan Syndrome mutations in the testes of
three unaffected men aged 65 (60649), 68 (60891), and 21 (63878) …..………..….…...47
Figure 13. Noonan Syndrome mutations with C>T and G>T mutations types in the
testes of three unaffected men aged 65 (60649), 68 (60891), and 21 (63878)...............…49
Figure 14. Distribution of control sites in the testes of three unaffected men aged
65 (60649), 68 (60891), and 21 (63878).………………..…………..….…..…….............51
Figure 15. Symmetric versus Asymmetric Stem Cell Divisions...................….……….................52
Figure 16. Histogram of simulations of Mx/Av values for the four models…...............................55
Figure 17. Cancer-associated PTPN11 mutation distribution in the testes of three
unaffected men aged 65 (60649), 68 (60891), and 21 (63878).........….……...................60
Figure 18. Schematic effect of SHP2 on STAT3 and stem cell self-renewal.....……....................64
Figure 19. Phosphate activity assay of wildtype and mutated SHP2 proteins
performed by Tartaglia et al., 2006.......………………..…………..….………...............67
6
Abstract
Noonan Syndrome (NS) is one of the most common disorders. Recent evidence suggests the most
common NS mutation (PTPN11 c.922 G>A) likely confers a selective advantage on newly mutant
spermatogonial stem cells (SSC) in unaffected men, however, there are additional NS mutations in
the PTPN11 gene. The fact that many other NS mutations are also recurrent raises the possibility that
the high incidence of NS results from the accumulated effects of germline selection at many sites.
We asked if the high frequency of sporadic NS cases in humans is due to a high mutation rate per
cell division (hot spot) at a small number of disease sites or if rare mutations at these sites confer a
selective advantage to the mutated SSCs leading to high numbers of sperm carrying the NS disease
mutations (selection). We measured the NS mutation frequency in exon 3 of PTPN11 in three testes
from unaffected men using a modified Safe Sequence System version of NGS. In the older donors
(65 and 68 years old), almost all of the NS mutations are in a small number of often-adjacent testis
pieces with very high mutation frequencies (maximum testis piece mutation frequency is 5.9x10
-3
).
Computational modeling and statistical analysis show that the NS mutation clusters are inconsistent
with the hot spot model but consistent with the selection model.
7
1 INTRODUCTION
1.1 SPONTANEOUS MUTATIONS
Mutation is a term that refers to a permanent change to the genome of a cell. During DNA
replication, errors can be made that lead to new mutations. If these new mutations happen in the
germline of the individual, then they are transmissible to the next generation. These inherited
mutations can be damaging or advantageous to the organism, both types drive evolution through
phenotypic variation and adaptability. Transmitted pathogenic mutations can lead to diseases in the
following generations which contributes to the genetic burden of our species. These spontaneous
mutations can arise by a number of events, leading to deletions, insertions, and duplications. The
mutations can be replication-dependent, arising due to misincorporation of nucleotides during DNA
replication, or replication-independent, arising, for example, due to the repair of spontaneous DNA
damage, such as non-homologous end joining (Heidenreich, 2007). My research focuses on
spontaneous nucleotide substitution mutations (single nucleotide mutations). Using trio studies,
where the parents’ genomes are directly compared to the child’s genome, the human mutation rate is
estimated to be ~1x10
-8
mutations/nucleotide/generation (Campbell and Eichler, 2013). This estimate
is in line with indirect approaches which compare the human, chimpanzee, and gorilla genomes
(~1.5x10
-8
, Scally et al., 2012).
Around 76% of these spontaneous mutations are found to arise primarily in the male parent
germline (Campbell and Eichler, 2013). It has been shown that the human male germline has a
greater neutral mutation frequency compared to the female germline, called male bias, (Haldane,
1946; Taylor et al, 2006) likely due to the increased mitotic cell divisions in the male germline
compared to the female germline. Appropriately, the germline of older males has a higher mutation
8
frequency than the germline for younger males, presumably due to the increased number of mitotic
divisions with age (Penrose, 1955). It is important to note that this male bias is seen primarily with
nucleotide substitutions and not with other mutation types. My work examines the increased
mutation frequency seen in the male germline as they age and what determines that increase.
1.2 GAMATEOGENESIS
1.2.1 Spermatogenesis
In order to understand germline mutations, it is necessary to understand germline
development. Spermatogenesis is the process of male gamete development through mitosis, meiosis,
and spermiogenesis. Men produce hundreds of millions of spermatozoa, or male gametes, every day
which are vital in the transmission of genetic information and maintaining the species; thus it is
important that the complex process of spermatogenesis is executed efficiently and accurately.
In humans, the tunica albuginea is a thick layer of connective tissue that encapsulates the
testis (Holstein et al., 2003). The mediastinum testis, a connective tissue, radiates outwards toward
the tunica of the testis separating the testis into lobuli (Figure 1). There are around 300 lobuli in the
testis and each lobule contains 1-4 seminiferous tubules. The seminiferous epithelium, which lines
the seminiferous tubules, contains spermatogenic cells and Sertoli cells, the primary testis-specific
somatic cell type. Sertoli cells are large and found in between the spermatogenic cells and provide
physiological support, such as, hormones to the spermatogenic cells to aid in spermatogenesis
(Witsuba et al., 2007). The number of Sertoli cells correlates with sperm output numbers (Ehmcke et
al., 2006). In the interstitium between the seminiferous tubules, there are Leydig cells, a testis-
specific somatic cell that releases androgens in response to luteinizing hormone. These hormonal
signals are released from the Leydig cells and transduced through the Sertoli cells, since the
9
spermatogonial cells do not have the receptors for these hormones. There are other cell types in the
testis, including, peritubular cells, blood vessels and fibroblasts (Schlatt et al., 2014).
Figure 1. Testis Anatomy. (A) Cross-section of the human testis. Drawing of a paraffin section. ×
2.5 magnification. (B) The epididymis and the arrangement of the seminiferous tubules in the lobuli.
Semi-schematic drawing. (C) Cross section of a seminiferous tubule of a fertile man 32 years of age.
Drawing of a semi-thin cross section. × 300 magnification. Modified from Holstein et al.
Reproductive Biology and Endocrinology 2003 1:107. doi:10.1186/1477-7827-1-107.
Testis development begins in the developing embryo with the migration of primordial germ
cells to the genital ridge, the future site of the testis. There the germ cells organize into the
seminiferous tubules where the future Sertoli cells and spermatogonial cells in different stages of
developmental maturity reside (Drost, 1993; Holstein et al., 2003). The most immature
spermatogonial cells are Type A spermatogonia and Type B spermatogonia. After puberty, there are
two Type A spermatogonial stem cell (SSC) types: Adark (Ad) and Apale (Ap). Ap SSCs divide
10
mitotically to produce Type B spermatogonia while renewing themselves to produce a daughter Ap
SSC. Ad spermatogonia rarely divide and are a reserve stem cell population (Clermont 1966). It has
been shown, in monkeys, that Ad cells could become Ap cells to repopulate the stem cell pool after
injury or depletion of SSCs (van Alphen et al., 1988).
In humans, Type B spermatogonia, produce primary spermatocytes (Fig. 2). In the
seminiferous tubules, the primary spermatocytes begin meiosis, with an elongated prophase. During
this elongated prophase, homologous recombination occurs. Secondary spermatocytes form as a
result of the first meiotic division and they move from the basal membrane of the seminiferous
tubule to the adluminal section of the seminiferous tubule, placing them inside a blood-testis barrier
created by the Sertoli cells (Hecht, 1998). The second, faster (~6 hours) meiotic division occurs
which results in spermatids (De Kretser, 1998). Spermatids mature into spermatoza by condensing
their chromatin and elongating in a process called spermiogenesis.
Throughout spermatogenesis the dividing cells remain connected by a thin bridge (due to
incomplete cytokinesis) in the epithelium of the seminiferous tubules forming a syncytium (Schlatt
and Ehmcke, 2014; Amann 2008). Due to different cohorts of germ cells developing through
spermatogenesis, there are layers of cohorts at different stages of spermatogenesis throughout the
seminiferous tubule with the youngest at the basal layer and the more mature at the tubule lumen, in
general (Amann, 2008). Once the process is completed, mature spermatozoa are released from the
epithelium into the seminiferous tubule lumen in a complex process called spermiation after which
they continue through the duct into the epididymis (O’Donnell et al., 2011). From one
spermatogonia stem cell division, a maximum of 16 (or 32) sperm cells can be derived (Schlatt and
Ehmcke, 2006 and 2014). The process of spermatogenesis takes around 74 days to complete and
stem cell divisions happen every 16 days (Heller and Clermont, 1964; Amann 2008). Within the
11
seminiferous tubules, a human male will release over 25,000 sperm cells every minute (Amann,
2008).
1.2.2 Oogenesis
Oogenesis is the process of gamete development in females and like spermatogenesis differs
among species. Unlike males, human females do not continuously produce new gametes; they are
born with their total number of ova. Oogonia are diploid precursors to an ovum, or egg cell, and all
are created via mitosis in the developing fetus (Sanchez and Smitz, 2012). Beginning at the 15
th

week of gestation, oogonia undergo genome duplication and enter into the first meiotic division.
Late in prophase I (diplotene), meiosis is paused during which most oocytes will die (Baker et al.,
1970). Once the female enters puberty, of the surviving oogonia (primary oocytes), one is released
during a particular menstrual cycle and will continue through meiosis I, resulting in a secondary
oocyte and a polar body, which is a much smaller haploid cell and is not fertilized (Fig. 2). This
oocyte enters the fallopian tube during ovulation where it can be fertilized by a sperm cell. If it is
fertilized by a sperm, it proceeds through meiosis II, and becomes a zygote. After fertilization, the
haploid nucleus of the ovum and the haploid nucleus of the sperm combine to form a diploid nucleus
in the zygote. The menstrual cycle will continue to release an oocyte every month until the ovum
population is exhausted (Sanchez and Smitz 2012).
12
Figure 2. Diagram of cell divisions in male and female germline development. Male germline
cell divisions are dependent on the number of stem cell divisions and thus the age of the male, while
the total number of female germline cell divisions is constant regardless of age. Some cell lineages
end in apoptosis. Ap, Apale; B, B spermatogonia; 1°, primary spermatocyte (preleptotene); 2°,
secondary spermatocyte. Based on Heller and Clermont, 1963; Drost and Lee, 1995.

13
Male germ cells have more divisions throughout gametogenesis compared to females, and
the increase in number of DNA replications and cell divisions could lead to an increase in mutations
introduced over time (Fig. 2). Unlike males, in females all of the stem cell divisions happen before
birth so the number of replication-dependent mutations present in the germline (except for mutations
that arise by chemical means, for example, deaminations of cytosine) is independent of age.
1.3 MALE BIAS AND THE PATERNAL AGE EFFECT
The ratio, α, is the ratio of the mutation rate of males to females and, in humans, there is a
male bias with males having, on average, a higher mutation rate than females (Taylor et al., 2006).
The ratio, α, was first calculated indirectly by comparing the mutation rates of a DNA segment in the
X chromosome (which spends about 66% more time in the female cell environment) to a segment in
an autosome (Arnheim and Calabrese, 2009; Makova and Li, 2002). Given that the autosomes spend
the same amount of evolutionary time in both the female and male environments and the X
chromosome spends 1/3 less time in the “higher-mutation-rate” background of the male
environment, then the X chromosome should have a lower mutation frequency than the autosomes
and the mutation bias between the sexes can be deduced from this comparison. Note that these
estimates assume that replication-dependent mutations impact the male bias more than replication-
independent mutations.
To estimate the number of germ cell division in an adult man of any age and knowing that a
post puberty human SSC divides every 16 days to give rise to an Ap cell, then there are around 23
divisions per year after the ~28 germ cell divisions before puberty (Drost and Lee, 1995). We can
estimate the number of SSC divisions (N) of a man age, Ma, as: N = 28 + 23(Ma – Map), where Map
is the age at puberty, ~13 years old (Lee 1980). Accordingly, a 26 year old man will have sperm that
originated from a stem cell that has a history of ~327 cell divisions since zygote formation.
14
Compared to a man, the ovum of a 26 year old female (average age of first time mother; Martin et
al., 2015) has only gone through ~31 cell divisions (Drost and Lee, 1995). Calculation of α would be
327/31 = 10.5. For a 65 year old, α would be 1,224/31 = 39.5, showing that α increases dramatically
as a man ages. The difference in α values as a man ages supports that the age of the father affects the
mutation rate. In humans, α has been estimated by many groups studying DNA evolution to be
between 2 and 8 (Taylor et al., 2006; Makova and Li, 2002; Nachman and Crowell, 2000). These
calculations assume that the evolutionary age of a reproducing man is between 14.5 and 22.6 years
old.
New mutations introduced during mitotic stem cell divisions increases the number of
mutations that accumulate over time and is dependent on the age of the male. The probability of an
offspring that inherits a new mutation increases as a man ages, and is called the paternal age effect
(PAE) (Penrose, 1955; Crow, 2000). Early studies by Wilhelm Weinberg (1912) showed that there
was a higher incidence of the disease achondroplasia in the last offspring of affected families and, in
1955, Lionel Penrose attributed this phenomenon to the advancing age of the father (Goriely and
Wilkie, 2012). In 1987, Risch et al. confirmed the PAE by analyzing the parental age data of 17
different autosomal-dominant disorders in the population. However, the increase in mutation rate
with advance paternal age is greater, for many of them, than would be expected if it were simply
replication-dependent (Crow 2000), suggesting that other processes may be contributing to the
increase in frequency.
The increasing birth rates in older couples highlights the significance of understanding the
PAE mechanism. Currently, childbearing after the age of 40 is associated with a decrease in fertility
as well as unfavorable pregnancy outcomes. With birth rates increasing for pregnant couples aged
15
35-39 and more than doubling for pregnant couples aged 40-44, it is important to understand the
increase in inherited mutations with increased paternal age.
1.4 PAST RESEARCH AND RAMP DISEASES
Some diseases exhibit a surprisingly high incidence in the population due to new mutations
entering the population every generation. Our laboratory has studied four well known examples: the
first is the most common form of disproportionate short stature (achondroplasia), the second
involves craniofacial abnormalities (Apert syndrome), the third causes childhood thyroid cancer
(multiple endocrine neoplasia type 2B), and the last causes Noonan syndrome (Shinde et al., 2013;
Yoon et al., 2009; Choi et al., 2012; Yoon et al., 2013). Noonan syndrome (NS) is characterized by
short stature, heart defects, intellectual disability, craniofacial abnormalities and a predisposition to
certain cancers (Noonan and Ehmke, 1963; Allanson, 1987). The aforementioned diseases have
features in common with one another (and to some other genetic diseases) that we call RAMP
features: 1) Recurrent de novo mutation at the same nucleotide, 2) Autosomal dominant
transmission, 3) Male-biased de novo mutations and 4) a Paternal age effect.
Most genetic disease mutations are extremely rare yet sporadic (new) cases of the four
diseases mentioned above occur 100-1,000 times more frequently than expected (Allanson, 1987;
Gardner, 1977; Brauckhoff et al., 2004; Cohen et al., 1992). The classical explanation for this high
incidence is that the mutation rate per cell division at the casual nucleotide(s) is, for some unknown
reason, exceptionally high (hotspot model). A newer selection model, however, argues that these
specific mutations confer a proliferative advantage to the spermatogonial stem cell (SSC) over the
wildtype stem cell. In order to test which of these models can explain the increased frequency, we
look to the distribution of mutated SSCs in the testes of unaffected men. Assuming that all SSCs
have an equal chance of being mutated at the disease site and that SSCs have a uniform distribution
16
throughout the testis (via the uniform distribution of the seminiferous tubules), the mutated cells
should also be evenly distributed throughout the testis if the hot spot model were true for a given
disease site (Fig.3). However, if the selection model were the mechanism responsible for the
increased frequency then, the mutated SSCs would be clustered, similar to a “mini-tumor”, since the
mutation would give the SSC a proliferative advantage (Fig. 3). Our lab (Arnheim and Calabrese)
has developed methods, statistical tests, and computational models to test whether the mutation
distribution we see in the testis agrees with the hotspot model or the selection model. We analyzed
the distribution of NS mutations in the testes of men of various ages to understand the increased
frequency of NS in the population.

Hotspot Model Selection Model
Figure 3. Hypothetical distribution of mutated SSCs based on the hotspot model or the
selection model. The distribution of cells containing mutations (red dot) at a single nucleotide site in
the testis according to the hot spot model; a uniform distribution on the tubules is predicted. The
clusters shown for the selection model could be explained by a few rare mutations which provide the
mutated self-renewing SSC with a proliferative selective advantage. Figure used from Ref. Arnheim
and Calabrese, 2009 with permission.
1.4.1 Noonan Syndrome
17
Noonan Syndrome (NS) is one of the most common Mendelian diseases with a 1/2,000 live
birth frequency. Arnheim and Calabrese estimated that the most common NS mutation (PTPN11 c.
922A>G) has a de novo mutation frequency 2,400 fold greater than the genome average mutation
frequency (Yoon et al., 2013). Mutations in genes, such as, SOS1, KRAS, RAF1, BRAF, and other
genes in the RAS/Mitogen-Activated Protein kinase (RAS/MAPK) pathway (Fig. 4) have all been
implicated in NS, but about 50% of all (sporadic and inherited) NS cases are due to mutations in the
protein tyrosine phosphatase, non-receptor type 11 (PTPN11) gene which codes for the SHP2
protein (Tartaglia et al., 2005 and 2010). SHP2 is a phosphatase which functions as a signaling
protein in the RAS/MAPK pathway, which is responsible for developmental processes, including
cell differentiation, proliferation, and apoptosis. SHP2 is activated by its N-SH2 domain binding to
phosphotyrosyl (pTyr) residues on interacting partners (Fig. 5). The protein tyrosine phosphatase
(PTP) domain of the protein is then no longer inhibited by the Src homology 2 (SH2) domain that
normally blocks the phosphatase active site (Hof et al., 1998).
18
Noonan Syndrome
LEOPARD
JMML
Figure 4. Diseases associated with the Ras-MAPK pathway. Noonan Syndrome is caused by
mutations in SHP2 (PTPN11), SOS1, KRAS-GTP, and RAF. JMML (Juvenile Myelomonocytic
Leukemia) is caused by mutations in SHP2, Neurofibromin-1, NRAS, and KRAS. Ninety-three
percent of individuals with LEOPARD syndrome have mutations in SHP2 or RAF.
Some NS mutations affect residues important to the stabilization of the SHP2 protein in the
inactive state. These mutations are predicted to increase the activity of the SHP2 protein due to the
shift from the inactive, closed state to more constitutively open state (Tartaglia et al., 2002). The
majority of NS mutations cluster in exons 3 and 8 of PTPN11. Exon 8 contains the most common
NS mutation, c. 922 A>G (N308D), within the PTP domain of the protein. Although the N308D
mutation does not directly affect the PTP/SHP2 interaction, it is involved in maintaining the overall
PTP structure and was found to facilitate interactions with substrates needed to open the protein to
the active state (Tartaglia et al., 2006; Darian et al., 2011; Lauriol et al., 2015).
19
Figure 5. Structure of SHP2 and activation of PTP after phosphotyrosyl-peptide binding. SHP2
contains the PTP catalytic domain and two SH2 domains, C-SH2 and N-SH2. The protein is found in
the inactive state when the phosphotyrosyl-peptides (pY) are not present. When the peptides are
present the SH2 domains unfold and expose the active site of the SHP2 protein allowing for
substrate to bind to the catalytic domain.
Exon 3 contains the N-SH2 domain which interacts with the PTP domain to maintain the
protein in the inactive conformation. Most NS mutations in exons 3 and 8 were found to interfere
with the N-SH2 and PTP interaction and cause the SHP2 protein to remain in the more open active
state (Tartaglia et al., 2002; Martinelli et al., 2008).
Many researchers have found that mutations affecting different amino acids of the SHP2
protein affect the protein activation level (Bentires-Alj et al., 2004, Martinelli et al., 2008, Yu et al.,
2014). In vitro assays measured the phosphatase activity of mutated SHP2 compared to the wild-type
when SH2 domain-binding pTyr peptides were added. The biochemical assay revealed that with the
wild-type SHP2 the activity was increased by 2-fold compared to when the pTyr peptide was not
added. The exon 8 germline N308D mutation increased the activation level of the protein to around
4-fold higher than basal levels, while the somatic exon 3 mutations A72V and E76K (only associated
with cancers (hematological, lung, etc.) and not with NS) greatly increased in the activation level to
around 5-fold and 8-fold higher, respectively, without the addition of the pTyr peptide and 9-fold
higher for both mutations with the addition of the pTyr peptide (Tartaglia et al, 2006). These results
show that different mutations affect the activity level of the SHP2 protein and these different activity
20
levels determine the severity of the disease phenotype. This will be discussed further in the
Discussion section.
1.4.2 Past Research on RAMP mutations using the PAP technique
Using a highly-sensitive PCR technique, pyrophosphorolysis-activated polymerization (PAP,
Liu and Sommer, 2004), the spatial distribution of the most common Noonan syndrome mutation
(PTPN11 c.922A>G) was analyzed (Fig. 6; Yoon et al., 2013). The PAP assay allows for single
nucleotide resolution and can detect a single nucleotide mutation with a false positive rate of 9.3 x
10
-7
. Using the PAP protocol, and the associated experimental, statistical, and computational tests
developed in our lab, we found, surprisingly, that the high incidence for this NS mutation was not
due to it being a hotspot. Rather, the mutations appeared to occur extremely rarely but, once a stem
cell in a man’s testis experiences one of these disease mutations, it proliferates producing a “mini-
tumor” in the testis. This “tumor” is not malignant but due to their increased numbers they produce
an increase in the number of sperm carrying the mutation.
For the younger donors (19-23 years old), there was a low average mutation frequency
(1.06x10
-5
) for all of the pieces, while for the middle aged donors (36-62 years old), there was a high
average mutation frequency (1.81x10
-3
). For the middle aged donors, the NS mutation showed a
cluster-like distribution in the testes, suggesting that this NS mutation gives the mutated SSC a
selective advantage (Fig. 6, Yoon et al., 2013). Surprisingly, there was a varied mutation frequency
for the older donors (68-80 years old), some were similar to the younger donors with a low mutation
frequency and others were similar to the middle aged donors with clusters of testes pieces with a
higher mutation frequency. This varying frequency in the older individuals was also seen for the
MEN2B mutations and was attributed to age-dependent cell death (Choi et al., 2012; Yoon et al.,
21
2013; Johnson, 1986). We found that the high incidence of the most common NS mutation in the
testes fits the selection model, but the hot spot model was rejected.
Figure 6. Distribution of the most common Noonan Syndrome mutation (PTPN11 c.922 A>G)
in the testes of 15 unaffected donors of various ages. The frequency of the NS mutant in each
testis piece. The testis is dissected into 6 slices and 32 pieces per slice for a total of 192 testis pieces.
The left column shows data from younger donors, aged 19-23, with a low mutation frequency. The
data in the middle column is from middle aged donors, aged 36-62, with a clustered mutation
distribution. The right column shows data from older donors, age 68-80. The frequency of the NS
mutation in each testis piece is indicated by color, the scale is shown in the heat map. Figure used
from Ref. Yoon et al., 2013 with permission.
22
The PTPN11 gene is known to have other DNA mutations that cause Noonan Syndrome
which are also highly recurrent (Yoon et al., 2013; Tartaglia and Gelb, 2008). Do all of these disease
mutations provide a selective advantage to the stem cell? And if so, will the sum proliferative
advantage of all of the PTPN11-associated NS mutations explain the high NS incidence in the
population? Also, given that different mutations affect the protein in different ways (constitutively
active, reduced function, etc.), do these activity level changes in the protein coincide with the level
of selection given to the SSCs in the testis? My dissertation proposal examines the distribution of
many additional mutations in exon 3 of the PTPN11 gene in the testes of unaffected men to help
answer the above questions.
1.5 NEXT GENERATION SEQUENCING
Next generation sequencing (NGS), or massively parallel sequencing, is a technology that
allows the interrogation of tens of millions of nucleotides at one time scaling up the throughput of
the traditional method of sequencing, Sanger Sequencing. NGS improves on Sanger sequencing by
creating millions of mini-reactions that are sequenced separately but in parallel. In 2005, Margulies
et al. described a technique for separating the mini-reactions into tiny individual wells which
allowed for the interrogation of over 25 million bases in a four-hour run (Margulies et al., 2005). The
technique (454 sequencing) is a method that binds individual DNA molecules onto beads which are
amplified in isolation within a droplet of a PCR-reaction-mixture in an oil emulsion. Sequencing of
each individual bead is performed using “pyrosequencing”, which releases a pyrophosphate
molecule and results in a flash of light when the correct nucleotide is incorporated (Nyren et al.,
1993).
The NGS platform that we are using is called Illumina sequencing. The Illumina sequencing
approach does not amplify in droplets but uses a solid support for a “bridge amplification” of each
23
single DNA molecule to create tens to hundreds of millions of clusters each composed of identical
DNA strands from different regions of the genome. The sequencing step itself is similar to Sanger
sequencing where four nucleotides, simultaneously, are added to the oligo-primed clusters, and
differentiating each incorporated nucleotide with a different fluorescent dye. There have been a
number of iterations of NGS technologies over the years which have improved the technology as
well as reduced its cost. The technology is now able to sequence a full human genome for $1000 in
about 3-4 days (Hayden, 2014).
1.6 SAFE SEQUENCE SYSTEM OVERVIEW
We previously used the PAP (pyrophosphorolysis-activated polymerization) protocol to
analyze the distribution of mutated SSCs in the testes of unaffected men. However, the PAP protocol
is difficult to optimize for different mutation nucleotides and is only able to interrogate one
nucleotide at a time. Unlike the PAP protocol, next generation sequencing (NGS) allows us to
interrogate more than one nucleotide site per reaction allowing us to investigate other disease
causing mutations in the same DNA fragment as well as control nucleotides within one run.
Although NGS has its benefits, there are some drawbacks, such as its high error rate of 0.1%.
The high NGS error rate is due, in part, to the polymerase error during PCR amplification, but
primarily by the incorporation errors during sequencing. Since the mutated SSCs are a very small
fraction of the overall cell count in the testis (1.2 x 10
9
total SSCs; V ogel and Motulsky, 2010), to
accurately detect these rare mutations, we need to have a very low false positive rate. We have
reduced the error rate by modifying a method originally developed in the V ogelstein lab, which
incorporates a unique identifier onto each original template DNA strand in order to track it through
PCR; the method is called the Safe Sequencing System (Kinde et al., 2011; Eboreime et al., 2016 in
submission).
24
The Safe-Sequence System (SSS) utilizes NGS but with the benefit of increased sensitivity
by the addition of a Unique IDentifier (UID). The UID sequence is a stretch of random nucleotides
which allows for each target DNA molecule to be uniquely tagged. Similar to a barcode, which
differs between samples, the UID is able to differentiate between original DNA molecules. One way
that we modified the original method was by putting a 14-mer UID on the forward primer and a 6-
mer UID on the reverse primer, totaling 20 nucleotide positions for the UID. Since there are 4
nucleotides, 4
20
possibilities would produce over 10
12
UID sequences. This reduces the possibility of
two DNA molecules getting tagged with the same UID combination. In Results, I will discuss how
we optimized this protocol and how we used it to detect and quantitate rare germline mutations.
25
2 METHODS
2.1 SOURCE OF TESTIS DNA
Testes from a 68 year old donor (60891), a 65 year old donor (60649), and 21 year old
(63878) for the PTPN11 exon 3 analysis and testis from a 36 year old donor (#62923) for the
PTPN11 exon 8 analysis were supplied by the National Disease Research Interchange (NDRI,
Philadelphia, PA) with the approval of the Institutional Review Board of the University of Southern
California. No donors were accepted if they had been treated with drugs known to interfere with
normal spermatogenesis. All samples were frozen within 10-12 h after death (Eboreime et al., 2016
in submission).
2.2 TESTIS DISSECTION/ DNA PURIFICATION AND DNA QUANTITATION
The testis dissection scheme has previously been described in detail, see Figure 1 in (Qin et
al., 2007). The epididymis was removed from the testis. Each testis was fixed in 70% ethanol at 4°C
for about 3 days, cut into 6 slices and each slice further divided into 32 pieces of approximately
equal size. Testis DNA was extracted from each testis piece using the Puregene DNA purification kit
(Gentra Systems). Real-time PCR (Bio-Rad Opticon 2 instrument) was used to determine the
concentration of DNA in all of the samples. On average, each testis piece contained DNA from about
2.5x10
7
cells (range 1.5x10
7
to 4.1x10
7
). This estimate was made using the average number of
genomes per testis piece and data on the number of germline and non-germline cells found in
individual human testes (Zhengwei et al., 1998) after accounting for the ploidy of pre- and post-
meiotic germline cells as well as non-germline cells (Sertoli cells) (Eboreime et al., 2016 in
submission).
26
2.3 MODIFIED SAFE SEQUENCE SYSTEM
A modified version of the Safe Sequence System (SSS) was used to tag and amplify exon 3
of the PTPN11 gene. The protocol is similar to that found in Eboreime et al., 2016 (Fig. 7). A million
human genomes (~3 µg) from each testis piece were used for the analysis. The forward 1
st
round
PCR primer contained a portion of the Illumina forward sequencing primer region, a universal
identifier (UID) of 14 bps, and a target specific sequence of 21bp that corresponds to a region
slightly upstream of exon 3 of the PTPN11 gene. The reverse 1
st
round PCR primer contained a
portion of the Illumina reverse sequencing primer region, a UID of 6bp, an 8-mer barcode to
designate the testis piece, and a different target specific sequence of 24bp that corresponds to a
region towards the 3’ region of exon 3 of the PTPN11 gene. The target sequence is 127 bps, not
including the primer sequences. The 1
st
round primer pair was used for two cycles of PCR in order to
introduce the UID and the universal sequence to the DNA molecule. Each of the 10 reactions (per
testis piece) contained 1x10
5
genomes of purified DNA, 1x GC Phusion Buffer (1.5mM MgCl2),
300nM of each of the first round primers, 80µM dNTP, Phusion Hot Start High Fidelity DNA
Polymerase (0.02U/µl), and 0.26x SYBR green, 2.6µM ROX solution in a 45µL reaction. After a
98°C 1 min initial denaturation, two cycles were run (98°C 30sec, 62°C 4min, 72°C 1min) followed
by 72°C for 5 minutes in a MJ Research Opticon 2 QPCR instrument. The 1
st
round PCR primers
were removed by adding 1.0x volume of homemade magnetic beads (Using Sera-Mag SpeedBeads;
ThermoScientific 6515-2105-050250; Rohland and Reich, 2012). After 5 minutes at room
temperature, the plate was placed on the magnet for 2 minutes, washed twice with 70% ethanol, and
allowed to dry for 9-10 minutes in a biosafety cabinet.
27
The DNA on the dried magnetic beads was resuspended in the second round PCR buffer
which was identical to the first round buffer except the first round primers were replaced by 550nM
of each second round primer also in a 45 µL reaction. Twenty-eight additional cycles of PCR were
performed (98°C 1 min initial denaturation, 98°C 10sec, 72°C 15 sec ending with 72°C for 5 min).
Both second round primers contained sequences complementary to the 5’ most portion of the first
round PCR product at the 3’ end. The 5’ end of the second round primers carried the sequences
needed for binding to the Illumina flow cell.
Following the second round, 0.8x PEG 8000 (20%) dissolved in 2.5M NaCl (required for the
bead-DNA binding step) was added to each reaction, allowed to sit at room temperature for 5
minutes, exposed to the magnet for 2 minutes, washed twice with 70% ethanol, and allowed to dry
for 5-6 minutes. The DNA was eluted in 10µL of water and the expected 317bp final product from
each of the ten reactions for each piece were pooled. For the final library, all 192 pooled PCR
products for the testis were combined in equal amounts based on Nanodrop analysis. The pooled and
concentrated (SpeedVac) sample was evaluated (Bioanalyzer 2100, Agilent Technologies) to
quantify the final amount of DNA in the library. The DNA was then loaded on a NextSeq 500 and
150bp paired-end sequencing was performed. Each testis had a similar number of UID families,
totaling 40 million UID families with an average family size of 7 for a Utility Ratio of ~10
(discussed below) for all three testes.
28
Figure 7. Safe Sequence System schematic. Note that this is a representation of 2 template
molecules out of a many millions. The numbered primers contain a gene specific sequence on the 3’
end (black), the unique identifier, or UID (numbered box), and a universal amplification sequence on
the 5’ end (magenta). Two cycles of PCR are performed in order to produce uniquely tagged, double
stranded DNA fragment from the template molecule. The unused UIDs are removed and the two new
universal primers are introduced (blue and magenta). These new primers are complementary to the
tails of the UID primers so it only amplifies products in the first round which acquire a UID. The
result of the Universal Primer amplification are UID families all originating from the same template
molecule.
Example of SSS Noonan Reverse Primer:
SSS Noonan E3 Reverse Barcode #65:
CGGCATTCCTGCTGAACCGCTCTTCCGATCTACAGCAATNNNNNNTCTTTTAATTGCCCGTGATGTTCC
SSS Noonan Forward Primer:
CGACGCTCTTCCGATCTNNNNNNNNNNNNNNTATTTGTCCCCTTGCCTCCCT
29
Second Round Forward Primer:
AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT*C*T
Second Round Reverse Primer:
CAAGCAGAAGACGGCATACGAGATCGGTCTCGGCATTCCTGCTGAACCGCTCTTCCG*A*T
Color Key:
FLOWCELL ADAPTER
SEQUENCING PRIMER/ UNIVERSAL SEQUENCE
UNIVERSAL IDENTIFIER
BARCODE
GENE SPECIFIC PRIMERS
2.4 BARCODES
The barcodes are added to differentiate between the different testis pieces. The 8-mer barcode
is on the reverse primer, as seen in the example primer shown above. The barcode was place on the
5’ side of the UID, right next to the sequencing primer. We selected the barcode sequences for the
different pieces such that if there is a mistake at any one site in the barcode then we can infer the
correct barcode, and if there is a mistake at two sites then we will know there has been a mistake and
it will not be incorrectly counted as a different barcode.
2.5 QUANTITATIVE METHODS
We wrote Perl scripts to analyze the raw sequencing data. We only considered reads with
fewer than 5% of bases different from the target sequences (maximum difference 6 bp). To measure
the sequencing error, we compared sites where reads #1 and #2 overlapped. If these two reads
differed and one of the reads agreed with the reference, we assumed the other read was a sequencing
mistake (the number of cases where both reads disagreed with the reference was too few to affect the
sequencing error frequencies). To form UIDs, we only considered those reads where all the sites in
the barcode and UID sequence had quality scores of 32 or greater. We clustered reads with the same
barcode and UID into UID families. We only considered families with at least 3 paired reads with the
same UID. We only considered sites in these families with at least 3 reads with quality scores 32 or
30
greater (reads with quality scores less than 32 at a site were not considered for that particular site). A
super-mutant is when 95% or more of the reads in a UID family agree with each other and disagree
with the reference at a site.
31
3. RESULTS
3.1 MODIFIED SAFE SEQUENCE SYSTEM
The modified SSS protocol was developed to reduce the error rate associated with NGS
sequencing. Figure 8 shows how the modified SSS protocol is able to differentiate between true
mutations and sequencing errors. Genomic DNA is shown in blue (Fig. 8B) with a true mutation (red
dot) in the gene of interest (gray box). The primers for the first round of PCR contain a universal
sequence, the UID, and a target-specific sequence (Fig. 8A). The method begins with two cycles of
PCR (Round #1) for the addition of the UIDs onto the copies of the template DNA molecules (Fig.
8C). During the first round, a PCR error could be made (yellow dot). In the following round of PCR
(Round #2), the second round primers (Fig. 8E) are targeted to the universal sequence in order to
amplify products that received a UID in the first round. The second round creates UID families
which allow us to differentiate between true germline mutations (red dot) and errors during PCR
Round 2 and sequencing errors (black star). However, errors during Round 1 (yellow dot) will be
amplified creating UID families which all have the same UID and, unfortunately, these Round 1
errors cannot be distinguished from true mutations.
32
Figure 8. Schematic of modified Safe Sequence System Strategy with true mutation versus
PCR error. In the first round, UIDs are added on to nascent genomic DNA strands via a two-cycle
PCR with primers A and B. A. The primers contain a gene specific region (GENEF, GENER), a
barcode (BCx), a UID sequence (UIDi, UIDj), and a sequencing primer (SEQ). Only the reverse
primer contains a barcode (BC). B. Duplex genomic DNA fragment carrying a gene target with a
true mutation (red dot). C. The two new daughter strands of the first denaturation, primer
hybridization and primer extension. The A42 and B28 represent specific UIDs on the forward and
reverse primers, respectively (in blue). The blue arrow indicates the direction of the primer extension
that includes the specific target region (black dashes). The A42 extension may introduce a new PCR
33
error (yellow dot). The green overhang represents the sequencing primer that will be used as the
template in the second round of 28 cycles of PCR. D. The final product of the first round of PCR
with the 5’ overhang sequence necessary for amplification in the second round (green). E. Primers C
and D are used to amplify the products made in the first round of PCR. The primers contain the
Illumina Sequencing primer (SEQ), which is complementary to the Sequencing primer of the first
round, and the Flowcell Adapter (FC). F. A PCR error may be introduced during the first two cycles
of PCR in the first round (yellow dot) or a true mutation (red dot) may have been present in the
original DNA duplex. The amplification of the first round product by Primer C and D (green arrow).
These primers are used for 28 cycles of amplification to create UID families. H. Representative UID
families which are all generated from the same starting DNA molecule. After analysis, the mutations
that are not seen in 95% of the UID family are removed. True mutations or mutations made in the
first two cycles of SSS are not removed by analysis and all are quantified as mutations.

3.1.1 Sequencing Error
In order to test the efficacy of the SSS protocol for the detection of rare germline mutations,
we needed to know the background false mutation frequency at different base pairs. We studied
regions from three different human genes, a previously studied region of the PTPN11 gene, MECP2,
and FGFR3 (Eboreime et al., 2016 in submission). For the PTPN11 and MECP2 SSS experiments,
the primer pairs directed to the gene specific region were close enough to have overlapping paired-
end sequencing reads. We looked at the mutation frequencies associated with the overlapping sites
for the MECP2 (23 nucleotides) and PTPN11 (32 nucleotides) libraries which helped us to identify
the sequencing error associated with the SSS protocol by estimating the disagreement between the
reads when the same molecule is sequenced from opposite ends in a paired end run. Nine PTPN11
libraries were made from 3 testes pieces from a 36 year old donor (#62923), while thirty-two
MECP2 libraries were made from thirty-two testis pieces from a 68 year old donor (#60891). We
compared our mutation frequencies to the quality scores defined by Illumina (Table 1) and found
that there was an inverse relationship between the two values. Based on these values, we decided to
34
only consider the reads where the nucleotide sites in the barcode and the UID have a quality score of
32 or greater.
Table 1. Sequencing error compared to the Illumina-derived quality score. Sequencing error
estimated by measuring the disagreement between nucleotides in the overlap region of paired end
reads compared to Illumina based Quality scores (Eboreime et al., 2016 in submission).
Quality score Sequence error
1 1.9 x 10
-3
5 6.1 x 10
-4
14 1.3 x 10
-4
21 5.9 x 10
-5
26 2.3 x 10
-5
32 5.3 x 10
-6
36 1.0 x 10
-6
39 3.9 x 10
-7
3.1.2 Supermutants and Error Rate
A Unique Identifier, or UID, labels PCR products that arose from the same original template
DNA molecule. A UID family, we have designated, is at least 3 reads with the same UID. We have
defined a supermutant site as a site where 95% or more of the reads in the UID family (with a quality
score of at least 32) agree with each other and also disagree with the reference allele. These
supermutants are considered real mutations present in the original DNA strand. However, these
supermutants can be created also in the first two rounds of PCR in the SSS protocol by PCR error, or
by sequencing error.
We can calculate the probability of falsely identifying a supermutant due to sequencing error.
For a quality score of 32 or higher, the sequencing error per site is 5.3x10
-6
(Table 1). For the
minimum UID family size, all three reads will have to have the same sequencing error at the same
nucleotide site, (5.3x10
-6
)
3
x (1/3)
2
= 1.7x10
-17
. Given this incredibly low probability, the probability
of an error due to the Phusion polymerase is much more likely (4.4x10
-7
; New England Biolabs).
Thus, the minimum error rate of the SSS protocol is limited by the error rate of the polymerase used
35
and after two cycles of PCR, there is a minimum PCR error rate of 4.4x10
-7
x 2 = 8.8x10
-7
. This
error rate is much lower than that of traditional NGS experiments (1x10
-3
) and is low enough to
detect rare germline mutation rates.
3.1.3 Germline Mutation Frequency Comparison of PAP and SSS
Using both PAP and SSS, we sequenced a region in exon 8 of the PTPN11 gene that spans
the most common Noonan mutation c.922A>G in the testes of unaffected men to compare the
mutation frequencies from both protocols. We found previously, with the PAP protocol, that there
were three adjacent testis pieces in a 36 year old donor (#62923) that had wide-ranging mutation
frequencies for this nucleotide site (Yoon et al., 2013). Using PAP, we analyzed testis piece #18 of
slice 5 of testis donor #62923 and found it had a very low mutation frequency of 0 mutants in
5,977,500 genomes studied. Analyzing this same piece with SSS, we found the mutation frequency
to be 0 mutants out of 49,655 UID families (95% confidence interval (ci), 0 – 7.7x10
-5
). For piece
#19 of the same slice, using PAP, we measured a mutation frequency of 3.7x10
-4
(close to the
average mutation frequency for the whole testis). Using SSS, we measured a mutation frequency of
1.1x10
-4
(95% ci, 7.9x10
-5
– 1.6x10
-4
). For the third piece, #20, the mutation frequency was the
highest one measured using PAP at 2.2x10
-2
, and using SSS, the frequency was measured at 7.3x10
-3

(95% ci, 7.0x10
-3
– 7.7x10
-3
). The frequencies measured by SSS and PAP align significantly or are
very similar even across a variety of mutation frequencies. SSS is a faster and more efficient
protocol than PAP and is able to accurately detect a low germline mutation frequency at hundreds of
nucleotide sites, not just one site (Eboreime et al., 2016 in submission).
3.1.4 Mutation Type Frequency
Using SSS, we were also able to investigate the mutation frequency of 87 nucleotides in exon
8 of PTPN11 in addition to the c.922 A>G NS mutation site. We found an unexpectedly high average
36
mutation frequency for these other sites (3.7x10
-5
). Upon further investigation, we found that the
mutation frequency greatly depended on the reference nucleotide; for C and G the mutation
frequency was 8.0x10
-5
, and for A and T the mutation frequency was 8.7x10
-6
, almost a whole order
of magnitude lower (Fig. 9). CpG sites, unsurprisingly, have an even higher mutation frequency than
the other C and G reference nucleotides (Fig. 9, green dots). We measured the mutation frequency as
the sum of all the mutations at that particular site, in other words, for a C reference nucleotide, the
mutation frequency is the sum of the C>A, C>T, and C>G frequencies.
Figure 9. PTPN11 mutation frequencies color-coded by reference base. Each dot represents an
average of frequencies over the nine libraries from three testis pieces for eighty-seven control
nucleotide sites, the bars represent the 95% confidence intervals. Position c. 922 is not shown due to
the extremely high mutation frequency which would have distorted the scale. The color denotes the
reference allele: Red is an A or T, Blue is a C or G, Green indicates a CpG (Eboreime et al., 2016 in
submission).
Even further, we found that C>A/G>T and C>T/G>A (only G>T and C>T will be used for
ease) mutation types were seen at much higher frequencies than the other mutation types (Table 2).
37
Table 2. Mutation frequencies for each mutation type.
Mutation type Mutation frequency
A>C/T>G 4.4 x 10
-7
C>G/G>C 1.1 x 10
-6
A>T/T>A 1.4 x 10
-6
A>G/T>C 3.3 x 10
-6
C>T/G>A (non-CpG) 1.7 x 10
-5
G>T/C>A 2.2 x 10
-5
C>T/G>A (CpG) 7.4 x 10
-5
The frequencies shown are a combination of the mutation frequencies for PTPN11, MECP2, and
FGFR3 (Eboreime et al., 2016 in submission).
The C>T and G>T mutation types would seem to correspond to DNA damage associated with the
common deamination of cytosine to uracil, or 5’methylcytosine to thymine, and the oxidation of
guanine to 8-oxoguanine leading to 8-oxoguanine mispairing to adenine during replication. This
DNA damage could have presented either during DNA preparation or during library preparation.
(See above; Eboreime et al., 2016 in submission).
3.1.5 Optimal UID Family Size and Utility Ratios
We used the SSS method to detect the mutation frequency of a region of MECP2 and FGFR3
in testis donor #60891, a 68 year old unaffected male (Eboreime et al., 2016 in submission). We
assayed a 171bp region of the FGFR3 gene and a 99bp region of the MECP2 gene in 32 pieces of
the testis (one slice). We can use this data to determine the optimal number of initial DNA molecules
in the reaction necessary to have the optimal UID family size.
Quality reads, or paired end reads with a quality score of at least 32, are wasted in two ways:
super large UID families or too small UID families with less than 3 members. In order to understand
this discrepancy in UID family size, we developed the Utility Ratio which is the ratio of quality
reads to the number of UID families. For example, the utility ratio for the PTPN11 library is 6.4x10
7
38
quality reads/6.1x10
5
UID families = 106. The average family size for the PTPN11 experiment is 105
and 18% of families have 100+ members and only 1% of reads have less than 3 members. Also, the
largest family size for the PTPN11 experiment is 50,000 members. The utility ratio of ~100 shows
that the PTPN11 experiment did not make efficient use of its reads, which is evident in the average
family size being much higher than 3. However, for the MECP2 and FGFR3 experiments, the utility
ratios are 10.4 and 11.4, respectively. The average UID family size for MECP2 and FGFR3 are 7.1
and 8.1, respectively, and less than 0.1% of the families have 100+ members with the largest family
size having 237 members. However, 32% of MECP2 and 30% of FGFR3 reads are wasted due to
them having less than 3 members per UID family. Comparing the utility ratios, the MECP2 and
FGFR3 experiments made more efficient use of their reads. We estimate the optimal obtainable
utility ratio to be 5.15, which is calculated in part by comparing the number of quality reads to the
number of initial molecules in each reaction (details in Eboreime et al., 2016 in submission).
A reason for the different utility ratios between the PTPN11 experiment and the MECP2 and
FGFR3 experiments is most likely the number of initial DNA molecules used in each of the
experiments since the number of initial molecules limits the number of UID families produced. Due
to the PTPN11 experiment having fewer initial molecules, a smaller number of UID families were
produced. The PTPN11, MECP2, and FGFR3 experiments had similar number of reads, but because
PTPN11 had fewer UID families compared to the other experiments, the additional reads contributed
to the size of the UID family, giving the PTPN11 experiment fewer UID families with much larger
UID family sizes. Inversely, the MECP2 and FGFR3 experiments had a larger number of starting
molecules so they produced a large number of UID families with less members in each family. The
ratio of quality reads to initial molecules was 27.8 for the PTPN11 experiment, and 0.24 and 0.88 for
the MECP2 and FGFR3 experiments, respectively. We have calculated that in order to achieve the
39
optimal utility ratio the ratio of quality reads to initial molecules should be 3.3 (details in Eboreime
et al., 2016 in submission).
3.2 MOLECULAR ANATOMY OF NOONAN MUTATIONS IN THE TESTES OF
UNAFFECTED MEN
We ask if the high frequency of sporadic cases of NS is due to the hot spot model (an
unusually high base substitution rate at a particular nucleotide) or due to the selection model. Under
the selection model, the mutation rate is low as expected, but when a mutation occurs it confers a
proliferative advantage to the mutated SSC. The single most common NS mutation (c. 922A>G) in
exon 8 cannot by itself fully explain the high incidence of PTPN11-associated NS in the population.
Since there are other recurrent NS mutations, it could be that the high incidence of NS results from
the accumulated effects of germline selection at these other sites. We analyzed a different region in
the PTPN11 gene, exon 3, which contains 35 codons of exon 3, of which 12 codons are associated
with NS (Table 3) and 20bp are in the intronic region. If many of the NS mutations in this region
show a distribution pattern aligning with a proliferative advantage, perhaps mutations in this region
of the protein manipulate the protein in a way which provides the cell with a selective advantage. In
Table 3, the actual number of NS cases due to new independent germline mutations (Germline
origin) are shown based on data collected by Tartaglia et al., 2006. We define high recurrent NS
mutations as those with more than 10 observed cases. Those sites in Table 3 that are marked (
╪
) have
also been associated with primarily blood cancers.
Table 3. Germline mutations in exon 3 of PTPN11
Amino Acid Wildtype
Amino Acid
Change Nucleotide Change
Germline Origin
(N=573)
N58 AAC H 172 A>C: CAC 1
D 172 A>G: GAC
╪
6
40
K 174 C>G: AAG 3
T59 ACT A 175 A>G: GCT a
G60 GGT S 178 G>A: AGT 1
A 179 G>C: GCT
╪
13
C 178 G>T: TGT
╪
b
deletion 1
D61 GAT N 181 G>A: AAT
╪
16
G 182 A>G: GGT
╪
23
H 181 G>C: CAT
╪
c
A 182 A>C:GCT d
deletion 2
Y62 TAC N 184 T>A: AAC 1
D 184 T>G: GAC
╪
17
C 185 A>G: TGC
╪
1
Y63 TA T C 188 A>G: TGT
╪
44
E69 GAG Q 205 G>C: CAG 2
F71 TTT L 211 T>C: CTT
╪
1
I 211 T>A: ATT 1
A72 GCC S 214 G>T: TCC 18
G 215 C>G: GGC 16
P 214 G>C: CCC e
T73 ACT I 218 C>T: ATT
╪
21
E76 GAG D
228 G>T: GAT; 228
G>C GAC
13
Q79 CAG P 236 A>C: CCG 2
R 236 A>G: CGG 30
The Table shows the Noonan Syndrome mutations in exon 3 of the PTPN11 gene. The germline
origin (N=573) shown were determined by Tartaglia et al., 2006. The sites marked (
╪
) are also
associated with cancer. For those germline origins marked a – e, they were found in other
publications. a. Ko et al., 2008. b. Limal et al., 2006. c. Houweling et al., 2010. d. Chen et al., 2009.
e. Lee et al., 2009. These latter mutations were not reported to cause NS by September 2005 when
Tartaglia et al. gathered the data.
41
3.2.1 Noonan Syndrome Mutation Distribution
We sectioned the testes of three unaffected men, aged 21, 65, and 68, into six slices and each
slice was further sectioned into 32 pieces, for a total of 192 pieces per testis (Fig.10A). We assayed a
172bp region of the PTPN11 gene which contains a portion of exon 3 as described above. The SSS
protocol was performed on each testis piece (1 million genomes per piece) in order to observe the
distribution of NS mutations in the testes. Together these data give us a multi-dimensional data set of
PTPN11 mutations at multiple sites in testes of different age men.
To introduce the data, first let's look at one NS mutation (c.181 G>C) for one donor (#60891,
68 yo). As shown in Figure 10B, most of the testis pieces have very low mutation frequencies. In
fact, 93% of the testis pieces are colored light gray indicating that the observed mutation frequency
is less than 1x10
-5
. In contrast with these low frequency pieces, almost all of the mutations are found
in just three testis pieces found in slice 2. The piece (#12) with the greatest mutation frequency (Mx)
is colored brown and has a mutation frequency of 5.94x10
-3
. This piece is adjacent to a piece (#20)
colored red with a mutation frequency of 2.80x10
-3
and a third adjacent piece (#19) colored pink
with a frequency of 1.36x10
-4
. Pieces 12, 19, and 20 are described as a cluster. For this work, a
cluster is characterized as adjacent testis pieces (or one testis piece) that have elevated mutation
frequencies (colored pink; greater than 1x10
-4
) with at least one of those pieces having a frequency
greater than 1x10
-3
(colored orange). These clusters are surrounded by pieces colored gray with very
low mutation frequencies. For this NS mutation, the average mutation frequency (Av) for the testis is
4.86x10
-5
. Due to the low mutation frequency of the majority of the pieces and the few adjacent
pieces with a high mutation frequency, we would describe this NS mutation as exhibiting a cluster-
like distribution for the testis. We would expect such a distribution if this mutation gives the SSC a
42
selective advantage (See Introduction). We will test this statistically in 3.2.4 Models and Statistical
Analysis.
Figure 10. Distribution of Noonan Syndrome mutation c.181 G>C in a 68 year old man (testis
#60891). A. Testis dissection scheme. B. The frequency of the NS mutation in each testis piece is
indicated by color, the scale is shown in the heat map to the right of the figure. The near white color
specifies those three pieces that did not have enough DNA to analyze (Missing Data).
Figure 11 shows the distribution of twelve NS mutations in the same testis, including NS site
c.181 G>C discussed in Figure 10. The sites shown are a mixture of high recurrent and low recurrent
NS mutations (Table 3). The high recurrent mutation are labeled with an asterisk. Similar to the
previous site (Fig. 10), for these twelve sites almost all of the testis pieces are colored light gray.
Seven of the sites have mutation clusters where at least one of the testis pieces is greater than 1x10
-3

(colored orange, red, or brown). Three of the sites (c.182 A>G, c.188 A>G, c.236 A>G) have
multiple independent clusters. Let’s focus on c.182 A>G. Slice 3 has one cluster; pieces 29 and 30
are colored orange and red, respectively. The second cluster spans slice 5 and 6. In the figure the
43
testis is depicted in two dimensions. Though, in 3 dimensions, the piece with a high mutation
frequency in slice 5 is touching the two orange/red colored pieces in slice 6. These two clusters are
separated by slice 4 where all the testis pieces have low mutation frequencies, colored gray. Three
sites (c. 172 A>G, c.175 A>G, c.184 T>G) have a few testis pieces with elevated mutation
frequencies colored pink, salmon, and light orange. These pieces do not have a high enough
frequency to be considered a cluster. All the testis pieces for two sites (c.205 G>C and c.214 G>C)
are colored gray showing that all of the testis pieces have very low mutation frequencies. The Av
values, Mx values, and other data for all of the NS sites can be found in Appendix 6.1.
The number of NS cases associated with a site does not determine whether it will show a
cluster-like distribution. Some low recurrent sites do not have any mutation clusters and all their
testis pieces are colored gray (c.205 G>C). Though, other low recurrent sites have testis pieces
colored red and brown and have mutation clusters (c.181 G>C). There are high recurrent sites (c.
179 G>C) which have mutation clusters for testis #60891, but do not have any clusters in testis
#60649 (Figure 12).
44
Figure 11. Distribution of twelve Noonan Syndrome mutation sites in testis #60891. Some NS
sites show a cluster-like distribution, while there are no pieces with an elevated mutation frequency
for sites c. 205 G>C and c.214 G>C. Those sites with an asterisk (*) have been observed in over 10
different Noonan Syndrome cases (Table 3). Sites marked with (
╪
) are also associated with cancer.
Mutation sites are order by their position in the gene.
Figure 12 compares twelve NS mutation distributions seen in testis #60891 to another older
donor (#60649, 65 yo) and a younger donor (#63878, 21 yo). Similar to the previous testis, almost
45
all of the testis pieces for this older donor are colored gray with a few clustered pieces with high
mutation frequencies. However, the NS mutations show a heterogeneous distribution between the
two older males. There are six NS sites that have clusters in one older donor but do not have clusters
in the other older donor. This is expected since mutations happen randomly and arbitrarily
throughout the lifespan of the male individual. Testis #60649 has seven sites with mutation clusters
where at least one of the testis pieces has a frequency greater than 1x10
-3
. The testis pieces for two of
the twelve mutation sites for this testis are all colored gray. The remaining three sites have testis
pieces with mutation frequency colored light orange, salmon and pink which are elevated but too
low to be considered clusters.
Unlike both of the older donors, there are no NS sites that have an elevated mutation
frequency in the young donor; none of the pieces exceed a frequency of 3.41x10
-4
(colored salmon).
This low frequency might be expected due to the younger donor having fewer SSC divisions as an
adult compared to the older donors. Having fewer cell divisions limits the number of chances for a
mutation to occur and also limits the size of the mutation cluster once a selective mutation is gained.
The difference in mutation frequencies between the younger and older donors shows that NS
mutations increase in the germline as a man ages.
46
47
Figure 12. Distribution of twelve Noonan Syndrome mutations in the testes of three unaffected
men aged 65 (60649), 68 (60891), and 21 (63878). The older individuals have clusters with high
mutation frequencies while the younger donor has no high mutation frequency clusters. Those sites
with an asterisk (*) have been observed in over 10 different Noonan Syndrome cases (Table 3). Sites
marked with (
╪
) are also associated with cancer. Mutation sites are order by their position in the
gene.
3.2.2 Increased Mutation Frequency with G>T and C>T mutation types
Similar to Table 2 (Eboreime et al., 2016 in submission), we found reference sites of C or G
to have a higher mutation frequency than A and T bases (Table 4; Fig. 13). We found a similar
pattern to the previous experiment where A>C mutation types had the lowest mutation frequency,
while C>T and G>T mutation types had the highest mutation frequencies. The region in exon 3 we
analyzed did not have any CpG sites so that frequency could not be measured.
Table 4. Mutation Frequency for each Mutation Type.
Mutation Type Mutation Frequency
A>C/ T>G 5.7 x 10
-7
C>G/ G>C 1.9 x 10
-6
A>T/ T>A 2.3 x 10
-6
A>G/ T>C 6.8 x 10
-6
C>T/ G>A (non-CpG) 4.5 x 10
-5
G>T/C>A 4.1 x 10
-5
C>T/ G>A (CpG) No CpG sites
The numbers are an average of the mutation frequencies for all the pieces within each testis: 60649,
60891, and 63878.
Fig. 13 shows the distribution for the NS sites which are C>T or G>T mutation types. For
these mutation types, there are more testis pieces with a slightly elevated mutation frequency
(colored pink and salmon) compared with the other mutation types where most of the testis pieces
are colored gray. Although there is an elevated mutation frequency for the majority of pieces, there
are pieces with an exceptionally elevated mutation frequency. One such testis piece is Slice 5 Piece
18 in testis #60649 for mutation c. 218 C>T. This testis piece has a mutation frequency of 2.04x10
-3
,
compared to the average mutation frequency of 1.35x10
-4
, which is an order of magnitude lower. As
48
discussed earlier, we believe this elevated mutation frequency for C>T and G>T mutation types is
due to DNA lesions caused by deamination of 5me-cytosine to thymine or cytosine to uracil as well
as 8-oxo guanine mismatching to adenine.
Figure 13. Noonan Syndrome mutations with C>T and G>T mutations types in the testes of
three unaffected men aged 65 (60649), 68 (60891), and 21 (63878). C>T and G>T mutation types
have an elevated background mutation frequency compared to the other mutation types. Those sites
with an asterisk (*) have been observed in over 10 different Noonan Syndrome cases (Table 3). Sites
marked with (
╪
) are also associated with cancer. Mutation sites are order by their position in the
gene.
3.2.3 Control Sites
There are 79 control sites which are defined as non-disease codons and synonymous
mutations that occur in the third position of a disease site. Figure 14 shows the mutation distribution
for a small fraction of these control sites. To discuss the top panel first (non C>T or G>T mutation
49
types), every testis piece is gray colored with a very low mutation frequency (<1x10
-5
). For control
sites with C>T and G>T mutation types (Figure 14, bottom panel), there is an elevated mutation
frequency. For most of these control sites, the majority of the testis pieces are colored gray, with only
a few pieces with an elevated mutation frequency greater than 2x10
-4
(colored salmon and pink).
However, for two of the sites shown (c.174 C>T, c.193 C>T) most of the testis pieces are salmon or
pink colored showing an elevated mutation frequency for all of the pieces. These sites have no testis
pieces with a mutation frequency higher than 9x10
-4
(colored light orange). We only saw this
elevated frequency for most of the testis pieces for testis #60649 and not the other two testes.
Unfortunately, we do not have an explanation for the even higher mutation frequency for C>T and
G>T mutations uniquely seen in testis #60649. There are no clusters for any of the control sites
suggesting that the mutation clusters seen with the NS mutations are special. All data for the control
sites can be found in Appendix 6.3.
50
Figure 14. Distribution of control site mutations in the testes of three unaffected men aged 65
(60649), 68 (60891), and 21 (63878). Sites shown are nucleotide changes that result in the same
amino acid, unless otherwise noted with (
AC
) and non-disease sites are noted with (#). The bottom
panel shows the C>T and G>T mutation types which have an elevated background mutation
frequency. Mutation sites are ordered by their position in the gene.
51
3.2.4. Models and Statistical Analysis
Using computational modeling and statistical analysis developed by Peter Calabrese, we
showed that the NS mutation clusters are inconsistent with the hotspot model and consistent with the
selection model. We tested four different models: two hot spot models and two selection models.
Each model assumes a different mode of stem cell division, either symmetric or asymmetric (Yoon at
el., 2013). Symmetric stem cell division is defined as the stem cell dividing to produce either two
committed daughter cells or renewing to yield two stem cells (Fig 15). Asymmetric stem cell
division is when the stem cell divides to produce one differentiated daughter cell and another stem
cell (Fig 15). The four models we tested are: symmetric hot spot, asymmetric hot spot, symmetric
selection, and asymmetric selection. Unlike the hot spot model, in the selection model, the mutated
SSC gains a new advantage which promotes cluster formation.
Figure 15. Symmetric versus Asymmetric Stem Cell Divisions. Symmetric renewal results in two
stem cells, symmetric commitment results in two differentiated daughter cells, and asymmetric
division results in one stem cell and one differentiated daughter cell. Red cells are stem cells and
white cells are differentiated daughter cells.
All of the models have two phases: a growth phase and an adult phase. The growth phase
models germline development from zygote formation to puberty. During this phase, for all of the
models, the germline cells divide symmetrically and increase in number exponentially. This allows
for a mutation in this phase to potentially produce a cluster. The germ cells in the growth phase
52
eventually form the self-renewing Apale (SrAp) cells of the adult phase (see Chapter 1). The adult
phase models germline development after puberty. During this phase, the SrAp cells divide every 16
days, providing many more opportunities for mutation than in the growth phase. Described below,
the models differ in how they model the adult phase.
The SrAp either divides asymmetrically or symmetrically. In both the asymmetric and the
symmetric models, SrAp cells continuously produce differentiated cells that will eventually become
sperm and also replenish the stem cell pool. For the asymmetric model, a mutation will only persist
in one stem cell lineage, since the SrAp cell will only produce one stem cell at a time. Therefore, the
asymmetric hot spot model, will not form clusters. For the symmetric hot spot model, half of the
SrAp cells will divide to produce differentiated cells, terminating the stem cell lineage before a
mutation cluster can form. However, the other half of the SrAp cells will divide symmetrically to
form two SrAp cells. Since the daughter SrAp cells remain near their ancestors, a mutation cluster
can grow if symmetric renewal divisions continue. Notably, these mutation clusters can diminish if
followed by symmetric commitment divisions which deplete the stem cell pool. Both hot spot
models have one free parameter: the mutation rate per cell division.
In the selection model, the mutated SrAp are different from the wildtype SrAp. In the
asymmetric selection model, the mutant SrAp occasionally divide symmetrically, enabling a cluster
to grow. In the symmetric selection model, the mutant SrAp are slightly more likely to renew
themselves than to differentiate, likewise enabling a mutation cluster to grow. Both selection models
have two free parameters: the mutation rate per cell division and the selection parameter.
In order to quantify the observed clustering, we computed the ratio of the maximum testis
piece mutation frequency (Mx) to the average mutation frequency for the entire testis (Av). If the
mutation frequency were the same in every testis piece this ratio (Mx/Av) would be 1. This ratio is
53
computed separately for each nucleotide site and for each testis. For those 18 nucleotide sites, among
the three testis, where the Mx value is greater than 2x10
-3
(colored red or brown), Mx/Av values
range from 11 to 189 (average value 83). These data points are depicted as red circles along the x-
axis of the four plots in Figure 16.
In order to test the models, we selected parameter values that best matched the target Av
value. We simulated the models many times with these parameter values and we only considered
those simulations where the simulated Av value was within 5% of the target value. We then
compared the simulated Mx/Av values to those values observed in the testis data. To simultaneously
study all 18 clustered data points, we considered one representative target Av value of 5x10
-5
.
Figure 16 shows that 95% of the simulations for the asymmetric hot spot model have Mx/Av
values between 1 and 4.5. In over one million simulations, all of the simulated Mx/Av values are less
than those values observed in the clustered data. Therefore, the asymmetric hot spot model can be
strongly rejected with p-values less than 10
-6
. This model can be rejected for another reason. In this
model, mutation clusters can only form during the growth phase. If clusters formed during the
growth phase, than all three donors should have clusters no matter their age. However, the 21 year
old donor does not have any mutation clusters like the two older donors, indicating that the mutation
clusters grow in the adult not during development.
54
Figure 16. Histogram of simulated Mx/Av values for the four models. Red circles along the x-
axis represent Mx/Av values observed in the clustered testis data. The two hot spot models are
strongly rejected with p-values less than 10
-6
, while the two selection models are consistent with the
testis data.
The range of the simulated Mx/Av values for the symmetric hot spot model is somewhat
greater than the asymmetric model: 95% of these values are between 4 and 9. However, these
simulated values are not nearly as high as those observed in the testes. For all but two of the data
points in Figure 16, the hot spot model can be rejected with a p-value less than 10
-6
and for those two
points with the smallest Mx/Av values this model can be rejected with p-value 0.01.
55
In contrast, the selection models cannot be rejected. For the asymmetric selection model,
95% of the simulations have Mx/Av values between 79 and 192. This range includes most of the
data points in Figure 16, including those with the largest Mx/Av values. All of the simulations were
performed with the same selection parameter: the probability a mutant SrAp divides symmetrically
is 0.01 (since the SrAp cells divide every 16 days, it’s about one symmetric division every 4 years).
Simulations with a smaller selection parameter could match those data points with smaller Mx/Av
values. Likewise, for the symmetric hot spot model, 95% of the simulated Mx/Av values are between
59 and 184. The peak near 200 in Figure 16 is due to those simulations where almost all of the
mutations are in one testis piece. These simulations were also all performed with only one selection
parameter: the probability a mutant SrAp symmetrically renews is 0.505 and the probability of a
symmetrically differentiating is 0.495 (the wild type SrAp are perfectly balanced with probabilities
0.5 and 0.5 for these two types of division). As in the other selection model, simulations with a
smaller selection parameter could match those data points with smaller Mx/Av values. Also, for both
selection models, the value for the model parameter of the mutation rate per cell division is such that
simulations without selection would produce the genome average mutation frequency. Thus
implying that it is selection and not an elevated mutation rate that produces the increased mutation
frequency.
3.3 CANCER-ASSOCIATED PTPN11 SITES
We observed that some sites not associated with NS within exon 3 of the PTPN11 gene also
had a cluster-like distribution in the testes. Strikingly, we found that these sites were associated with
cancer mutations (Table 5). These PTPN11 mutations are associated with Juvenile Myelomonocytic
Leukemia (JMML), Acute Myeloid Leukemia (AML), Childhood Acute Lymphoblastic Leukemia
(ALL), and Neuroblastoma. SHP2 (the protein product of PTPN11) is a proto-oncogene, or a gene
56
that normally helps to regulate cell growth but when mutated gives rise to an oncogene which leads
to unregulated cell divisions. Most somatic PTPN11 mutations result in hematologic malignancies
but have been seen in solid tumors, as well (Chan et al., 2008, Hou et al., 2008, Jongmans et al.,
2011, Loh et al., 2015). Patients with PTPN11-associated NS have an increased risk of cancers, with
hematological malignancies presenting most frequently (Jongmans et al., 2011). Additionally, 35%
of sporadic JMML is due to PTPN11 mutations, most of which are distinct from NS-associated
PTPN11 sites (Tartaglia et al., 2003).
Table 5 shows those mutations in exon 3 of PTPN11 that are associated with cancers. The
somatic origin are the number of documented cases of cancer for that mutation type as determined
by Tartaglia et al., 2006. Although some of these sites are also associated with NS, it does not mean
that those patients have both NS and cancer. Of the cases below, those patients that have presented
with JMML and NS had mutations D61G (Kratz et al., 2005), Y62C (Bentires-Alj et al., 2004), and
T73I (Kratz et al., 2005; Niihori et al., 2005).
Table 5. Cancer-associated mutations in exon 3 of PTPN11
Amino
Acid Wildtype
Amino Acid
Change
Nucleotide
Change
Somatic
Origin
(N=256) Cancer Type
N58 AAC D 172A>G: GAC
╪
a
Low grade glioneural tumor
Y 172A>T : TAC
1
JMML; ALL
G60 GGT A 179 G>C: GCT
╪
1
Neuroblastoma;
Dysembryoplastic neuro-
epithelial tumor; Common
B ALL
R 178G>C: CGT
3
JMML
V 179 G>T: GTT
15
MDS
C 178G>T: TGT
╪
b
AML M4

D61 GAT N 181G>A: AAT
╪
3
AML
G 182A>G: GGT
╪
1
JMML
H 181G>C: CAT
╪
1
AML
57
Y 181G>T: TAT
19
JMML, ALL
V 182A>T: GTT
15
JMML; MDS; ALL

Y62 TAC D 184T>G: GAC
╪
c
JMML
C 185 A>G: TGC
╪
1 Neuroblastoma
C 185 A>C: TCC d Neuroblastoma

Y63 TA T C 188A>G: TGT
╪
1
Precursor B-ALL

E69 GAG V 206A>T:GTG
1
hematological malignancy
K 205G>A: AAG
10
JMML; MDS; AML; ALL

F71 TTT L 213T>A: TTA
4
MDS; AML
K 211T>C:CTT
╪
2 AML

A72 GCC T 214G>A: ACC
20
JMML; AML; ALL
V 215C>T: GTC
23
JMML; AML; ALL
D 215C>A:GAC 1 Pre-B ALL

T73 ACT I 218 C>T: ATT
╪
6
JMML; AML

L74 TTG M 220 T>A: ATG
e
AML

E76 GAG K 226G>A: AAG
47
JMML; AML; ALL
Q 226G>C: CAG
6
JMML; AML
V 227A>T: GTG
6
JMML
G 227A>G: GGG
23
JMML; AML
A 227A>C: GCG 5
MDS; Behcet's disease with
MDS; ALL

L77 TTG V 229T>G: GTG
1
hematological malignancy
Sites marked with an asterisk (
╪
) are also associated with Noonan Syndrome. The somatic origin
(N=256) shown were determined by Tartaglia et al., 2006. For those somatic origins marked a – e,
they were found in other publications. a. Jongmans et al. 2011; Sherman 2009. b. Jongmans et al.
2011. c. Tartaglia et al., 2003. d. Jongmans et al. 2011. e. Hou et al., 2008. These latter mutations
were not reported to cause cancer by September 2005 when Tartaglia et al. gathered the data.
Figure 17 shows a sample of mutations in exon 3 of PTPN11 which are only associated with
cancer mutations. First, let’s look at the top panel showing mutation types other than C>T and G>T.
Similar to the NS sites, most of the testis pieces are colored gray. Testis #60891 has a mutation
58
cluster for all three sites, while testis #60649 and #63878 do not have any pieces over 1x10
-4

(colored pink). Now let’s look at the cancer mutations which are C>T or G>T mutation types. Even
with the elevated background mutation frequency, most sites have mutation clusters in both of the
older donors. Sites c.179 G>T for #60649 is the only site that does not have a piece with a frequency
above 1x10
-3
. The highest mutation frequency for a cancer-associated site for the 68 year old donor
(60891) was 7.53x10
-3
(c.182 A>T), which, interestingly, is higher than the highest NS mutation
frequency (5.94x10
-3
; c.182 A>G) for that same testis. Notably, for those low recurrent NS sites that
had mutation clusters (c.181 G>C, c.211 T>C), we found that they are associated with cancers (Fig.
12). For the younger donor, there are sites with an elevated background mutation frequency (colored
salmon and pink) but no site shows a cluster-like distribution in the testis. Comparable to the NS
sites, the cancer-associated sites also increase in the germline as the man ages. Due to its cluster-like
distribution in the testis, this increase is predicted by the selection model. All data for the cancer sites
can be found in Appendix 6.2.
Although in the same exon, the cancer-associated PTPN11 mutations have a more drastic
effect on the function of the protein (Zhang et al, 2009), which we conclude is why they lead to a
more aggressive proliferation leading to a cancer phenotype in the offspring. In the Discussion, I
will talk more about these cancer sites and the effect of the mutation on the activity of the SHP2
protein.
59
Figure 17. Cancer-associated PTPN11 mutation distribution in the testes of three unaffected
men aged 65 (60649), 68 (60891), and 21 (63878). The older individuals have clusters with high
mutation frequencies while the younger donor has no high mutation frequency clusters. The bottom
panel shows the C>T and G>T mutation types which have an elevated background mutation
frequency. Those sites with an asterisk (*) have been observed in over 10 different cancer cases
(Table 5). Mutation sites are order by their position in the gene.
60
4. DISCUSSION
My dissertation has investigated the distribution of Noonan Syndrome mutations in the testes
of men of various ages in order to determine if the increased mutation frequency at these recurrent
sites may be due to the hot spot model or the selection model. The hot spot model argues that the
particular site in the genome has a higher mutation frequency than the genome average while the
selection model argues that the site has an average mutation frequency but once a stem cell is
mutated it gains a proliferative advantage over its wildtype counterparts.
4.1 ANATOMICAL CONCLUSIONS
We found that the NS-associated mutations in exon 3 of the PTPN11 gene show a clustered
distribution in the testes of older men and have no clusters in the younger donor which agrees with
the hypothesis that these RAMP mutations give the SSC a selective advantage. Based on their
distributions in the testes, the different mutations arose independently in the testes. In addition, the
mutation clusters were of different sizes. The size of the cluster could be indicative of the time of
mutation occurrence or the strength of selection, however, we are unable to determine which of the
two possibilities with the type of data we have generated. In vivo experiments would be necessary to
test these two possibilities. In future studies, we could compete mouse SSCs with different PTPN11
mutations in vivo to test the strength of selection for each mutation in the mouse but how that would
reflect on humans is unclear.
These NS mutations are highly recurrent in the population and could all be contributing to
the increased frequency of NS in the population through germline selection. This germline selection
can explain the increased frequency and also the male bias and PAE. We found that NS-associated
PTPN11 mutations as well as some cancer-associated NS PTPN11 mutations appear to give SSCs a
selective advantage. Both of these mutation types provide a gain-of-function mutation to the SHP2
61
protein so it is not surprising that they affect the protein in a similar manner. In contrast, LEOPARD
syndrome (LS; multiple Lentigines, Electrocardiographic conduction abnormalities, Ocular
hypertelorism, Pulmonic stenosis, Abnormalities of genitalia, Retardation of growth, and Deafness,
also known as Noonan Syndrome with Multiple Lentigines) mutations, also found in the PTPN11
gene, reduce the catalytic activity of the SHP2 protein (Kontaridis et al., 2006). LS and NS have
similar clinical features, however, they affect SHP2 in different ways. Do these LS mutations also
provide the SSC with a selective advantage or is that reserved only for the gain-of-function
mutations? Recent studies have shown that although the LS mutations have reduced catalytic
activity, the mutant proteins are more readily activated by other pTyr ligands and they localize to
substrates for elongated spans of time causing them to exhibit a gain-of-function phenotype (Yu et
al., 2014). This could explain why NS and LS have similar clinical phenotypes with disparate effects
on the SHP2 protein. Interestingly, this could also lend itself to LS mutations displaying a germline
selection. The SSS protocol should be performed on exons 7 and 12 of the PTPN11 gene to
determine if the LS mutations have a cluster-like distribution in testes thus suggesting germline
selection.
4.2 SPERMATOGONIAL SELECTION
Our data show that the NS mutations found in exon 3 of PTPN11 are distributed through the
testes in a cluster-like pattern and we have concluded, using statistical tests and modeling, that this
pattern agrees with the selection model and the hot spot model is rejected. The selection model states
that these mutations cause the cell to have a proliferative advantage over their wildtype counterparts,
but what exactly is the proliferative advantage?
The role of SHP2 is to remove phosphate groups from phosphorylated tyrosines of target
proteins which either promotes or inhibits certain signaling pathways. SHP2 is involved in the
62
positive regulation of the RAS/MAPK pathway, phosinositide-3-kinase (PI3K) pathway, and
negatively regulates STAT3 (signal transducer and activator of transcription 3) (Chan et al., 2008).
STAT3 has been shown in mice to have an important role in the differentiation of many tissues and
in particular mouse testes (Kisseleva et al., 2002, Murphy et al., 2004). In addition, Song and
Wilkinson found that, in mice, STAT3 promotes SSC differentiation and reducing the
phosphorylation of STAT3 increases the number of SSCs without increasing the total number of cells
(Oatley et al., 2010; rev. in Song and Wilkinson, 2014). The inhibition of STAT3 in mouse SSC
cultures inhibits differentiation but at the same time increases SSC self-renewal.
Interestingly, STAT3 has been shown to be a direct target of SHP2 in other cell types. Zhang
et al. showed that there were reduced levels of phosphorylated STAT3 in human peripheral blood
cells with NS-associated PTPN11 mutations. Specifically, they found that SHP2 E76K and SHP2
N308D lowered the level of phosphorylated STAT3 in bone marrow cells. Further, they found that
the stronger activating cancer mutation E76K had a more robust effect on phosphorylated STAT3
levels compared to the germline Noonan mutation N308D (Zhang et al, 2009).
Until recently, there were no studies showing the direct impact of SHP2 mutations on SSC
maintenance. In 2013, Puri et al. showed that SHP2 knockout mice (removal of exon 4) have
reduced spermatogenesis causing an inhibition in the proliferation of SSCs and undifferentiated
spermatogonia (Puri et al., 2013). They found that SHP2 was essential for the production of
undifferentiated spermatogonia from SSCs, however, germ cells beyond this stage were able to
complete spermatogenesis. Together, these data demonstrate that SHP2 is necessary for the self-
renewal and proliferation of SSCs and most likely functions through the negative regulation of
STAT3.
63
In order for germ cell numbers to be maintained, there needs to be a balance between self-
renewal and differentiation. If there is too much self-renewal, there will be too many stem cells with
no sperm production and if there is too much differentiation, the stem cell pool will be depleted and
new sperm cells will not be produced. The activating PTPN11 mutations could alter this balance ever
so slightly to the side of self-renewal causing the SSCs with PTPN11 mutations to self-renew more
often than wildtype SSCs (Fig. 18) thus increasing the number of PTPN11 mutated SSCs in the testis
over time.
Figure 18. Schematic effect of SHP2 on STAT3 and stem cell self-renewal. NS-associated SHP2
mutations increase SHP2’s negative regulation of STAT3 slightly increasing the self-renewal of
SSCs. Cancer-associated SHP2 mutations greatly increase the negative regulation of STAT3 leading
to unregulated cell growth and self-renewal leading to cancers.
4.3 CANCER-ASSOCIATED PTPN11 MUTATIONS
Most somatic PTPN11 mutations often result in hematologic malignancies (Chan et al., 2008,
Hou et al., 2008, Jongmans et al., 2011, Loh et al., 2015). Researchers have found that germline
PTPN11 mutations increase the risk of cancer and are associated with cancer predisposition
(Tartaglia et al., 2005). In fact, the most common cause of Juvenile myelomonocytic leukemia
64
(JMML) is somatic PTPN11 mutations (Loh et al, 2004; Tartaglia et al., 2003). PTPN11 mutations
are also seen in acute myelogenous leukemia (AML), chronic myelomonocytic leukemia (CMML),
and other myeloid neoplasms. Although, PTPN11 mutations are not often seen in solid tumors they
have been found in neuroblastomas (Bentires-Alj et al., 2004).
Like NS-associated PTPN11 mutations, cancer related mutations are gain-of-function
mutations which have been found to activate the SHP2 protein even more than germline mutations
(See Biochemical Studies below. Martinelli et al., 2008, Tartaglia et al., 2005, Niihori et al., 2005,
Keilhack et al., 2005). Unsurprisingly, we saw that some of these cancer-associated PTPN11
mutations had a cluster-like distribution in testes similar to the NS-associated PTPN11 mutations.
Some NS-associated cancer mutations are associated with the predisposition to cancer but those
mutations that are strictly cancer mutations never present in the germline. Given that strictly cancer
mutations are seen in the testis at an increased frequency, it would be expected that these mutations
would present in live births unless these cancer-associated mutations affect the developing fetus
more severely than NS mutations, the sperm carrying these cancer-associated mutations never
fertilize an ovum, or SSCs with these mutations cannot produce sperm.
4.3.1 Biochemical Studies
Different PTPN11 mutations cause different phenotypes in the individual, including NS, LS,
JMML, AML, and other cancers. This suggests that these different mutations affect the protein in
various ways and thus present as diverse phenotypes in the individuals which carry them. There have
been biochemical studies performed that look at the different PTPN11 mutations and how they affect
the activation level of the SHP2 protein. Mutated and wildtype SHP2 proteins were expressed in
bacteria, purified, and their phosphatase activity was measured in vitro with and without BTAM
peptide (protein tyrosine phosphatase non-receptor type substrate 1, bisphosphotyrosyl-containing
65
activation motif) stimulation (Martinelli et al., 2008). They compared the activation level for
different amino acid substitutions for T42 and E139 to the wildtype protein. The T42A and E139D
mutations are recurring NS mutations in exons 2 and 4, respectively. All but one of the mutants
(T42P) had a similar level of activation without BTAM peptide stimulation. With BTAM
stimulation, most mutations were either activated to a similar level as the wildtype protein or were
not activated at all. However, with BTAM peptide stimulation only the NS-associated mutations,
T42A and E139D, had a highly elevated level of activation of ~20 fold and ~14 fold, respectively.
This demonstrated that the NS mutations have a higher ligand affinity and activation ability than the
other mutants not associated with a disease phenotype (Martinelli et al., 2008).
To understand the mechanism of increased activation, Martinelli et al. compared the crystal
structure of the T42 mutant protein to the wildtype. They found that both the NS T42A and the non-
disease associated mutation T42I interacted with the pTyr phosphate, so that wasn’t the cause of the
increased activation. In contrast, they found a significant difference in the total mobility of the Cα
atoms in the peptides, which could be an indication of the overall stability of the protein-peptide
interaction. They concluded that the NS T42A mutation leads to a stabilization of the protein in the
open conformation thus increasing the level of activation (Martinelli et al., 2008).
Tartaglia et al. compared the biochemical properties of NS-associated PTPN11 mutations to
cancer-associated PTPN11 mutations (Tartaglia et al., 2006). They looked at twelve mutated sites
throughout the PTPN11 gene. Looking specifically at exon 3, they investigated two NS sites (A72S,
E76D), one site that is seen in NS patients that are at a high risk of developing JMML (T73I), and
two that are only seen in leukemia (A72V , E76K). Note that for the NS and cancer sites, they are the
same residue substituted with different amino acids. They compared the activation level of the
proteins with and without stimulation with BTAM peptide. They found that the leukemia-associated
66
mutants had the highest phosphatase activity both with and without stimulation (Fig. 19). The NS-
associated mutants had a slightly increased basal activity and had increased activity after BTAM
peptide stimulation higher than wildtype but significantly lower than the leukemia-associated
mutants. Remarkably, the NS/JMML mutant showed a basal activity level similar to the NS mutants
but a BTAM peptide stimulated activity similar to the leukemia-associated mutants. NS/JMML
patients show a milder course of the JMML, which could be explained by the intermediate effects of
these NS/JMML mutations on the SHP2 protein (Choong et al., 1999).
Figure 19. Phosphate activity assay of wildtype and mutated SHP2 protein performed by
Tartaglia et al., 2006. In vitro assays where activity was measured as picomoles of phosphate
released. White bars represent basal levels and black bars are after BTAM peptide stimulation. The
values are expressed in means ±SD. Modified from Tartaglia et al., 2006 with permission.
Given that the leukemia-associated mutations are never seen as germline events, these data
support that the cancer-associated mutations have a more dramatic effect on the SHP2 protein. As
suggested above, one possibility is that the mutation may have a more severe phenotype resulting in
fetal lethality events. There have been records of fetal lethality in NS, so perhaps this idea is correct.
4.4 ADV ANTAGES AND DRAWBACKS TO THE SSS PROTOCOL
The Safe Sequence System (SSS) is easy and fast to implement on any area of the genome.
The UID and barcode are added on to the template DNA strands via a two-cycle PCR which allows
67
for the targeted capture of the gene of interest. The two-cycle PCR step is followed by a quick
purification step and a second amplification round allows for rapid library preparation which takes
on average 4-5 hours to produce 8 targeted libraries (1 million initial molecules per library) on one
PCR machine. Using a high fidelity polymerase (Hot Start Phusion, in our case) also helps to lower
the rate of PCR errors. SSS has another benefit, a low false positive rate of 3.3x10
-6
for the
mutations that are not C>T and G>T, which have a false positive rate that is an order of magnitude
higher.
In our journey, we found that our data revealed the presence of a higher mutation frequency
for C>T and G>T mutation types, which we concluded are due to deamination of methyl-cytosine
(or cytosine) and the oxidation of guanine to 8-oxo-guanine, respectively. Unfortunately, the two
PCR cycles needed to add on the UID and barcode at the beginning of the protocol negate the ability
to identify when and how these errors are produced. Our results favor the hypothesis that these errors
arise during the first two cycles when the UIDs are being attached. Although there are high
temperatures during the first two cycles, which could lead to deamination and oxidation, the
different temperatures and spans of time at these high temperatures amongst the experiments did not
greatly affect the C>T and G>T mutation rates. It has also been proposed that different NGS library
preparation procedures can cause an increase in the number of C>T and G>T mutations (Costello et
al., 2013) before the amplification step. Due to the addition of the UIDs, in the second round, the
SSS protocol elucidates the presence of PCR and polymerase errors associated with most PCR
reactions. However, the first two cycles of PCR, are done before both UIDs are attached and thus
make it impossible to track those errors.
4.5 ALTERNATIVE PROTOCOLS
68
It would be possible to parse out which strand the error occurred if we could track both
strands of the double stranded molecule together throughout the entire protocol. To do this, the UID
and barcode need to be ligated on to the DNA duplex before any PCR cycles have occurred. In
theory, this seems straightforward, however, due to the low number of starting molecules and also
the targeted nature of this experiment it was very difficult to achieve.
The Duplex Sequencing (DS) method, developed in the Loeb laboratory (Kennedy et al.,
2014), addresses the need for the DNA duplex to be tracked together throughout the protocol
however the DS method was shown to be successful only on small genomic regions of <1Mb, which
is much smaller than the human genome (9x10
9
bps). Also, the DS method has been shown to reduce
the number of mutations detected as due to strand specific lesions such as deamination of cytosine
and oxidation of guanine (Kirsch and Klein, 2012). Subsequently, the decrease in mutation
frequency seen in DS reactions supports the hypothesis that the elevated mutation frequency is due
to PCR errors introduced in the first two cycles of PCR during the SSS protocol.
In an effort to apply the DS method to exon 3 of the PTPN11 gene, I attempted a number of
different methods to ligate on mirror UID adapters onto a targeted region of unamplified human
genomic DNA. Many variations of ligations were attempted including, circle ligation, long-overhang
ligations, A/T ligations, streptavidin bead pull-down, PNA openers, and more. Through all of these
experiments, not one of them produced enough ligated starting molecules to produce enough
resulting UID families to attain any viable data. Similar methods in the past have used mitochondrial
DNA or yeast DNA, which have much smaller genomes than the human genome, and still had low
efficiency (Kennedy et al., 2014; Yu et al., 2013). Future studies aimed at applying the DS method to
a targeted region in the human genome would be greatly beneficial in reducing the error rate
associated with deep targeted NGS sequencing.
69
70
5. REFERENCES
Allanson, J E. “Noonan Syndrome.” Journal of Medical Genetics 24, no. 1 (January 1, 1987): 9–13.
doi:10.1136/jmg.24.1.9.
Amann, R. P. “The Cycle of the Seminiferous Epithelium in Humans: A Need to Revisit?” Journal of
Andrology 29, no. 5 (April 17, 2008): 469–87. doi:10.2164/jandrol.107.004655.
Arnheim, Norman, and Peter Calabrese. “Understanding What Determines the Frequency and
Pattern of Human Germline Mutations.” Nature Reviews Genetics 10, no. 7 (July 2009):
478–88. doi:10.1038/nrg2529.
Bentires-Alj, M. “Activating Mutations of the Noonan Syndrome-Associated SHP2/PTPN11 Gene
in Human Solid Tumors and Adult Acute Myelogenous Leukemia.” Cancer Research 64, no.
24 (December 15, 2004): 8816–20. doi:10.1158/0008-5472.CAN-04-1923.
Brauckhoff, Michael, Oliver Gimm, Carl-Ludwig Weiss, JÖrg Ukkat, Carsten Sekulla, Katrin
Brauckhoff, Phuong Nguyen Thanh, and Henning Dralle. “Multiple Endocrine Neoplasia 2B
Syndrome due to Codon 918 Mutation: Clinical Manifestation and Course in Early and Late
Onset Disease.” World Journal of Surgery 28, no. 12 (December 2004): 1305–11. doi:
10.1007/s00268-004-7637-4.
Campbell, Catarina D., and Evan E. Eichler. “Properties and Rates of Germline Mutations in
Humans.” Trends in Genetics 29, no. 10 (October 2013): 575–84. doi:10.1016/j.tig.
2013.04.005.
Chan, Gordon, Demetrios Kalaitzidis, and Benjamin G. Neel. “The Tyrosine Phosphatase Shp2
(PTPN11) in Cancer.” Cancer and Metastasis Reviews 27, no. 2 (June 2008): 179–92. doi:
10.1007/s10555-008-9126-y.
Chan, R. J. “Human Somatic PTPN11 Mutations Induce Hematopoietic-Cell Hypersensitivity to
Granulocyte-Macrophage Colony-Stimulating Factor.” Blood 105, no. 9 (May 1, 2005):
3737–42. doi:10.1182/blood-2004-10-4002.
Check Hayden, Erika. “Is the $1,000 Genome for Real?” Nature, January 15, 2014. doi:10.1038/
nature.2014.14530.
Chen, C. H., T. H. Chen, S. J. Kuo, C. P. Chen, D. J. Lee, Y . Y . Ke, K. T. Yeh, et al. “Genetic
Evaluation and Management of Fetal Chylothorax: Review and Insights from a Case of
Noonan Syndrome.” Lymphology 42, no. 3 (September 2009): 134–38.
Choong, K., M. H. Freedman, D. Chitayat, E. N. Kelly, G. Taylor, and A. Zipursky. “Juvenile
Myelomonocytic Leukemia and Noonan Syndrome.” Journal of Pediatric Hematology/
Oncology 21, no. 6 (December 1999): 523–27.
71
Clermont, Yves. 1966. “Renewal of Spermatogonia in Man.” American Journal of Anatomy 118 (2):
509–24. doi:10.1002/aja.1001180211.
Choi, Soo-Kyung, Song-Ro Yoon, Peter Calabrese, and Norman Arnheim. “Positive Selection for
New Disease Mutations in the Human Germline: Evidence from the Heritable Cancer
Syndrome Multiple Endocrine Neoplasia Type 2B.” Edited by Bret A. Payseur. PLoS
Genetics 8, no. 2 (February 16, 2012): e1002420. doi:10.1371/journal.pgen.1002420.
Cohen, M. M., S. Kreiborg, E. J. Lammer, J. F. Cordero, P. Mastroiacovo, J. D. Erickson, P. Roeper,
and M. L. Martínez-Frías. “Birth Prevalence Study of the Apert Syndrome.” American
Journal of Medical Genetics 42, no. 5 (March 1, 1992): 655–59. doi:10.1002/ajmg.
1320420505.
Costello, M., T. J. Pugh, T. J. Fennell, C. Stewart, L. Lichtenstein, J. C. Meldrim, J. L. Fostel, et al.
“Discovery and Characterization of Artifactual Mutations in Deep Coverage Targeted
Capture Sequencing Data due to Oxidative DNA Damage during Sample Preparation.”
Nucleic Acids Research 41, no. 6 (April 1, 2013): e67–e67. doi:10.1093/nar/gks1443.
Crow, J. F. “The Origins, Patterns and Implications of Human Spontaneous Mutation.” Nature
Reviews. Genetics 1, no. 1 (October 2000): 40–47. doi:10.1038/35049558.
Darian, Eva, Olgun Guvench, Bing Yu, Cheng-Kui Qu, and Alexander D. MacKerell. “Structural
Mechanism Associated with Domain Opening in Gain-of-Function Mutations in SHP2
Phosphatase.” Proteins: Structure, Function, and Bioinformatics 79, no. 5 (May 2011):
1573–88. doi:10.1002/prot.22984.
De Kretser, D. M., K. L. Loveland, A. Meinhardt, D. Simorangkir, and N. Wreford.
“Spermatogenesis.” Human Reproduction (Oxford, England) 13 Suppl 1 (April 1998): 1–8.
Drost, J. B., and W. R. Lee. “Biological Basis of Germline Mutation: Comparisons of Spontaneous
Germline Mutation Rates among Drosophila, Mouse, and Human.” Environmental and
Molecular Mutagenesis 25 Suppl 26 (1995): 48–64.
“Effect of Advanced Paternal Age on Fertility and Pregnancy.” Accessed November 11, 2014. http://
www.uptodate.com/contents/effect-of-advanced-paternal-age-on-fertility-and-pregnancy.
Gardner, R. J. M. “A New Estimate of the Achondroplasia Mutation Rate.” Clinical Genetics 11, no.
1 (April 23, 2008): 31–38. doi:10.1111/j.1399-0004.1977.tb01274.x.
Goriely, Anne, and Andrew O.M. Wilkie. “Paternal Age Effect Mutations and Selfish
Spermatogonial Selection: Causes and Consequences for Human Disease.” The American
72
Journal of Human Genetics 90, no. 2 (February 2012): 175–200. doi:10.1016/j.ajhg.
2011.12.017.
Haldane, J. B. S. “The Mutation Rate of the Gene for Haemophilia, and Its Segregation Ratios in
Males and Females.” Annals of Eugenics 13, no. 4 (June 1947): 262–71.
Hecht, Norman B. “Molecular Mechanisms of Male Germ Cell Differentiation.” BioEssays 20, no. 7
(July 1, 1998): 555–61. doi:10.1002/(SICI)1521-1878(199807)20:7<555::AID-
BIES6>3.0.CO;2-J.
Heidenreich, Erich. “Adaptive Mutation in Saccharomyces Cerevisiae.” Critical Reviews in
Biochemistry and Molecular Biology 42, no. 4 (January 2007): 285–311. doi:
10.1080/10409230701507773.
Heller, C. G., and Y . Clermont. “Spermatogenesis in Man: An Estimate of Its Duration.” Science
(New York, N.Y.) 140, no. 3563 (April 12, 1963): 184–86.
Heller, C. H., and Y . Clermont. “KINETICS OF THE GERMINAL EPITHELIUM IN MAN.”
Recent Progress in Hormone Research 20 (1964): 545–75.
Holstein, Adolf-Friedrich, Wolfgang Schulze, and Michail Davidoff. “Understanding
Spermatogenesis Is a Prerequisite for Treatment.” Reproductive Biology and Endocrinology:
RB&E 1 (November 14, 2003): 107. doi:10.1186/1477-7827-1-107.
Hof, Peter, Scott Pluskey, Sirano Dhe-Paganon, Michael J Eck, and Steven E Shoelson. “Crystal
Structure of the Tyrosine Phosphatase SHP-2.” Cell 92, no. 4 (February 1998): 441–50. doi:
10.1016/S0092-8674(00)80938-1.
Hou, H-A, W-C Chou, L-I Lin, C-Y Chen, J-L Tang, M-H Tseng, C-F Huang, et al.
“Characterization of Acute Myeloid Leukemia with PTPN11 Mutation: The Mutation Is
Closely Associated with NPM1 Mutation but Inversely Related to FLT3/ITD.” Leukemia 22,
no. 5 (May 2008): 1075–78. doi:10.1038/sj.leu.2405005.
Houweling, A. C., Y . M. de Mooij, I. van der Burgt, H. G. Yntema, A. M. A. Lachmeijer, and A. T. J.
I. Go. “Prenatal Detection of Noonan Syndrome by Mutation Analysis of the PTPN11 and
the KRAS Genes.” Prenatal Diagnosis 30, no. 3 (March 1, 2010): 284–86. doi:10.1002/pd.
2458.
Johnson, Larry. “Spermatogenesis and Aging in the Human.” Journal of Andrology 7, no. 6
(November 12, 1986): 331–54. doi:10.1002/j.1939-4640.1986.tb00943.x.
Jongmans, Marjolijn C J, Ineke van der Burgt, Peter M Hoogerbrugge, Kees Noordam, Helger G
Yntema, Willy M Nillesen, Roland P Kuiper, et al. “Cancer Risk in Patients with Noonan
73
Syndrome Carrying a PTPN11 Mutation.” European Journal of Human Genetics 19, no. 8
(August 2011): 870–74. doi:10.1038/ejhg.2011.37.
Kinde, I., J. Wu, N. Papadopoulos, K. W. Kinzler, and B. V ogelstein. “Detection and Quantification
of Rare Mutations with Massively Parallel Sequencing.” Proceedings of the National
Academy of Sciences 108, no. 23 (June 7, 2011): 9530–35. doi:10.1073/pnas.1105422108.
Keilhack, H., F. S. David, M. McGregor, L. C. Cantley, and B. G. Neel. “Diverse Biochemical
Properties of Shp2 Mutants: IMPLICATIONS FOR DISEASE PHENOTYPES.” Journal of
Biological Chemistry 280, no. 35 (September 2, 2005): 30984–93. doi:10.1074/
jbc.M504699200.
Kennedy, Scott R, Michael W Schmitt, Edward J Fox, Brendan F Kohrn, Jesse J Salk, Eun Hyun
Ahn, Marc J Prindle, et al. “Detecting Ultralow-Frequency Mutations by Duplex
Sequencing.” Nature Protocols 9, no. 11 (October 9, 2014): 2586–2606. doi:10.1038/nprot.
2014.170.
Kirsch, S., and C. A. Klein. “Sequence Error Storms and the Landscape of Mutations in Cancer.”
Proceedings of the National Academy of Sciences 109, no. 36 (September 4, 2012): 14289–
90. doi:10.1073/pnas.1212246109.
Kisseleva, T, S Bhattacharya, J Braunstein, and C.W Schindler. “Signaling through the JAK/STAT
Pathway, Recent Advances and Future Challenges.” Gene 285, no. 1–2 (February 2002): 1–
24. doi:10.1016/S0378-1119(02)00398-0.
Ko, Jung Min, Jae-Min Kim, Gu-Hwan Kim, and Han-Wook Yoo. “PTPN11, SOS1, KRAS, and
RAF1 Gene Analysis, and Genotype-Phenotype Correlation in Korean Patients with Noonan
Syndrome.” Journal of Human Genetics 53, no. 11–12 (2008): 999–1006. doi:10.1007/
s10038-008-0343-6.
Kontaridis, M. I., K. D. Swanson, F. S. David, D. Barford, and B. G. Neel. “PTPN11 (Shp2)
Mutations in LEOPARD Syndrome Have Dominant Negative, Not Activating, Effects.”
Journal of Biological Chemistry 281, no. 10 (March 10, 2006): 6785–92. doi:10.1074/
jbc.M513068200.
Kratz, Christian P., Charlotte M. Niemeyer, Robert P. Castleberry, Mualla Cetin, Eva Bergsträsser,
Peter D. Emanuel, Henrik Hasle, et al. “The Mutational Spectrum of PTPN11 in Juvenile
Myelomonocytic Leukemia and Noonan Syndrome/myeloproliferative Disease.” Blood 106,
no. 6 (September 15, 2005): 2183–85. doi:10.1182/blood-2005-02-0531.
Lauriol, Jessica, Fabrice Jaffré, and Maria I. Kontaridis. “The Role of the Protein Tyrosine
Phosphatase SHP2 in Cardiac Development and Disease.” Seminars in Cell &
Developmental Biology 37 (January 2015): 73–81. doi:10.1016/j.semcdb.2014.09.013.
74
Lee, Ka, B Williams, K Roza, H Ferguson, K David, K Eddleman, J Stone, et al. “PTPN11 Analysis
for the Prenatal Diagnosis of Noonan Syndrome in Fetuses with Abnormal Ultrasound
Findings.” Clinical Genetics 75, no. 2 (February 1, 2009): 190–94. doi:10.1111/j.
1399-0004.2008.01085.x.
Lee, P. A. “Normal Ages of Pubertal Events among American Males and Females.” Journal of
Adolescent Health Care: Official Publication of the Society for Adolescent Medicine 1, no. 1
(September 1980): 26–29.
Limal, Jean-Marie, Béatrice Parfait, Sylvie Cabrol, Damien Bonnet, Bruno Leheup, Stanislas
Lyonnet, Michel Vidaud, and Yves Le Bouc. “Noonan Syndrome: Relationships between
Genotype, Growth, and Growth Factors.” The Journal of Clinical Endocrinology and
Metabolism 91, no. 1 (January 2006): 300–306. doi:10.1210/jc.2005-0983.
Liu, Qiang, and Steve S. Sommer. “Detection of Extremely Rare Alleles by Bidirectional
Pyrophosphorolysis-Activated Polymerization Allele-Specific Amplification (Bi-PAP-A):
Measurement of Mutation Load in Mammalian Tissues.” BioTechniques 36, no. 1 (January
2004): 156–66.
Loh, M. L. “Mutations in PTPN11 Implicate the SHP-2 Phosphatase in Leukemogenesis.” Blood
103, no. 6 (March 15, 2004): 2325–31. doi:10.1182/blood-2003-09-3287.
Lou, D. I., J. A. Hussmann, R. M. McBee, A. Acevedo, R. Andino, W. H. Press, and S. L. Sawyer.
“High-Throughput DNA Sequencing Errors Are Reduced by Orders of Magnitude Using
Circle Sequencing.” Proceedings of the National Academy of Sciences 110, no. 49
(December 3, 2013): 19872–77. doi:10.1073/pnas.1319590110.
Makova, Kateryna D., and Wen-Hsiung Li. “Strong Male-Driven Evolution of DNA Sequences in
Humans and Apes.” Nature 416, no. 6881 (April 11, 2002): 624–26. doi:10.1038/416624a.
Margulies, Marcel, Michael Egholm, William E. Altman, Said Attiya, Joel S. Bader, Lisa A.
Bemben, Jan Berka, et al. “Genome Sequencing in Microfabricated High-Density Picolitre
Reactors.” Nature, July 31, 2005. doi:10.1038/nature03959.
Martin JA, Hamilton BE, Osterman MJK, et al. Births: Final data for 2013. National vital statistics
reports; vol 64 no 1. Hyattsville, MD: National Center for Health Statistics. 2015.
Martinelli, S., P. Torreri, M. Tinti, L. Stella, G. Bocchinfuso, E. Flex, A. Grottesi, et al. “Diverse
Driving Forces Underlie the Invariant Occurrence of the T42A, E139D, I282V and T468M
SHP2 Amino Acid Substitutions Causing Noonan and LEOPARD Syndromes.” Human
Molecular Genetics 17, no. 13 (March 12, 2008): 2018–29. doi:10.1093/hmg/ddn099.
75
Murphy, Katherine, Luis Carvajal, Leo Medico, and Melissa Pepling. “Expression of Stat3 in Germ
Cells of Developing and Adult Mouse Ovaries and Testes.” Gene Expression Patterns 5, no.
4 (April 2005): 475–82. doi:10.1016/j.modgep.2004.12.007.
Nachman, M. W., and S. L. Crowell. “Estimate of the Mutation Rate per Nucleotide in Humans.”
Genetics 156, no. 1 (September 2000): 297–304.
Niihori, Tetsuya, Yoko Aoki, Hirofumi Ohashi, Kenji Kurosawa, Tatsuro Kondoh, Satoshi
Ishikiriyama, Hiroshi Kawame, et al. “Functional Analysis of PTPN11/SHP-2 Mutants
Identified in Noonan Syndrome and Childhood Leukemia.” Journal of Human Genetics 50,
no. 4 (April 2005): 192–202. doi:10.1007/s10038-005-0239-7.
Noonan, Jacqueline A., and Dorothy A. Ehmke. “Complications of Ventriculovenous Shunts for
Control of Hydrocephalus: Report of Three Cases with Thromboemboli to the Lungs.” New
England Journal of Medicine 269, no. 2 (July 11, 1963): 70–74. doi:10.1056/
NEJM196307112690203.
Nyren, P., B. Pettersson, and M. Uhlen. “Solid Phase DNA Minisequencing by an Enzymatic
Luminometric Inorganic Pyrophosphate Detection Assay.” Analytical Biochemistry 208, no.
1 (January 1993): 171–75. doi:10.1006/abio.1993.1024.
Oatley, J. M., A. V . Kaucher, M. R. Avarbock, and R. L. Brinster. “Regulation of Mouse
Spermatogonial Stem Cell Differentiation by STAT3 Signaling.” Biology of Reproduction 83,
no. 3 (September 1, 2010): 427–33. doi:10.1095/biolreprod.109.083352.
O’Donnell, Liza, Peter K Nicholls, Moira K O’Bryan, Robert I McLachlan, and Peter G Stanton.
“Spermiation.” Spermatogenesis 1, no. 1 (2011): 14–35. doi:10.4161/spmg.1.1.14525.
Puri, Pawan, Bart T. Phillips, Hitomi Suzuki, Kyle E. Orwig, Aleksandar Rajkovic, Philip E.
Lapinski, Philip D. King, Gen-Sheng Feng, and William H. Walker. “The Transition from
Stem Cell to Progenitor Spermatogonia and Male Fertility Requires the SHP2 Protein
Tyrosine Phosphatase: Stem Cell to Spermatogonia Transit Requires SHP2.” STEM CELLS
32, no. 3 (March 2014): 741–53. doi:10.1002/stem.1572.
Qin, Jian, Peter Calabrese, Irene Tiemann-Boege, Deepali Narendra Shinde, Song-Ro Yoon, David
Gelfand, Keith Bauer, and Norman Arnheim. “The Molecular Anatomy of Spontaneous
Germline Mutations in Human Testes.” Edited by Jim Crow. PLoS Biology 5, no. 9 (August
28, 2007): e224. doi:10.1371/journal.pbio.0050224.
Sánchez, Flor, and Johan Smitz. “Molecular Control of Oogenesis.” Biochimica et Biophysica Acta
(BBA) - Molecular Basis of Disease 1822, no. 12 (December 2012): 1896–1912. doi:
10.1016/j.bbadis.2012.05.013.
76
Saxton, Tracy M., Mark Henkemeyer, Stephan Gasca, Randy Shen, Derrick J. Rossi, Fouad Shalaby,
Gen-Sheng Feng, and Tony Pawson. “Abnormal Mesoderm Patterning in Mouse Embryos
Mutant for the SH2 Tyrosine Phosphatase Shp-2.” The EMBO Journal 16, no. 9 (May 1,
1997): 2352–64. doi:10.1093/emboj/16.9.2352.
Scally, Aylwyn, Julien Y . Dutheil, LaDeana W. Hillier, Gregory E. Jordan, Ian Goodhead, Javier
Herrero, Asger Hobolth, et al. “Insights into Hominid Evolution from the Gorilla Genome
Sequence.” Nature 483, no. 7388 (March 7, 2012): 169–75. doi:10.1038/nature10842.
Schlatt, Stefan, and Jens Ehmcke. “Regulation of Spermatogenesis: An Evolutionary Biologist’s
Perspective.” Seminars in Cell & Developmental Biology 29 (May 2014): 2–16. doi:10.1016/
j.semcdb.2014.03.007.
Sherman, Courtney B., Afshan Ali-Nazir, Ignacio Gonzales-Gomez, Jonathan L. Finlay, and Girish
Dhall. “Primary Mixed Glioneuronal Tumor of the Central Nervous System in a Patient With
Noonan Syndrome: A Case Report and Review of the Literature.” Journal of Pediatric
Hematology/Oncology 31, no. 1 (January 2009): 61–64. doi:10.1097/MPH.
0b013e31818ab2cf.
Shinde, D. N., D. P. Elmer, P. Calabrese, J. Boulanger, N. Arnheim, and I. Tiemann-Boege. “New
Evidence for Positive Selection Helps Explain the Paternal Age Effect Observed in
Achondroplasia.” Human Molecular Genetics 22, no. 20 (October 15, 2013): 4117–26. doi:
10.1093/hmg/ddt260.
Song, Hye-Won, and Miles F. Wilkinson. “Transcriptional Control of Spermatogonial Maintenance
and Differentiation.” Seminars in Cell & Developmental Biology 30 (June 2014): 14–26. doi:
10.1016/j.semcdb.2014.02.005.
Tartaglia, Marco, Kamini Kalidas, Adam Shaw, Xiaoling Song, Dan L. Musat, Ineke van der Burgt,
Han G. Brunner, et al. “PTPN11 Mutations in Noonan Syndrome: Molecular Spectrum,
Genotype-Phenotype Correlation, and Phenotypic Heterogeneity.” The American Journal of
Human Genetics 70, no. 6 (June 2002): 1555–63. doi:10.1086/340847.
Tartaglia, Marco, Charlotte M Niemeyer, Alessandra Fragale, Xiaoling Song, Jochen Buechner,
Andreas Jung, Karel Hählen, Henrik Hasle, Jonathan D Licht, and Bruce D Gelb. “Somatic
Mutations in PTPN11 in Juvenile Myelomonocytic Leukemia, Myelodysplastic Syndromes
and Acute Myeloid Leukemia.” Nature Genetics 34, no. 2 (June 2003): 148–50. doi:10.1038/
ng1156.
Tartaglia, Marco, Simone Martinelli, Ivano Iavarone, Giovanni Cazzaniga, Monica Spinelli,
Emanuela Giarin, Valentina Petrangeli, et al. “Somatic PTPN11 Mutations in Childhood
Acute Myeloid Leukaemia.” British Journal of Haematology 129, no. 3 (May 2005): 333–39.
doi:10.1111/j.1365-2141.2005.05457.x.
77
Tartaglia, Marco, Simone Martinelli, Lorenzo Stella, Gianfranco Bocchinfuso, Elisabetta Flex,
Viviana Cordeddu, Giuseppe Zampino, et al. “Diversity and Functional Consequences of
Germline and Somatic PTPN11 Mutations in Human Disease.” The American Journal of
Human Genetics 78, no. 2 (February 2006): 279–90. doi:10.1086/499925.
Tartaglia, M., G. Zampino, and B.D. Gelb. “Noonan Syndrome: Clinical Aspects and Molecular
Pathogenesis.” Molecular Syndromology 1, no. 1 (2010): 2–26. doi:10.1159/000276766.
Taylor, James, Svitlana Tyekucheva, Michael Zody, Francesca Chiaromonte, and Kateryna D.
Makova. “Strong and Weak Male Mutation Bias at Different Sites in the Primate Genomes:
Insights from the Human-Chimpanzee Comparison.” Molecular Biology and Evolution 23,
no. 3 (March 2006): 565–73. doi:10.1093/molbev/msj060.
van Alphen, M. M. A., H. J. G. van de Kant, and D. G. de Rooij. “Repopulation of the Seminiferous
Epithelium of the Rhesus Monkey after X Irradiation.” Radiation Research 113, no. 3 (March
1988): 487. doi:10.2307/3577245.
Witsuba, Joachim, Jan-Bernd Stukenborg, and C. Marc Luetjens. “Mammalian Spermatogenesis.” In
Functional Development and Embryology, 2:99–117. Mammalian Spermatogenesis 1, 2007.
Yoon, Song-Ro, Jian Qin, Rivka L. Glaser, Ethylin Wang Jabs, Nancy S. Wexler, Rebecca Sokol,
Norman Arnheim, and Peter Calabrese. “The Ups and Downs of Mutation Frequencies
during Aging Can Account for the Apert Syndrome Paternal Age Effect.” Edited by Bruce
Walsh. PLoS Genetics 5, no. 7 (July 10, 2009): e1000558. doi:10.1371/journal.pgen.
1000558.
Yoon, Song-Ro, Soo-Kung Choi, Jordan Eboreime, Bruce D. Gelb, Peter Calabrese, and Norman
Arnheim. “Age-Dependent Germline Mosaicism of the Most Common Noonan Syndrome
Mutation Shows the Signature of Germline Selection.” The American Journal of Human
Genetics 92, no. 6 (June 2013): 917–26. doi:10.1016/j.ajhg.2013.05.001.
Yoon, Song-Ro, Soo-Kung Choi, Jordan Eboreime, Bruce D. Gelb, Peter Calabrese, and Norman
Arnheim. “Age-Dependent Germline Mosaicism of the Most Common Noonan Syndrome
Mutation Shows the Signature of Germline Selection.” The American Journal of Human
Genetics 92, no. 6 (June 2013): 917–26. doi:10.1016/j.ajhg.2013.05.001.
Yu, Zhi-Hong, Ruo-Yu Zhang, Chad D. Walls, Lan Chen, Sheng Zhang, Li Wu, Sijiu Liu, and
Zhong-Yin Zhang. “Molecular Basis of Gain-of-Function LEOPARD Syndrome-Associated
SHP2 Mutations.” Biochemistry 53, no. 25 (July 2014): 4136–51. doi:10.1021/bi5002695.
Zhang, Leying, Darya Alizadeh, Michelle Van Handel, Marcin Kortylewski, Hua Yu, and Behnam
Badie. “Stat3 Inhibition Activates Tumor Macrophages and Abrogates Glioma Growth in
Mice.” Glia 57, no. 13 (October 2009): 1458–67. doi:10.1002/glia.20863.
78
Zhang, W., R. J. Chan, H. Chen, Z. Yang, Y . He, X. Zhang, Y . Luo, et al. “Negative Regulation of
Stat3 by Activating PTPN11 Mutants Contributes to the Pathogenesis of Noonan Syndrome
and Juvenile Myelomonocytic Leukemia.” Journal of Biological Chemistry 284, no. 33
(August 14, 2009): 22353–63. doi:10.1074/jbc.M109.020495.
79
6. APPENDIX
6.1 Noonan-associated mutations in exon 3 of PTPN11
80
6.2 Cancer-associated mutations in exon 3 of PTPN11
81
6.3 Control sites in exon 3 of PTPN11
82
83
84

Abstract (if available)

Abstract Noonan Syndrome (NS) is one of the most common disorders. Recent evidence suggests the most common NS mutation (PTPN11 c.922 G>A) likely confers a selective advantage on newly mutant spermatogonial stem cells (SSC) in unaffected men, however, there are additional NS mutations in the PTPN11 gene. The fact that many other NS mutations are also recurrent raises the possibility that the high incidence of NS results from the accumulated effects of germline selection at many sites. We asked if the high frequency of sporadic NS cases in humans is due to a high mutation rate per cell division (hot spot) at a small number of disease sites or if rare mutations at these sites confer a selective advantage to the mutated SSCs leading to high numbers of sperm carrying the NS disease mutations (selection). We measured the NS mutation frequency in exon 3 of PTPN11 in three testes from unaffected men using a modified Safe Sequence System version of NGS. In the older donors (65 and 68 years old), almost all of the NS mutations are in a small number of often-adjacent testis pieces with very high mutation frequencies (maximum testis piece mutation frequency is 5.9x10⁻³). Computational modeling and statistical analysis show that the NS mutation clusters are inconsistent with the hot spot model but consistent with the selection model.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Identification of CBP/FOXM1 as a molecular target in triple negative breast cancer

PDF

Role of the bone marrow niche components in B cell malignancies

PDF

Modeling SynGAP1 truncating mutations in neurodevelopmental disease using iPSC-derived neurons

PDF

Identification of small molecular inhibitors targeting G13D mutant K-ras

PDF

Evolutionary genomic analysis in heterogeneous populations of non-model and model organisms

PDF

Molecular signature of aggressive disease and clonal diversity revealed by single-cell copy number analysis of prostate cancer cells across multiple disease states

PDF

The role of glucose-regulated proteins in endometrial and pancreatic cancers

PDF

Targeted cell imaging by in-situ assembly and activation of hot spot SERS nanoprobes using split-fluorescent protein scaffolds

PDF

The role of adipocytes in acute lymphoblastic leukemia cell migration and survival against daunorubicin

PDF

The roles of endoplasmic reticulum chaperones in regulating liver homeostasis and tumorigenesis

PDF

Genomic variant analysis in whole blood of recurrent early-onset major depression patients and differential expression in BA25 of suicide completers with major depression

PDF

Detecting joint interactions between sets of variables in the context of studies with a dichotomous phenotype, with applications to asthma susceptibility involving epigenetics and epistasis

PDF

Uncovering the protective role of protein glycosylation in Parkinson's disease utilizing protein semi-synthesis

Asset Metadata

Creator Eboreime, Jordan (author)

Core Title Molecular anatomy of Noonan syndrome mutations in the testes of unaffected men

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Publication Date 07/15/2016

Defense Date 04/21/2016

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

Tag deep sequencing,genetics,germline mutations,human genetics,male bias,next generation sequencing,Noonan syndrome,OAI-PMH Harvest,paternal age effect,rare mutations,selection,selective advantage,Spermatogenesis,targeted sequencing

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Arnheim, Norman (committee chair), Calabrese, Peter (committee member), Dean, Matthew (committee member), Shibata, Darryl (committee member)

Creator Email jordaneboreime@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-268097

Unique identifier UC11281150

Identifier etd-EboreimeJo-4549.pdf (filename),usctheses-c40-268097 (legacy record id)

Legacy Identifier etd-EboreimeJo-4549.pdf

Dmrecord 268097

Document Type Dissertation

Format application/pdf (imt)

Rights Eboreime, Jordan

Type texts

Source 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 a...

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