Wright State University CORE Scholar
Browse all Theses and Dissertations Theses and Dissertations
2019
Genotyping for Response to Physical Training
Stacy Simmons Wright State University
Follow this and additional works at: https://corescholar.libraries.wright.edu/etd_all
Part of the Molecular Biology Commons
Repository Citation Simmons, Stacy, "Genotyping for Response to Physical Training" (2019). Browse all Theses and Dissertations. 2109. https://corescholar.libraries.wright.edu/etd_all/2109
This Thesis is brought to you for free and open access by the Theses and Dissertations at CORE Scholar. It has been accepted for inclusion in Browse all Theses and Dissertations by an authorized administrator of CORE Scholar. For more information, please contact [email protected].
GENOTYPING FOR RESPONSE TO PHYSICAL TRAINING
A thesis submitted in partial fulfillment of
the requirements for the degree of Master of Science
By
STACY SIMMONS
B.S., Wright State University, 2014
______2019 Wright State University
WRIGHT STATE UNIVERSITY
GRADUATE SCHOOL
July 29, 2019
I HEREBY RECOMMEND THAT THE THESIS PREPARED UNDER MY SUPERVISION BY
Stacy Simmons ENTITLED Genotyping for Response to Physical Training BE
ACCEPTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
Master of Science.
______Michael Markey, Ph.D. Thesis Director
______Madhavi P. Kadakia, Ph.D. Committee on Chair, Department of Biochemistry Final Examination and Molecular Biology
______Michael Markey, Ph.D.
______Richard Chapleau, Ph.D.
______Madhavi Kadakia, Ph.D.
______Barry Milligan, Ph.D. Interim Dean of the Graduate School
ABSTRACT
Simmons, Stacy. M.S. Department of Biochemistry and Molecular Biology, Wright State University 2019. Genotyping for Response to Physical Training
Understanding the inter-individual variability in physical fitness performance has been the focus of scientific research for decades especially in the United States military. Injury and physical inadequacy cost the U.S military millions of dollars every year. The project PHITE (Precision High
Intensity Training through Epigenetics) was funded to investigate this personal complex trait by combing the genetic and epigenetic (non-shared environmental factors) contributions into a single model for physical training response. This project is set up as having 150 male and female recruits between the ages of 18-27 years old. Each participant is randomly put blind into either a high intensity or moderate intensity (group A or group B) 12-week training program. The training response was measured by various means including % change in chest press, lat pull,
VO2max, among others and muscle biopsies and blood samples were collected regularly. The genetic contribution was examined by using Taqman qPCR genotyping for common single nucleotide polymorphisms (SNP) found in the literature to contribute to physical performance.
By combining the genotyping data with the training response data, it was hypothesized that the training response would be different depending on the training group assigned and that genotype can be used to develop a predictor for training response. There were two main aims that were investigated for this hypothesis: the performance difference between the high and moderate intensity; and whether genotype could be used as a predictor of this training response. Individual training responses and overall performance in each training group was investigated using ANOVA analysis. It was determined that there wasn’t a significant difference
iii | P a g e
in individual training responses for each group, but overall performance in group B was significantly higher than group A (p=0.0059). Five different analyses were used to investigate whether genotype could be used as a predictor for training response: ANOVA of beneficial allele score, ANOVA of performance score, Fisher’s exact test, ANOVA of individual performance results, and GWAS analysis. 16 SNPs were found to be associated with either individual training responses or overall performance response. These results support the importance of considering genotype in any model of physical training response.
iv | P a g e
Table of Contents
I. INTRODUCTION ...... 1 A. Physical Fitness in the Military ...... 1 B. Genotyping Using Single Nucleotide Polymorphisms ...... 2
1. Aerobic Capacity/Endurance/VO2max ...... 3 2. Physical Strength/Response to Training/Muscle Volume ...... 4 3. Blood Pressure/Cardiovascular/Heart Rate ...... 4 4. Increased BMI/Obesity/Diabetes ...... 5 5. Endurance/Strength ...... 5 6. Other (Bone Density/Height/Metabolism) ...... 6 7. Epigenetic Modifiers ...... 6 II. MATERIALS AND METHODS ...... 7 A. Subjects and Training Protocol ...... 7 B. DNA Isolation and Genotyping Analysis ...... 9 C. Beneficial Allele Score ...... 10 D. Statistical Analysis ...... 11 1. Chi Squared Analysis ...... 11 2. K-Means Clustering and Performance Score ...... 13 3. Fisher’s Exact Test ...... 13 4. Analysis of Variance (ANOVA) ...... 14 5. Bonferroni Correction ...... 14 6. Genome Wide Association Study (GWAS) ...... 14 III. RESULTS...... 15 A. AIM I. Training Group A Versus B Physical Performance Parameters ...... 15 1. Group A Versus B: Individual Performance Parameters ...... 15 1. Group A Versus B: Performance Score ...... 15 B. AIM II. Investigate if genotype can be used to predict phenotype in individual training responses and in overall training response ...... 19 1. Beneficial Allele Score ...... 20 2. Performance Score ...... 22 3. Fisher’s Exact Test ...... 34
v | P a g e
4. ANOVA Analysis ...... 38 5. GWAS Analysis ...... 44 IV. DISCUSSION ...... 46 V. FUTURE DIRECTIONS ...... 55 VI. SUPPLEMENTARY DATA ...... 56 VII. BIBLIOGRAPHY ...... 83
vi | P a g e
LIST OF TABLES
Table 1: Summary of SNPs Chosen for PHITE ...... 3 Table 2: Physical Performance Parameters (% Change 0-12 Weeks) ...... 9 Table 3: Chi Squared Analysis of all 42 Completers ...... 19 Table 4: SNPs Shown to be Associated with Performance Score ...... 22 Table 5: Bonferroni Corrected Fisher’s Exact Test SNP Results ...... 34 Table 6: SNPs Shown to Predict Specific Training Responses by ANOVA ...... 38 Table 7: SNPs Shown to Predict Specific Training Responses by GWAS ...... 44 Table 8: SNPs Beneficial Alleles Associated with Overall Performance ...... 49 Table 9: Summary Table for all SNPs Found to be Possible Training Response Predictors ...... 53
vii | P a g e
LIST OF FIGURES
Figure 1: The Moderate and High Intensity Training Protocols ...... 8 Figure 2: Ethnicity Data for the First 42 Completers ...... 12 Figure 3: Group A Versus Group B Comparison for Arm Lean Mass % Change...... 17 Figure 4: A Versus B Training Group Comparison for Performance Score ...... 18 Figure 5: Beneficial Allele Score versus Vertical Jump Categorization ...... 21 Figure 6: SNP rs11096663 genotypes versus the Performance Score ...... 24 Figure 7: SNP rs1546570 genotypes versus the Performance Score ...... 25 Figure 8: SNP rs1801394 genotypes versus the Performance Score ...... 27 Figure 9: SNP rs2276731 genotypes versus the Performance Score ...... 29 Figure 10: SNP rs328 genotypes versus the Performance Score ...... 31 Figure 11: SNP rs7964046 genotypes versus the Performance Score ...... 32 Figure 12: Beneficial Allele Score Versus Performance Score ...... 33 Figure 13: Non-Responders Versus High-Responders for rs11689011 – EPAS1 for Knee Peak Power % Change ...... 35 Figure 14: Non-Responders Versus High-Responders for rs1801394 – MTRR for Lat Pull % Change ...... 36 Figure 15: Non-Responders Versus High-Responders for rs2979481 – RBPMS for Knee Extension% Change ...... 37 Figure 16: SNP rs10248479 Association with Anaerobic Peak Power % Chg 0-12 By ANOVA ...... 39 Figure 17: SNP rs10248479 Association with Relative Peak Power % Chg 0-12 By ANOVA ...... 40 Figure 18: SNP rs1532723 Association with Squat % Chg 0-12 By ANOVA ...... 41 Figure 19: SNP rs4994 Association with Vertical Jump % Chg 0-12 By ANOVA ...... 42 Figure 20: SNP rs7386139 Association with Chest Press % Chg 0-12 By ANOVA ...... 43 Figure 21: GWAS Plots for the Training Responses that had Significant SNP Associations ...... 45
viii | P a g e
LIST OF EQUATIONS
Equation 1: Example of How Beneficial Allele Score was Calculated ...... 10 Equation 2: Hardy-Weinberg Equilibrium Equation ...... 11 Equation 3: Equation for the Corrected Minor Allele Frequency Calculation (cMAF) ...... 11
ix | P a g e
LIST OF SUPPLEMENTARY TABLES
Table S1: Raw Genotyping Data for the First 16 Completers ...... 56 Table S2: Raw Genotyping Data for the Completers 17-29 ...... 61 Table S3: Raw Genotyping Data for the Completers 30-42 ...... 65 Table S4: SNPs Associated with Aerobic Capacity ...... 70 Table S5: SNPs Associated with Physical Strength/Response to Training ...... 71 Table S6: SNPs Associated with Cardiovascular Health ...... 72 Table S7: SNPs Associated with Increased BMI/Obesity ...... 74 Table S8: SNPs Associated with Endurance and Strength ...... 75 Table S9: SNPs associated with Other Aspects of Physical Fitness ...... 75 Table S10: SNPs Associated with Epigenetic Modifiers ...... 75 Table S11: ANOVA Analysis predictive SNPs for Anaerobic Peak Power % Change 0-12 Weeks ...... 76 Table S12: ANOVA Analysis predictive SNPs for Relative Peak Power % Change 0-12 Weeks ...... 76 Table S13: ANOVA Analysis predictive SNPs for Chest Press % Change 0-12 Weeks ...... 77 Table S14: ANOVA Analysis predictive SNPs for Knee Extension % Change 0-12 Weeks ...... 77 Table S15: ANOVA Analysis predictive SNPs for Overhead Press % Change 0-12 Weeks ...... 78 Table S16: ANOVA Analysis predictive SNPs for Knee Peak Power % Change 0-12 Weeks ...... 78 Table S17: ANOVA Analysis predictive SNPs for Lat Pull % Change 0-12 Weeks ...... 78 Table S18: ANOVA Analysis predictive SNPs for Vertical Jump % Change 0-12 Weeks ...... 79 Table S19: ANOVA Analysis predictive SNPs for Squat % Change 0-12 Weeks ...... 79
Table S20: ANOVA Analysis predictive SNPs for VO2max (ml/kg/min) % Change 0-12 Weeks ...... 80 Table S21: ANOVA Analysis predictive SNPs for Arm Lean Mass LM Change 0-12 Weeks (grams) ...... 80 Table S22: ANOVA Analysis predictive SNPs for Max Voluntary Contraction % Change 0-12 Weeks ...... 80 Table S23: ANOVA Analysis predictive SNPs for Bodyfat % Change 0-12 Weeks ...... 81 Table S24: ANOVA Analysis predictive SNPs for Thigh Lean Mass, Change 0-12 Weeks (grams) ...... 81 Table S25: ANOVA Analysis predictive SNPs for Visceral Fat Mass Change 0-12 Weeks (grams) ...... 82 Table S26: ANOVA Analysis predictive SNPs for Total Lean Mass Change 0-12 Weeks (grams) ...... 82
x | P a g e
ACKNOWLEGDGMENT
There are many people I want to thank for supporting me through the process of writing a thesis and obtaining a master’s degree in the Biochemistry program. First and foremost, I want to thank my advisor, Dr. Michael Markey for everything he has done for me the last two years in the Biochemistry program at Wright State University. He was very knowledgeable, patient, and encouraging during the completion of my research project and defense. I would also like to thank my committee members, Dr. Madhavi Kadakia and Dr. Richard Chapleau for all the guidance and support. Moreover, I want to show my great appreciation for everyone in the
BMB department, all the teachers were wonderful and very helpful and supportive to not only their own students, but to everyone in the program. I would also like to thank the grant that funded the project that encompasses my thesis work: ONR MURI Grant N00017-16-1-3159.
I would also like to personally thank my lab members, Karleigh Kessler and Abdullah Alatawi, and all my classmates that helped support me through all the lab work, class work and life. I would also like to thank my boss, Justin Metcalf, for allowing me to pursue this higher education and dealing with my erratic schedule the last two years.
Lastly, I would like to thank all my personal family and friends who supported me throughout this process and always were there to encourage and love me. I would like to especially thank my husband, CJ Simmons, for being my rock and who has always supported me through every crazy adventure. Thank you all!
xi | P a g e
I. INTRODUCTION
A. Physical Fitness in the Military
Physical fitness is determined by numerous genes and epigenetic modifications to those genes and many previous studies have used genetic markers known as single nucleotide polymorphisms (SNPs) to study the differences in physical fitness among individuals (Piotr
Gronek, 2013). The challenge with this kind of study is only 66% of variance in physical fitness can be explained with genetic markers; the other 34% is from non-shared environmental factors that produce epigenetic variation (Bamman & Broderick, 2018). Inter-individual variability in physical fitness has always been a focus in scientific research specifically in a military setting. Injury and physical inadequacy lead to a huge challenge when it comes to training candidates for special missions. In one United States Air Force Combat Rescue Officer
Training, 70% of washouts occurred in the physically intensive development part of the course.
It has also been reported that 20% of all program washouts occur during day 5 pass/fail physical fitness test (Chappelle, 2015). The washout rate /high injury rate cost the military millions of dollars every year, therefore, being able to predict response to various physical training protocols, assess peak performance, and avoiding overtraining and injury is a necessity
(Bamman & Broderick, 2018).
The project called PHITE (Precision High Intensity Training through Epigenetics) is a multi-year project investigating relevant genotype contribution and the changes in DNA methylation, chromatin/histones, miRNA, and nuclear transcription factors to changes in performance from physical training. The focus of this thesis is to determine the genetic contribution to physical
1 | P a g e
fitness in the first 42 completers of the PHITE training group. By combining the genotyping data with the training response data, it was hypothesized that the training response would be different depending on the training group assigned and that genotype can be used to develop a predictor for training response. There were two main aims that were investigated for this hypothesis: the performance difference between the high and moderate intensity; and whether genotype could be used as a predictor of this training response.
B. Genotyping Using Single Nucleotide Polymorphisms
As previously mentioned, single nucleotide polymorphisms or SNPs are common genetic markers that occur naturally and are the most common type of genetic variation among individuals. SNPs are widely used as genetic markers for a variety of studies and are great tools to use to determine genetic contribution for physical fitness. Polymorphisms potentially associated with physical fitness and/or response to physical training were gathered from a review of the literature. Many SNPs come from Bray et al. 2009 (Bray MS1, 2009) , Yoo et al.
2016 (Yoo J, 2016), Timmons et al. 2010 (Timmons, 2009), and Rankinen et al. 2012 (Rankinen T
S. Y., 2012), as well as many other studies as indicated in Table S4- S10. During this review, SNPs were excluded if they:
• Came from a study with an N of 1, or only in a single family
• Were associated with severe congenital phenotypes
• Were associated with diseases such as ALS, cystic fibrosis, or other diseases that directly
impact physical capability
• Were discounted by follow up studies
2 | P a g e
• Had an unknown minor allele frequency
An additional ten SNPs were chosen within the genes encoding proteins responsible for epigenetic modification. While these had not been demonstrated to play roles in physical performance, they are related to the epigenetic portion of PHITE; several of these have been demonstrated to affect protein function and therefore global DNA methylation (e.g. rs2276731 in FTHFD (Min-Ae Song, 2016) or rs1801131 and rs1801133 in MTHFR (Castro, 2004). A total of
181 SNPs and one base pair insertion/deletion were chosen and classified based on what physical attributes each affect in association with physical fitness. The beneficial alleles were also determined during this literature review and are indicated in Table S4-10. In some cases, the heterozygous genotype is indicated instead of a specific allele, this is due to either allele being beneficial to physical training response. Table 1 shows a summary of all the SNPs chosen for this study.
Table 1: Summary of SNPs Chosen for PHITE
SNPs Associated with N= Aerobic Capacity/Endurance/VO2max 35 Physical Strength/Response to Training/Muscle Volume 28 Blood Pressure/Cardiovascular/Heart Rate 69 Increased BMI/Obesity/Diabetes 30 Endurance/Strength 2 Other (Bone Density/Height/Metabolism) 8 Epigenetic Modifiers 10 Total 182
1. Aerobic Capacity/Endurance/VO2max
One of the major contributing factors to physical fitness is aerobic capacity. Aerobic capacity is most universally measured by using the maximal rate of oxygen consumption (VO2max).
3 | P a g e
VO2max is defined as the oxygen uptake attained during maximal exercise intensity that could not be increased despite increase in exercise load which defines the limits of the cardiorespiratory system (Megan Hawkins, 2007). VO2max is the gold standard measure of aerobic endurance and cardiorespiratory fitness (Yoo J, 2016). A group of SNPs were identified to be associated with aerobic capacity, endurance and increases in VO2max.
2. Physical Strength/Response to Training/Muscle Volume
A large contributor to physical fitness variation is physical/muscle strength and overall response to physical training. Muscle strength has been measured in various studies by using peak muscle torque, or the maximum force applied by the muscles through a “moment arm”
(calculation associated with the force generated by the muscle and the length of the limb or muscle group applying the force) and at a given angle to the joint associated (Yoo J, 2016). A group of SNPs were identified to be associated with physical strength, response to training, and muscle volume.
3. Blood Pressure/Cardiovascular/Heart Rate
Cardiovascular health is closely associated with physical fitness and ability to respond to physical training. It has been shown that with regular physical activity there are beneficial changes in cardiac function. These changes have been commonly measured using heart rate and blood pressure as physiological indicators of general cardiac health (Rankinen T S. Y., 2012). A group of SNPs were identified to be associated with physical strength, response to training, and muscle volume.
4 | P a g e
4. Increased BMI/Obesity/Diabetes
Increased BMI and Obesity is associated with increased risk of diabetes, hypertension, and heart disease. Genetic markers have been reported to be associated with increased BMI and resistance to weight loss and have been found to negatively impact response to physical training (Jiang Y1, 2004).
5. Endurance/Strength
Most SNPs have one associated beneficial allele for physical fitness, but some SNPs variations both attribute to physical fitness. One of the most studied examples of this is the insertion/deletion in the angiotensin converting enzyme I (ACE). There is a 287-base pair insertion (I) or deletion (D) in intron 16 that has been strongly connected to elite performance in various sprint/power sports. The ACE protein is involved in the regulation of blood pressure by converting angiotensin I into angiotensin II, which is a potent vasoconstrictor and a muscle growth factor involved in overload-induced muscle hypertrophy (Ioannis D. Papadimitriou,
2016). Studies have shown that the D allele is associated with high fiber mean cross sectional area, which has been associated with power/sprint oriented events, whereas the I allele is associated with a greater density of mitochondria, which has been associated with endurance oriented events (Flück M, 2019).
Another well studied gene is ACTN3, which encodes for the sarcomere protein α-actinin-3 in skeletal muscle fibers. There is a change from a C to a T that changes an arginine (R) at position
577 to a premature stop codon (X) in the ACTN3 gene. The ACTN3 protein is a structural protein found exclusively in the fast type II muscle fibers used during power activities. Studies have
5 | P a g e
shown that the R allele is associated with an increase in fast type II muscle fibers, which has been associated with power/sprint-oriented events, whereas the X allele is associated with an increase in oxidative muscle phenotype (muscle fiber types I and IIA) which are responsible for postural and endurance-oriented events (Ioannis D. Papadimitriou, 2016).
6. Other (Bone Density/Height/Metabolism)
There are many contributing genetic factors associated with physical fitness. SNPs associated with more of the obvious contributions have been discussed, including cardiovascular factors, muscle strength/volume, BMI/obesity, and others. There are other subtler factors associated with physical fitness, such as height, bone density, and metabolism. These were included in an attempt to cover all aspects of physical training response.
7. Epigenetic Modifiers
Along with the genetic marker contribution, epigenetics contributes to an individual’s overall physical fitness. Epigenetics is defined as changes in organisms caused by modification of gene expression rather than alteration of the genetic code itself. The proteins that modify gene expression are called epigenetic modifiers and include histone deacetylases, DNA methylases, and histone acetyltransferases. These SNPs have never been shown to be associated with physical performance, which is why the beneficial allele is marked n/a in Table S10, and if associated they could be novel candidates for physical fitness response and were also included to tie into the epigenetic section of the project.
6 | P a g e
II. MATERIALS AND METHODS
A. Subjects and Training Protocol
The PHITE participant protocol and recruitment is being completed by The University of
Alabama at Birmingham. The project is set up to include 150 men and women, between the ages of 18-27 years old, recruited from the University of Alabama at Birmingham, which is located close to the Sumpter Smith Air Nation Guard Base (Birmingham, AL). Currently there are 42 completed participants that will be included in the results. Each participant was put at random into one of two 12-week training programs: moderate intensity or high intensity. These training groups are blind, so they are coded group A (n=22) and group B (n=20) and it is not known which group is which training protocol. Figure 1 shows the difference between the two groups and details associated with each. Peripheral blood draws and muscle biopsies were taken before, during and after the training protocols were completed. The blood was used for
DNA extraction and SNP genotyping. Many physical performance parameters were tested at defined intervals, including before and after the 12 weeks. For all the analysis, the change from
0 to 12 weeks was used for each performance parameter. Table 2 contains the 16 physical performance parameters used during the analysis of the first 42 completers.
7 | P a g e
Figure 1: The Moderate and High Intensity Training Protocols
Differences in the two training groups (group A and group B); Right side shows-High Intesity, Left side-Moderate Intesity
8 | P a g e
Table 2: Physical Performance Parameters (% Change 0-12 Weeks)
Physical Performance Test Definition V02peak (ml/kg/min) Numerical measurement of your body’s ability to consume oxygen Anaerobic Peak Power (AP) Represents the highest mechanical power generated during any 3- 5 second interval of the test Relative Peak Power (RP) Division of AP by body mass Squat Strength exercise focused on developing lower body muscles Chest Press Strength exercise focused on the pectoral muscles Knee Extension Resistance exercise focused on the thigh muscles Overhead Press Strength exercise focused on the shoulders and arms Lat Pull Strength exercise focused on the latissimus dorsi muscle Maximum Voluntary Contraction (MVC) Standardized method for the measurement of muscle strength Knee Peak Power (PP) Maximum power that the knee can sustain for a short time Vertical Jump Measurement of how high an individual can elevate off the ground from a standstill Thigh Lean Mass, Both Legs (grams)* The amount of weight on the thighs that is not fat Arm Lean Mass (grams* The amount of weight on the arms that is not fat Total Lean Mass (grams)* The amount of weight from the entire body that is not fat % Bodyfat* Total mass of fat divided by the total body mass of an individual Visceral Fat Mass (grams)* Amount of body fat that is stored within the abdominal cavity *Not measured as a percent change due to minimum changes from 0-12 weeks; all lean/visceral mass categories are change in grams and the bodyfat is the % bodyfat change.
These physical parameters were used as the phenotypic data in all the analyses with separating
groups A versus B and comparison with the genotype data.
B. DNA Isolation and Genotyping Analysis
According to manufactures’ instructions, DNA was extracted and purified from peripheral blood
using a GeneJet Whole Blood Genomic DNA Purification Mini Kit (ThermoFisher Scientific). The
quantity and quality of the DNA extracted was determined using a Nanodrop
spectrophotometer. Genotyping was performed with a Taqman Assay allele discrimination
assay using standard thermal cycling conditions as defined by the manufacturer
(ThermoFisher). The genotypes were determined with a QuantStudio 7 Flex (Applied
Biosystems) quantitative polymerase chain reaction (qPCR) system by measuring
9 | P a g e
the fluorescence signals VIC, FAM, or both in each well, which corresponds to heterozygous genotype. Taqman Genotyper Software (Applied Biosystems) was used for genotype determination and verification (Supplementary Tables 1-3 contain the raw genotyping data for the first 42 completers).
C. Beneficial Allele Score
Beneficial allele classifications were gathered from the literature for all relevant SNPs described, and a beneficial allele score was calculated (Equation 1).
Equation 1: Example of How Beneficial Allele Score was Calculated
SNP ID Hetero Beneficial Person 1 Person 2 Person 3 SNP 1 A/G A A/A (2) G/G (0) G/G (0) SNP 2 C/T T T/T (2) C/C (0) T/T (2) SNP 3 G/A G G/A (1) G/G (2) G/A (1) Beneficial Allele Score 5 2 3
In general, homozygous genotypes consisting of alleles beneficial to physical fitness were coded as 2, homozygous genotypes consisting of alleles harmful/neutral to physical fitness were coded as 0, and heterozygous genotypes were coded as 1. The SNPs associated with strength/endurance genotypes (Table S8) and the epigenetic modifiers (Table S10) were not included in this analysis. If either allele is marked as beneficial, then all of genotypes were scored 1. In the case of SNP rs1799945, the beneficial genotype is the heterozygous C/G, in this case the heterozygous was coded as 2 and both homozygous genotypes were coded as 1.
10 | P a g e
D. Statistical Analysis
1. Chi Squared Analysis
For genotyping, a population is in Hardy-Weinberg Equilibrium (HWE) if the allele and genotype frequencies follow the following equation:
Equation 2: Hardy-Weinberg Equilibrium Equation
푝2 + 2푝푞 + 푞2 = 1
Determining whether the population of participants in PHITE is in HWE was completed using the chi squared (χ2) statistic and the known minor allele frequencies for each of the SNPs. Chi squared is used to measure how expectations compare to actual observed data. For the expected results, the minor allele frequency (MAF) from dbSNP was used for each of the 182
SNPs used. It is known that ethnicity can greatly impact genotype frequency; therefore, ethnicity was corrected for in the MAF. Figure 2 represents the ethnicity information of the 42 completed participants. The gnomAD-Exomes MAF for Global was used for the White population (73.8% of PHITE subjects) and the gnomAD-Exomes Asian MAF was used for the
Asian population (16.7%). Using the reported percentages of each in the 42 completers, the corrected MAF was calculated (Equation 3) for each SNP and used for the chi squared calculation.
Equation 3: Equation for the Corrected Minor Allele Frequency Calculation (cMAF)
퐶표푟푟푒푐푡푒푑 푀푖푛표푟 퐴푙푙푒푙푒 퐹푟푒푞푢푒푛푐푦 (푐푀퐴퐹) = (푀퐴퐹 퐺푙표푏푎푙 ∗ 0.738) + (푀퐴퐹 퐴푠푖푎푛 ∗ 0.167)
11 | P a g e
Figure 2: Ethnicity Data for the First 42 Completers
Pie chart showing the ethnicity data for the first 42 completers; purple-white (73.8%), blue-Asian (16.7%), red-Black or African American (7.1%), green-multiple races (2.4%)
12 | P a g e
2. K-Means Clustering and Performance Score
K-means clustering is an unsupervised machine learning algorithm that groups individuals in a population together by similarities. This method was used according to (Yoo J, 2016) to group individuals in each of the 16 performance tests into 2 clusters: medium responders and high responders. Any performance change that is equal to 0 or less than 0 was classified into another cluster: non-responder. These clusters were also used to determine a performance score for each individual based on their cluster categorization for each physical performance test. This performance score was calculated by doing the following: each non-responder categorization was given a 0, medium responder was given a 1, and high responder was given a score of 2. All the scores for the 16 physical performance parameters were added up and make up the performance score.
3. Fisher’s Exact Test
The fisher’s exact test is a statistical analysis used to determine if there is a nonrandom association between two categorical variables. This test is similar to the chi squared analysis test but is more accurate when there is a small sample size. For the Fisher’s exact test in this study, two 2 by 2 contingency tables are used to compare the observed values of alleles in two sub groups and the calculated allele expected values for both groups (McDonald, 2014). The groups of SNP alleles compared in this study are between training groups A and B, and the K means clustered determined Non and High Responders groups for each of the physical training responses.
13 | P a g e
4. Analysis of Variance (ANOVA)
To compare the physical performance tests with each of the SNPs genotypes, an analysis of variance (ANOVA) was performed. ANOVA is a statistical technique that is used to check if the means of two or more groups, in this case genotypes, are significantly different from one another. Each SNP genotype was plotted against the performance parameters and ANOVA was used to determine if any one genotype is associated with higher performance in the 16 training responses.
5. Bonferroni Correction
When testing multiple hypotheses at once, there is an increased probability of observing at least one significant result due to chance. In this case, the Bonferroni correction is used to decrease the chances of obtaining false-positive results when multiple tests are performed on a single data set (Goldman, 2008). The Bonferroni correction is simple: original p- value (0.05) divided by the number of tests performed (182 for the 182 SNPs in this study). This gives the new significance value of 0.0003 for any of the analyses in this study with multiple tests run.
6. Genome Wide Association Study (GWAS)
A common way to investigate the association of genetic variants to specific complex phenotype is a genome wide association study or GWAS. This is an observational study tool that can study hundreds or thousands of SNPs for the determination of alleles that correlate to different complex diseases or traits. This was done using the linear regression model that models
14 | P a g e
relationships between two variables by fitting a linear equation to the observed data. These determined associations are then plotted: each SNP is plotted against the negative log scale of the p-values. This kind of plot shows the association of each SNP clearly; the higher the point on the plot, the greater the association (lower the p-value) In this study, the 182 SNPs were compared to each of the 16 performance responses for association.
III. RESULTS
A. AIM I. Training Group A Versus B Physical Performance Parameters
1. Group A Versus B: Individual Performance Parameters
The first analysis completed was simply comparing each of the 16 physical performance parameters described in Table 2 against the blinded training groups A and B (a moderate intensity and high intensity). Significance in training response was determined by ANOVA. In this analysis, 15 of the 16 physical performance training responses did not significantly differ between training group A and B. Arm Lean Mass % Change was the only single physical parameter that had an overall higher response in training group A than training group B (shown in Figure 3).
1. Group A Versus B: Performance Score
Along with determining the differences for each individual physical performance parameter, a comparison of physical performance for each of the 42 completers was done. Determining overall performance was done by using the performance score determined by the performance clusters created using k-means clustering. Completers were separated by A or B training group
15 | P a g e
and the difference in performance score was found to be significantly different (Figure 4). Both results together show that even though the physical performance parameters individually did not significantly differ between the completers of training group A versus B (expect in the Arm
Lean Mass % Change), in overall measure of performance, training group B completers were significantly more successful than the completers of training group A. Genetic variation in the completers of both groups also must be determined to see if the genetic contribution is significant to the performance response differences.
16 | P a g e
Group A Group B
N = 22 N = 20
*
12 12 (grams)
-
ean Mass ean LM Change 0 Arm L Arm
Figure 3: Group A Versus Group B Comparison for Arm Lean Mass % Change
Box and whisker plot separating group A versus group B for Arm Lean Mass % Change response. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points, * p value ˂ 0.05
17 | P a g e
Group A Group B
N = 22 N = 20 **
PerformanceScore
Figure 4: A Versus B Training Group Comparison for Performance Score
Box and whisker plot separating group A versus group B for Performance Score. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points; ** p value ˂ 0.01
18 | P a g e
B. AIM II. Investigate if genotype can be used to predict phenotype in individual training responses and in overall training response
To investigate a usable genotype/allele predictor for the general population, the population of completers for PHITE had to be analyzed. To do this, chi squared analysis was completed to determine if any SNPs were out of equilibrium using an ethnicity-corrected minor allele frequency for comparison. Table 3 shows the results of this analysis and 14 SNPs had significant p values.
Table 3: Chi Squared Analysis of all 42 Completers Gene(s) SNP ID Minor Allele Corrected MAF Chi squared p-value FAT2 rs10072841 C 0.38359468 6.21403 1.27E-02 (intergenic) rs17231506 T 0.25437562 4.58778 3.22E-02 (intergenic) rs1799768 - 0.247065 8.01014 4.70E-03 CETP rs1800775 C 0.43784954 32.36959 0.00E+00 LIPG rs2000813 T 0.2265238 4.92748 2.64E-02 FTHFD rs2002287 G 0.3373774 3.86362 4.93E-02 ESR1 rs2234693 C 0.4144992 12.46752 4.00E-04 NFKB1 rs28362491 * 0.3862216 4.91928 2.66E-02 EPAS1 rs4035887 A 0.3199796 6.71720 9.50E-03 TYMS rs502396 T 0.3848382 7.46675 6.30E-03 (intergenic) rs5082 G 0.2567174 8.20291 4.20E-03 (intergenic) rs699947 A 0.32672976 5.29288 2.14E-02 GNAS rs7121 C 0.40546854 5.27060 2.17E-02 TCERG1l rs7893895 C 0.4336136 5.32885 2.10E-02
This means that those SNPs genotypes are out of equilibrium according to our corrected minor allele frequency and could lead to false positives of significance in the later analyses. It is for this reason any of the SNPs in Table 3 will not be considered if significant in any of the five- following predictor analyses described.
19 | P a g e
1. Beneficial Allele Score
The first analysis performed was to determine if the beneficial allele score calculated for each completer could be used to determine the k-means clustering group (High, Medium or Non-
Response) for individual performance of each of the 16 tests in Table 2. To do this, each of the performance response k-means clustering groups were plotted against the beneficial allele score. Figure 5 shows the beneficial allele score versus the clusters for the vertical jump categorization. The means of each cluster for beneficial allele score increase with the correct clusters (the highest median is the high response group) but this change was not determined to be significant (p=0.2405) using ANOVA. In the other 15 training responses, similar trends were seen but none were statistically significant. This shows that, with the current completers in consideration, the beneficial allele score is not able to correctly determine the k-means clustering groups.
20 | P a g e
N=17
N=10
N=15 BeneficialAllele Score
Figure 5: Beneficial Allele Score versus Vertical Jump Categorization
Box and whisker plot of beneficial allele score versus the k- means clustering categorizations for vertical jump. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points. The means increase with the clusters; the highest median for beneficial allele is in the high responder group for this trend is not significant, p= 0.2405
21 | P a g e
2. Performance Score
Using the performance score calculated using the k-means clustering groups (Non, Medium, and High-Responders) the next strategy was to determine whether any individual SNP genotype could predict overall performance. This was done by plotting the performance score against each of the 182 SNP genotypes and using ANOVA to determine significance. This analysis resulted in 6 SNPs that had a specific genotype associated with higher/lower performance score and are shown in Table 4.
Table 4: SNPs Shown to be Associated with Performance Score SNP ID Gene P value rs11096663 PUM2 | RHOB 0.0316 rs1546570 DIS3L 0.0063 rs1801394 MTRR 0.0162 rs2276731 FTHFD 0.0360 rs328 LPL 0.0182 rs7964046 PDE3A 0.0302
The Bonferroni correction was applied to the ANOVA analysis, which resulted in having no SNPs significantly associated with performance score. For further investigation, the p value of 0.05 will be used for the performance score and the 6 SNPs in Table 4 will be further characterized.
Each of the six SNPs were investigated further and plotted for determination of the beneficial allele associated with the performance score.
The first SNP is rs11096663, an A/G SNP in the intergenic region between gene PUM2 (Pumilio homolog 2) and RHOB (Ras Homolog B) in chromosome 2. In the literature this SNP has been shown to be a predictor of high response in VO2max (Yoo J, 2016). Figure 6 shows the genotypes for rs11096663 plotted against the performance score. The box and whisker plot
22 | P a g e
show that the G/G genotype is associated with higher overall performance and is significantly different from the A/G and A/A genotypes.
Without considering the outliers in the A/A genotype, the median is higher than the heterozygote, but the A/A genotype did also contain the lowest performance score outliers compared to either of the other genotypes.
SNP rs1546570 is an A/C intron variant in the DIS3L (DIS3 like exonuclease 1). In the literature, this SNP has been shown to be a predictor of high response in VO2max (Timmons, 2009). Figure
7 shows the genotypes for rs1546570 plotted against the performance score. The more beneficial genotype is A/C with the highest median, and the C allele is more beneficial than the
A allele significantly with a p=0.0063. This association could be skewed due to only three A/A genotypes in the population; more completers could help strength this associated with overall performance.
23 | P a g e
Performance Score vs. rs11096663 (PUM2 | RHOB)
* N=7 * N=17 N=18
PerformanceScore
Figure 6: SNP rs11096663 genotypes versus the Performance Score
Box and whisker plot of Performance Score versus rs11096663 a SNP in the intergenic region between genes PUM2 and RHOB. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points; * p value ˂ 0.05
24 | P a g e
Performance Score vs. rs1546570 (DIS3L) * N=13
** N=26
N=3
PerformanceScore
Figure 7: SNP rs1546570 genotypes versus the Performance Score
Box and whisker plot of Performance Score versus rs1546570. Indicated are the median, upper, and lower quartiles and minimum and maximum values ; *p value ˂ 0.05 and ** p value ˂ 0.01
25 | P a g e
The SNP rs1801394 is an A/G missense variant in the MTRR (5-methyltetrahydrofolate- homocysteine methyltransferase reductase). When the A allele is present, isoleucine at position
22 and when the G allele is present, there is a switch from isoleucine to methionine. The change to methionine doesn’t create a non-functional protein, but it does have a lower affinity for its target and affects the body’s homocysteine levels (Olteanu H, 2002). In the literature, this SNP has been shown to be a predictor of lower BMI and decrease chance of obesity (Bokor
S, 2011). Figure 8 shows the genotypes for rs1801394 plotted against the performance score. In the first 42 completers, the A/G genotype and the A allele are shown to be more beneficial to overall performance. The outliers also confirm this conclusion since the lower performance scores fall into the G allele genotypes.
Another SNP identified to associated with overall performance is rs2276731. This SNP is a C/T intron variant in the FTHFD (aldehyde dehydrogenase 1 family member L1) gene. This gene encodes a protein involved in one carbon metabolism and is an epigenetic modifier that plays a vital role in DNA methylation (Min-Ae Song, 2016). This SNP has never been shown to be a predictor of any physical training responses and could be a novel association.
26 | P a g e
Performance Score vs. rs1801394 (MTRR)
N=12 **
N=21
N=9 PerformanceScore
Figure 8: SNP rs1801394 genotypes versus the Performance Score
Box and whisker plot of Performance Score versus rs1801394. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points; ** p value ˂ 0.01
27 | P a g e
Figure 9 shows the genotypes for rs2276731 plotted against the performance score. In this case, the C/C genotype seems to be the beneficial for overall performance, but in the first 42 completers there was only one with the genotype C/C for this SNP. This makes it difficult to determine if the C/C genotype is associated with higher performance or if this is a false positive association. The addition of more completers could be very beneficial in determining the extent of this novel SNP association with overall physical training response. The SNP rs328 is a C/G stop gained variant in the LPL (lipoprotein lipase) gene. The C allele sequence creates a full- length protein, whereas the G allele creates a premature stop codon, leading to a shortened transcript. This short variant creates low activity LPL that can lead to and increased chance in insulin resistance. In the literature, this SNP has been shown to be a predictor of increased BMI and an increased chance of obesity (Zhang S, 2015).
28 | P a g e
Performance Score vs. rs2276731 (FTHFD)
* N=32
N=1
N=9
PerformanceScore
Figure 9: SNP rs2276731 genotypes versus the Performance Score
Box and whisker plot of Performance Score versus rs2276731 Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points; * p value ˂ 0.05
29 | P a g e
Figure 10 shows the genotypes for rs328 plotted against performance score. This shows that the beneficial allele is most likely the G allele due to the increase in performance score seen between the C/C and the C/G box plots. Without any completers that have the G/G genotype, it is hard to confirm this association, so increasing the sample size for genotyping would be beneficial.
The last SNP that was shown to be a predictor of overall performance is rs7964046. This SNP is a C/T intron variant in PDE3A (phosphodiesterase 3A) gene. In the literature, this SNP has been shown to be a predictor of a lower heart rate during exercise (Rankinen T S. Y., 2012). Figure 11 shows the genotypes for rs7964046 plotted against performance score. This figure shows that the T/T genotype is clearly beneficial for overall performance response compared to the other genotypes. In conclusion, all six SNPs determined by the performance score to be significantly associated with overall performance have specific genotypes that could be used as part of a model for predicting training response.
The performance score was also plotted against the beneficial allele score to determine if beneficial allele increases with the overall performance. This was done by doing a linear regression between the beneficial allele score and performance score (Figure 12). It was not found to be a significant association, with the R value = 0.027, but is trending in the right direction, which could become more pronounced with a greater number of completers in this study.
30 | P a g e
Performance Score vs. rs328 (LPL) ** N=7
N=35
PerformanceScore
Figure 10: SNP rs328 genotypes versus the Performance Score
Box and whisker plot of Performance Score versus rs328. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points; ** p value ˂ 0.01
31 | P a g e
Performance Score vs. rs7964046 (PDE3A)
* N=17
N=17
N=8
PerformanceScore
Figure 11: SNP rs7964046 genotypes versus the Performance Score
Box and whisker plot of Performance Score versus rs7964046. Indicated are the median, upper, and lower quartiles, minimum and maximum values, and outliers (which are values 1.5 times the difference between the upper and lower quartiles indicated as points; * p value ˂ 0.05
32 | P a g e
lAllele Score Beneficia
Performance Score
Figure 12: Beneficial Allele Score Versus Performance Score
A linear regression plot of beneficial allele score versus performance score with a best of fit line plotted in red (R2 = 0.027) R value is explained variation / total variation
33 | P a g e
3. Fisher’s Exact Test
The fisher’s exact test was completed to compare the number of alleles for each of the SNPs between two categorical variables. This was completed between groups A and B and between the k-means clustered categories High Responders and Non-Responders. The was completed by comparing the observed number of alleles in each category and comparing them to calculated expected alleles. This analysis determined 83 SNPs that were significantly different than what was expected in either of the two categories with a p-value of less than 0.05. When this data set used the Bonferroni correction, three SNPs were still significant: rs11622895 (EPAS1), rs1801394 (MTRR), and rs2979481 (RBPMS). These SNPs were plotted against the Performance categories they were found to be associated with to determine which allele was associated with which k-means clustered group (high or non-responders). Figures 13-15 show these results and
Table 5 summarizes the associations.
Table 5: Bonferroni Corrected Fisher’s Exact Test SNP Results
SNP ID Gene Performance Results P value rs11622895 EPAS1 Knee Peak Power T (High Response) 0.0002 rs1801394 MTRR Lat Pull A (High Response) 0.0003 rs2979481 RBPMS Knee Ext C (High Response) 0.0003
This analysis results in 3 SNPs that have alleles associated with high performance for specific training responses that could be used in a predictor for physical training response.
34 | P a g e
Non-Responders High-Responders Knee Peak Power % Change Knee Peak Power % Change rs11689011 – EPAS1 rs11689011 – EPAS1
C T C 33% 50% 50% T 67%
Figure 13: Non-Responders Versus High-Responders for rs11689011 – EPAS1 for Knee Peak Power % Change
Response group allele distributions for SNPs identified as significant by Fisher’s exact test, p value ˂0.0003
35 | P a g e
Non-Responders High-Responders Lat Pull % Change Lat Pull % Change rs1801394 - MTRR rs1801394 - MTRR
G A 36% 43% A G 64% 57%
Figure 14: Non-Responders Versus High-Responders for rs1801394 – MTRR for Lat Pull % Change
Response group allele distributions for SNPs identified as significant by Fisher’s exact test, p value ˂0.0003
36 | P a g e
Non-Responders High-Responders
Knee Extension % Change Knee Extension % Change rs2979481 – RBPMS rs2979481 – RBPMS
C C 30% 37%
T T 70% 63%
Figure 15: Non-Responders Versus High-Responders for rs2979481 – RBPMS for Knee Extension% Change
Response group allele distributions for SNPs identified as significant by Fisher’s exact test, p value ˂0.0003
37 | P a g e
4. ANOVA Analysis
ANOVA analysis is an easy way of comparing categorical data to continuous data. This was done to determine whether an individual SNP genotype could be used as a predictor of each of the 16 training responses. In this case, each of the SNP genotypes (182 SNPs) were plotted against each of the training responses (16 different tests). This analysis lead to 183 times a SNP was predictive of one of the training responses (see supplemental tables S11-S26 for all the SNPs, the associated training response, and the p-value determined by the ANOVA). Due to the large number of tests run on a single data set, the
Bonferroni correction was used to attempt to avoid any false positive associations. Using the new significant p-value of 0.0003, there were only 5 times a SNP was predictive of one of the training responses and these are described in Table 6.
Table 6: SNPs Shown to Predict Specific Training Responses by ANOVA
SNP ID Gene Performance P value rs10248479 TFEC AP, RP 0.0002, 0.0002 rs1532723 ANO4 Squat 0.0001 rs4994 ADRB3 Vertical Jump 0.0001 rs7386139 DEPDC6 Chest Press 0.0002
Figures 16-20 show the genotypes plotted for the 5 SNPs associated with higher performance in one of the training responses by ANOVA analysis. In conclusion, these four SNPs determined by the ANOVA analysis to be significantly associated with specific training responses could be used in a predictor for the associated training response.
38 | P a g e
Anaerobic Peak Power % Chg 0-12 vs. rs10248479 (TFEC) * N=1 **
**
12 - N=5
N=36 Anaerobic Peak Anaerobic Power Chg % 0
Figure 16: SNP rs10248479 Association with Anaerobic Peak Power % Chg 0-12 By ANOVA
Box and whisker plot of AP versus rs10248479 genotypes. Indicated are the median, upper, and lower quartiles and minimum and maximum values and the number of people in each group as indicated above the box and whisker plots; *p value ˂ 0.05
** p value ˂ 0.01
39 | P a g e
Relative Peak Power % Chg 0-12 vs. rs10248479 (TFEC) N=1 * **
**
12 - N=5
N=36 RelativePower Peak % Chg0
Figure 17: SNP rs10248479 Association with Relative Peak Power % Chg 0-12 By ANOVA
Box and whisker plot of RP versus rs10248479 genotypes. Indicated are the median, upper, and lower quartiles and minimum and maximum values and the number of people in each group as indicated above the box and whisker plots; *p value ˂ 0.05, ** p value ˂ 0.01
40 | P a g e
Squat % Chg 0-12 vs. rs1532723 (ANO4)
*
N=40
12 -
SquatChg % 0 N=2
Figure 18: SNP rs1532723 Association with Squat % Chg 0-12 By ANOVA
Box and whisker plot of RP versus rs1532723 genotypes. Indicated are the median, upper, and lower quartiles and minimum and maximum values and the number of people in each group as indicated above the box and whisker plots; *p value ˂ 0.05
41 | P a g e
Vertical Jump % Chg 0-12 vs. rs4994 (ADRB3)
12 ** - N=39
N=3
Jump Vertical % Chg 0
Figure 19: SNP rs4994 Association with Vertical Jump % Chg 0-12 By ANOVA
Box and whisker plot of RP versus rs4994 genotypes. Indicated are the median, upper, and lower quartiles and minimum and maximum values and the number of people in each group as indicated above the box and whisker plots; **p value ˂ 0.01
42 | P a g e
Chest Press % Chg 0-12 vs. rs7386139 (DEPDC6)
** N=2 **
12 -
N=33
ChestPress Chg % 0 N=7
Figure 20: SNP rs7386139 Association with Chest Press % Chg 0-12 By ANOVA
Box and whisker plot of RP versus rs7386139 genotypes. Indicated are the median, upper, and lower quartiles and minimum and maximum values and the number of people in each group as indicated above the box and whisker plots; **p value ˂ 0.01
43 | P a g e
5. GWAS Analysis
An alternate technique to the ANOVA analysis is an observational test called the GWAS. This test was also used to determine if an individual SNP could be used in a predictor to one of the 16 training responses. The Bonferroni correction was also used for this test, so the p-value for association was
0.0003. Figure 21 shows the Manhattan plots for the training responses that have SNPs with significant associations seen above the solid red lines on the plots. Table 7 summarizes the SNPs found with significant associations and the training responses that associated with each.
Table 7: SNPs Shown to Predict Specific Training Responses by GWAS
SNP ID Gene Performance P Value rs1044498 ENPP1 Squat 0.0003 rs1800562 (intergenic) Lat Pull 0.00025 rs1950902 MTHFD1 Arm LM 0.00028 rs2228059 IL15RA Lat Pull 0.00026 rs3025684 CREBBP Squat 0.00027 rs3770991 NRP2 Arm LM 0.00025 rs8192678 PPARGC1A AP, RP 0.00001, 0.00001 rs9340799 ESR1 Arm LM 0.00025
In conclusion, all eight SNPs determined by the GWAS analysis to be significantly associated with specific training responses could be used in a predictor for the associated training response.
44 | P a g e
A
(P)
10
log -
B C
D E
Figure 21: GWAS Plots for the Training Responses that had Significant SNP Associations
GWAS Manhattan Plots, where the orange line represents the Bonferroni corrected p value = 2.75x10-4; significant associations are labeled with dbSNP ID.
45 | P a g e
IV. DISCUSSION
Understanding the role that genetics play in physical training response is an important topic in military research. Investigating this connection and using it to create models of training response could save the military millions of dollars by determining risk of injury or inadequate physical fitness leading to dropouts. To investigate the genetic contribution to training response, genotyping of several SNPs previously identified to be associated with different parts of individual physical training response was completed for 42 individuals. These SNPs were chosen to cover a wide range of physical performance responses including aerobic capacity, endurance/strength, muscle volume, etc. (Table 2). Some SNPs that were chosen were also indicative of traits that would contribute negatively to physical performance such as increased
BMI or lower bone density. These SNP genotypes were compared to the physical training responses that were measured for each of the 42 completers of the blinded training groups
(high versus moderate intensity) and are described in Figure 1. Connecting the SNP genotypes to the physical training responses was used to investigate the differences in training groups and whether genotype can be used to determine phenotype (training response).
Due to the training group being blinded, it is not known which protocol each of the participants completed. It is for this reason that it was of great interest to investigate the performance response in each of the training groups: Group A versus Group B. This was first done by separating each of the 16 training responses by A and B and comparing the two. ANOVA was used to determine that there was no significant difference between the two, except in the case of the Arm Lean Mass Change (Figure 3). This shows that on an individual performance level,
46 | P a g e
the participants virtually responded the same no matter what training group was completed.
Next, overall performance was considered by using k-means clustering groups (non, medium, and high responders) for each of the 16 training responses to score overall physical performance (performance score). When the performance score was separated by group A and
B, there was a significant difference between the two groups (p=0.02). Group B had a much higher overall performance score than group A (Figure 4). This shows that even though individual phenotype responses were similar for completers in both groups, the overall performance of the participants was much greater if they completed training group B than group A. Due to the greater capacity to see larger increases in training response in males compared to females, the gender composition for each group was investigated. It was found that the gender make-up of each group wasn’t significantly different with group A having 13 females and 9 males and group B having 15 females and 5 males. This shows that the difference in training response is not driven by a gender disequilibrium between the two groups. In conclusion, individual training responses did not differ between the two training groups significantly, but the overall performance of the completers was significantly different between the two groups, with group B having the higher performance.
Determining whether genotyping can be used as a predictor for training response was the next aim of this project. Using the 182 determined physical performance SNP genotypes, it was hypothesized that this could be used as a model of prediction for training response. To investigate this hypothesis, five different analyses were used to investigate whether genotype could be used as a predictor for training response: ANOVA of beneficial allele score, ANOVA of
47 | P a g e
performance score, Fisher’s exact test, ANOVA of individual performance results, and GWAS analysis
The beneficial allele score was determined using the literature-defined beneficial alleles. This beneficial allele score was calculated for each of the 42 completers and compared to the training response k-means categories by ANOVA. This was done to determine whether the beneficial allele score could predict whether a completer would fall into the k-means clustered determined non, medium, or high-responder performance categories for each of the training responses. It was found that the beneficial allele score could not significantly predict the performance response. Even though it was not found to be significant, the beneficial allele score did seem to trend in the right direction, with the higher beneficial allele scores falling into the high-responders. A future direction for this study would be to perform predication statistics, this would allow for a determination of accuracy of how many times the beneficial allele score can correctly determine high performance. This study currently has a rather small sample size and with more completers, this trend might become clearer and could be statistically significant. This result could also be re-evaluated by investigating the each of the
SNPs beneficial allele determination. Some of the literature determined beneficial alleles were difficult to classify and might be conflicting with the data gathered from this study. Another approach that can be taken in this case would be to determine the beneficial alleles according to the data gathered in this study and then use that corrected beneficial allele score moving forward with more completers. This could be a more accurate beneficial allele score that might be able to be used as a model for training response. Another future direction for the beneficial allele score could be to separate the SNPs into their respective physical fitness categories and
48 | P a g e
use separate beneficial allele scores for specific physical training responses. For example, take only the SNPs for aerobic capacity/endurance and calculate a beneficial allele score using those
SNPs. This would then be used to compare against VO2peak categories (non, medium, and high response) by seeing if it can predict which category a person would fall into. This could be done with multiple physical training responses and might be a better way to use the beneficial allele score as a predictor.
Another tactic used toward determining a model for training response was using the performance score. This was done by plotting the performance score by every individual SNP genotype. The associations were analyzed using ANOVA and any significantly associated genotypes were analyzed for positive or negative correlation. Using a p-value of 0.05, this analysis found 6 SNPs that had a specific genotype associated with performance score. In this analysis, it is acknowledged that if the Bonferroni correction was applied to this set of associations, then there were no significant SNPs associated with overall performance for the first 42 completers. For the 6 SNPs found, each SNP was analyzed, and the beneficial genotypes were determined due to the results of this study and compared to the literature determined beneficial alleles (Table 8).
Table 8: SNPs Beneficial Alleles Associated with Overall Performance
SNP ID Gene Beneficial Allele Literature Beneficial Allele rs11096663 PUM2 | RHOB G A rs1546570 DIS3L C A rs1801394 MTRR A G rs2276731 FTHFD C n/a rs328 LPL G C rs7964046 PDE3A T T
49 | P a g e
As seen in Figures 6-11 show each of the SNPs genotypes compared to the performance score. It is recognized that there is a large variation is performance score for each of the genotypes which causes a lot of overlap between the quartiles of the box and whisker plots. This separation could improve with more completers and make a clearer association. In Table 8, the beneficial alleles determined in our study do not match up with the beneficial alleles determined in the literature review. This could be due to the amount of information about these SNPs found in the current literature. Many of these SNPs were shown to be associated with physical fitness/training response but were not specific in the allele that was associated. If information was not provided in the literature for which allele was specifically associated, the minor allele was chosen. This might explain some of the discrepancies between this study’s determined beneficial alleles and the literature determined. A future direction for this part of the study could be to create a new beneficial allele categorization based on the first 42 completers training responses and use to this new classification against the rest of the completers in this study. This also could be used to investigate the beneficial allele score’s ability to predict which k-means clustering group each completer would fall into.
The performance score was also plotted against the beneficial allele score to determine with the beneficial allele score could predict overall performance (Figure 12). This linear regression best fit line had an R2 value of 0.027. The R2 value explains the amount of variance the line shows compared to the total variance. In a perfect correlation, the R2 value would be 1, so the closer the value is to 1 the better the line of best fit. In this case, the R2 value is very low which means the beneficial allele score does not clearly predict the overall performance in this study.
50 | P a g e
The trouble with the determined SNP associations for performance score is the small sample size in this study. This lead to a few SNPs’ minor alleles only having a couple completers, which could result in false positive or false negative associations. More completers would greatly increase the confidence of the 6 SNPs determined to be associated with overall performance score. In summary, the SNPs in Table 8 are associated with overall performance significantly
(p=0.05) and could be used in a predictor to overall training response performance, but additional subjects are needed to decrease the chance that these are false positive associations.
The fisher’s exact test model was used to investigate whether specific alleles for the 182 SNPs were associated with high or non-response in the 16 different training responses, group A versus group B, or overall performance scores high versus non-responders. This statistical analysis determined several SNPs that had alleles associated with several different training responses (Table 5). These alleles could potentially be used in a predictor for physical training response. The three SNP associations determined with the Bonferroni corrected p value
(0.0003) were just at the significance cutoff. An increase in the number of completers could increase the strength of some of the associations seen. The SNP rs2979481 association with high response for Knee Extension % Change 0-12 weeks had only 5 completers in the non- response category, which could make the C allele association be a false positive for the high response category. An increase in the number of completers could increase the strength of some of the associations seen and create new associations with some of the other 83 SNPs that had significance above the Bonferroni correction.
51 | P a g e
ANOVA analysis was another approach taken to determine a model for physical training response. This test was used to determine whether an individual SNP could predict an individual training response. This analysis determined 183 times a single SNP was predictive of a single training response. With a large number of tests being performed on the same data set, the Bonferroni correction was used to determine a more specific set of associations. The correction narrowed down the list to 5 matches (Table 6). Figures 16-20 show the genotypes for the 5 associated SNPs plotted against the appropriate training response. Several of SNPs only had one or two completers for the minor homozygote which could lead to false positive associations. For example, rs10248479 the G/G genotype showed a very high performance for
AP and RP, but there was only one completer that had this genotype. Due to this, the G/G genotype in this study is associated with high performance for AP and RP but with more completers, this association might be disproved. Further completers will give a clearer picture of the significant associations with the ANOVA analysis.
These SNPs could be used as part of a prediction model of specific training responses. An alternative test, called a GWAS, was also performed to determine association for each allele.
This test also used the Bonferroni correction to narrow down the significant associations to the strongest correlations (Table 7). This analysis also found significant associations between SNPs and specific training responses. These similar techniques were able to determine possible predictors of training response (genotypes and alleles) but each list of SNPs is unique. This difference is most likely due to the difference in calculations of significance between the two different analyses. The ANOVA uses multiple t-tests to determine if each genotype significantly differs from the other genotypes. The GWAS uses the SNP allele frequencies of two or more
52 | P a g e
separated groups (in this case the training response groups) to determine significance. This variation in calculation leads to different results in associations. Both the ANOVA and the GWAS analyses determined SNPs are good candidates for looking at individual training response predictors.
In summary, this study determined that the completers of the two different training protocols do not respond significantly different on individual phenotype responses, but overall performance is significantly greater in training group B compared to group A. Using multiple analysis, SNP performance associations were determined for individual training responses, such as chest press or lat pull, and for overall performance taking into consideration all the training responses; these are summarized in Table 9. All these SNPs are candidates for a genotypic model of training response and could be used to predict an individual’s success in a physically demanding situation, such as basic training, before the situation is started.
Table 9: Summary Table for all SNPs Found to be Possible Training Response Predictors
SNP ID Gene Performance Analysis P value rs11096663 PUM2 |RHOB Overall Performance Performance Score 0.0316
rs1546570 DIS3L Overall Performance Performance Score 0.0063 rs1801394 MTRR Overall Performance Performance Score 0.0162 Lat Pull Fisher’s Exact Test 0.0003 rs2276731 FTHFD Overall Performance Performance Score 0.0360 rs328 LPL Overall Performance Performance Score 0.0182 rs7964046 PDE3A Overall Performance Performance Score 0.0302 rs11622895 EPAS1 Knee Peak Power Fisher’s Exact Test 0.0002 rs2979481 RBPMS Knee Ext Fisher’s Exact Test 0.0003 rs10248479 TFEC AP, RP ANOVA 0.0002, 0.0002 rs1532723 ANO4 Squat ANOVA 0.0001 rs4994 ADRB3 Vertical Jump ANOVA 0.0001 rs7386139 DEPDC6 Chest Press ANOVA 0.0002 rs1044498 ENPP1 Squat GWAS 0.0003 rs1800562 (intergenic) Lat Pull GWAS 0.00025
53 | P a g e
SNP ID Gene Performance Analysis P value rs1950902 MTHFD1 Arm LM GWAS 0.00028 rs2228059 IL15RA Lat Pull GWAS 0.00026 rs3025684 CREBBP Squat GWAS 0.00027 rs3770991 NRP2 Arm LM GWAS 0.00025 rs8192678 PPARGC1A AP, RP GWAS 0.00001, 0.00001 rs9340799 ESR1 Arm LM GWAS 0.00025
Another aim of this project was to determine if genotype could predict whether an individual would perform better in different training groups: moderate intensity versus high intensity. Due to the number of completers being small, this could not be investigated further at this time.
When the current data were split into the training groups for genotypes, the power was not sufficient to determine true differences in each genotype for each SNP (data not shown). It was also determined that due to lower minor allele frequencies, when the data were split into group A versus group B, sometimes the minor alleles only showed up in one group or the other due to few completers having this genotype. This lead possible to false positives for minor alleles being associated with only one group or the other. This could be a possible future direction of this part of the study that could lead to model predictors of what type of training would be best for different individuals.
Overall this study shows the importance of considering genotype in any model of physical performance and moving forward should be considered in the PHITE project. In the end, the overall aim would be to incorporate the genotyping date and the epigenetic data together into a single model for physical training response.
54 | P a g e
V. FUTURE DIRECTIONS
The project PHITE is set to continue until the total completers is 150 people. Currently, there are only plans to continue the epigenetic side of the project. The genotyping using the blood samples drawn at time zero will only be done until the current supply of materials is used. The data presented in this thesis supports the importance in considering genotype in any model of physical training response. The candidate phenotype predictors found in AIM II of this project are important to pursue further and could be used as part of an overall physical training response model in the future. In general, the overall final aim of PHITE is to determine a physical training response model for different types of training. It is believed that the best way to complete this is by incorporate the genotyping data and the epigenetic data together into a single model of physical training response.
55 | P a g e
VI. SUPPLEMENTARY DATA
Supplementary Tables S1 through S3 contain the raw data for the first 42 completers. The genotypes are colored coded using the following: Major Homozygote, Heterozygous, and
Minor Homozygote.
Table S1: Raw Genotyping Data for the First 16 Completers
Group A A B B A A B B A A A B B B B A Gene(s) SNP ID 1 4 6 8 15 17 20 21 23 25 31 34 38 39 41 42 FAT2 rs10072841 MATN2 rs10099863 PDK4 rs10247649 TFEC rs10248479 ADRB2 rs1042713 ADRB2 rs1042714 ENPP1 rs1044498 TTN rs10497520 PPP1CC rs1050587 PPARD rs1053049 CYP2D6 rs1065852 PLA2G2A rs10732279 HBB rs10768683 (intergenic) rs10861553 AMN1 rs11051548 (intergenic) rs1107946 (intergenic) rs11096663 LEPR rs1137101 NME7 rs1138486 HIF1A rs11549465 ID3 rs11574 RNASE1 rs11622895 EPAS1 rs11689011 (intergenic) rs11770725 (intergenic) rs12079745 DIO1 rs12095080 (intergenic) rs12535708 (intergenic) rs12535747 GABPB1 rs12594956
56 | P a g e
KCNH7 rs12613181 TMEM195 rs12672644 YWHAQ rs12692388 API5 rs12789205 CREBBP rs130021 LOC100288831 rs13060995 HDAC2 rs13212283 LEP rs13245201 (intergenic) rs1341439 (intergenic) rs1349419 STS rs13648 LOC730100 rs1451462 ANO4 rs1532723 DIS3L rs1546570 GPRIN3 rs1560488 (intergenic) rs1570360 NPY rs16139 EPS1 rs17039192 LOC100129278 rs17044554 (intergenic) rs17231506 TTC17 rs17508783 AMPD1 rs17602729 (intergenic) rs17782313 (intergenic) rs1799768 FABP2 rs1799883 HFE rs1799945 COL1A1 rs1800012 CNTF rs1800169 PPARA rs1800234 TGFB1 rs1800470 HFE rs1800562 (intergenic) rs1800588 (intergenic) rs1800629 CETP rs1800775 LOC541472 rs1800795 UCP3 rs1800849 MTHFR rs1801131 MTHFR rs1801133 ADRB1 rs1801252 ADRB1 rs1801253 MTRR rs1801394 MYOSTAIN rs1805086 MTR rs1805087
57 | P a g e
ACTN3 rs1815739 SPATA17 rs1832544 EPAS1 rs1867785 PTPRT rs1885831 MTHFD1 rs1950902 LIPG rs2000813 FTHFD rs2002287 VEGFA rs2010963 PPARD rs2016520 p300 (EP300) rs20551 GCH1 rs2057368 NOS3 rs2070744 RAB6A rs2140892 IL15RA rs2228059 VDR rs2228570 ESR1 rs2234693 H19 rs2251375 REEP3 rs2252578 CREB1 rs2253206 DOCK2 rs2270895 FTHFD rs2276731 TFAM rs2306604 LRP5 rs2306862 HYDIN rs235987 CREB1 rs2360969 NRF1 rs2402970 SLC22A3 rs2457571 (intergenic) rs2542729 BTAF1 rs2792022 NFKB1 rs28362491 MYLK rs28497577 ACVR1B rs2854464 (intergenic) rs2854744 RBPMS rs2979481 CREBBP rs3025684 TOM1L2 rs3183702 LPL rs320 CHRM2 rs324640 LPL rs328 RYR2 rs34967813 NUAK1 rs3741886 ATP1B1 rs3766031 ATP1B1 rs3766044
58 | P a g e
NRP2 rs3770991 PDK4 rs3779478 GRHL1 rs3791749 EPAS1 rs4035887 TUSC1 rs4246861 PPARA rs4253778 CPVL rs4257918 APOE rs429358 PHF2 rs4498613 (intergenic) rs4522375 COMT rs4680 PIWIL1 rs4759659 ADD1 rs4961 LRP5 rs4988300 ADRB3 rs4994 TYMS rs502396 AGT rs5050 (intergenic) rs5082 EDN1 rs5370 GNB3 rs5443 NTRK2 rs615189 LRP5 rs634008 YWHAQ rs6432018 RTP1 rs6444210 SVIL rs6481619 GABARAPL2 rs6564267 UST rs6570913 (intergenic) rs659366 UCP2 rs660339 PON1 rs662 APOA1-AS rs670 p300 rs67744457 MIR538 rs6865159 NRF1 rs6949152 AGT rs699 (intergenic) rs699947 CETP rs708272 GNAS rs7121 IGF1 rs7136446 TMX1 | FRMD6 rs7154161 GABPB1 rs7181866 CYP19A1 rs727479
59 | P a g e
MIPEP rs7324557 DEPDC6 rs7386139 APOE rs7412 C7orf74 rs7792872 TCERG1l rs7893895 NHLRC2 rs7901769 PDE3A rs7964046 GABPB1 rs8031031 LOC100131241 rs8069419 CKMM rs8111989 PRLHR rs8192524 PPARGC1A rs8192678 DRG2 rs854762 DRG2 rs854813 OR6N2 rs857838 ALX4 rs903514 MYLIP rs909562 ESR1 rs9340799 FOXQ1 rs9378283 SLFN11 rs938298 HDAC2 rs9488289 RTP1 rs9869355 RTP1 rs9872701 FTO rs9930506 ACE rs4646994
60 | P a g e
Table S2: Raw Genotyping Data for the Completers 17-29
Group A B A B B A B B A A B A B Gene(s) SNP ID 43 44 45 47 48 49 50 51 54 56 58 62 63 FAT2 rs10072841 MATN2 rs10099863 PDK4 rs10247649 TFEC rs10248479 ADRB2 rs1042713 ADRB2 rs1042714 ENPP1 rs1044498 TTN rs10497520 PPP1CC rs1050587 PPARD rs1053049 CYP2D6 rs1065852 PLA2G2A rs10732279 HBB rs10768683 (intergenic) rs10861553 AMN1 rs11051548 (intergenic) rs1107946 (intergenic) rs11096663 LEPR rs1137101 NME7 rs1138486 HIF1A rs11549465 ID3 rs11574 RNASE1 rs11622895 EPAS1 rs11689011 (intergenic) rs11770725 (intergenic) rs12079745 DIO1 rs12095080 (intergenic) rs12535708 (intergenic) rs12535747 GABPB1 rs12594956 KCNH7 rs12613181 TMEM195 rs12672644 YWHAQ rs12692388 API5 rs12789205 CREBBP rs130021 LOC100288831 rs13060995 HDAC2 rs13212283 LEP rs13245201 (intergenic) rs1341439 (intergenic) rs1349419
61 | P a g e
STS rs13648 LOC730100 rs1451462 ANO4 rs1532723 DIS3L rs1546570 GPRIN3 rs1560488 (intergenic) rs1570360 NPY rs16139 EPS1 rs17039192 LOC100129278 rs17044554 (intergenic) rs17231506 TTC17 rs17508783 AMPD1 rs17602729 (intergenic) rs17782313 (intergenic) rs1799768 FABP2 rs1799883 HFE rs1799945 COL1A1 rs1800012 CNTF rs1800169 PPARA rs1800234 TGFB1 rs1800470 HFE rs1800562 (intergenic) rs1800588 (intergenic) rs1800629 CETP rs1800775 LOC541472 rs1800795 UCP3 rs1800849 MTHFR rs1801131 MTHFR rs1801133 ADRB1 rs1801252 ADRB1 rs1801253 MTRR rs1801394 MYOSTAIN rs1805086 MTR rs1805087 ACTN3 rs1815739 SPATA17 rs1832544 EPAS1 rs1867785 PTPRT rs1885831 MTHFD1 rs1950902 LIPG rs2000813 FTHFD rs2002287 VEGFA rs2010963 PPARD rs2016520 p300 (EP300) rs20551
62 | P a g e
GCH1 rs2057368 NOS3 rs2070744 RAB6A rs2140892 IL15RA rs2228059 VDR rs2228570 ESR1 rs2234693 H19 rs2251375 REEP3 rs2252578 CREB1 rs2253206 DOCK2 rs2270895 FTHFD rs2276731 TFAM rs2306604 LRP5 rs2306862 HYDIN rs235987 CREB1 rs2360969 NRF1 rs2402970 SLC22A3 rs2457571 CDH2 | DSC3 rs2542729 BTAF1 rs2792022 NFKB1 rs28362491 MYLK rs28497577 ACVR1B rs2854464 (intergenic) rs2854744 RBPMS rs2979481 CREBBP rs3025684 TOM1L2 rs3183702 LPL rs320 CHRM2 rs324640 LPL rs328 RYR2 rs34967813 NUAK1 rs3741886 ATP1B1 rs3766031 ATP1B1 rs3766044 NRP2 rs3770991 PDK4 rs3779478 GRHL1 rs3791749 EPAS1 rs4035887 TUSC1 rs4246861 PPARA rs4253778 CPVL rs4257918 APOE rs429358 PHF2 rs4498613 (intergenic) rs4522375
63 | P a g e
COMT rs4680 PIWIL1 rs4759659 ADD1 rs4961 LRP5 rs4988300 ADRB3 rs4994 TYMS rs502396 AGT rs5050 (intergenic) rs5082 EDN1 rs5370 GNB3 rs5443 NTRK2 rs615189 LRP5 rs634008 YWHAQ rs6432018 RTP1 rs6444210 SVIL rs6481619 GABARAPL2 rs6564267 UST rs6570913 (intergenic) rs659366 UCP2 rs660339 PON1 rs662 APOA1-AS rs670 p300 (EP300) rs67744457 MIR538 rs6865159 NRF1 rs6949152 AGT rs699 (intergenic) rs699947 CETP rs708272 GNAS rs7121 IGF1 rs7136446 TMX1 | FRMD6 rs7154161 GABPB1 rs7181866 CYP19A1 rs727479 MIPEP rs7324557 DEPDC6 rs7386139 APOE rs7412 C7orf74 rs7792872 TCERG1l rs7893895 NHLRC2 rs7901769 PDE3A rs7964046 GABPB1 rs8031031 LOC100131241 rs8069419 CKMM rs8111989
64 | P a g e
PRLHR rs8192524 PPARGC1A rs8192678 DRG2 rs854762 DRG2 rs854813 OR6N2 rs857838 ALX4 rs903514 MYLIP rs909562 ESR1 rs9340799 FOXQ1 rs9378283 SLFN11 rs938298 HDAC2 rs9488289 RTP1 rs9869355 RTP1 rs9872701 FTO rs9930506 ACE rs4646994
Table S3: Raw Genotyping Data for the Completers 30-42
Group B A A A A B A A B B B A A Gene(s) SNP ID 65 67 72 75 78 79 80 83 84 87 91 95 102 FAT2 rs10072841 MATN2 rs10099863 PDK4 rs10247649 TFEC rs10248479 ADRB2 rs1042713 ADRB2 rs1042714 ENPP1 rs1044498 TTN rs10497520 PPP1CC rs1050587 PPARD rs1053049 CYP2D6 rs1065852 PLA2G2A rs10732279 HBB rs10768683 (intergenic) rs10861553 AMN1 rs11051548 (intergenic) rs1107946 (intergenic) rs11096663 LEPR rs1137101 NME7 rs1138486 HIF1A rs11549465 ID3 rs11574 RNASE1 rs11622895
65 | P a g e
EPAS1 rs11689011 (intergenic) rs11770725 (intergenic) rs12079745 DIO1 rs12095080 (intergenic) rs12535708 (intergenic) rs12535747 GABPB1 rs12594956 KCNH7 rs12613181 TMEM195 rs12672644 YWHAQ rs12692388 API5 rs12789205 CREBBP rs130021 LOC100288831 rs13060995 HDAC2 rs13212283 LEP rs13245201 (intergenic) rs1341439 (intergenic) rs1349419 STS rs13648 LOC730100 rs1451462 ANO4 rs1532723 DIS3L rs1546570 GPRIN3 rs1560488 (intergenic) rs1570360 NPY rs16139 EPS1 rs17039192 LOC100129278 rs17044554 (intergenic) rs17231506 TTC17 rs17508783 AMPD1 rs17602729 (intergenic) rs17782313 (intergenic) rs1799768 FABP2 rs1799883 HFE rs1799945 COL1A1 rs1800012 CNTF rs1800169 PPARA rs1800234 TGFB1 rs1800470 HFE rs1800562 (intergenic) rs1800588 (intergenic) rs1800629 CETP rs1800775 LOC541472 rs1800795 UCP3 rs1800849
66 | P a g e
MTHFR rs1801131 MTHFR rs1801133 ADRB1 rs1801252 ADRB1 rs1801253 MTRR rs1801394 MYOSTAIN rs1805086 MTR rs1805087 ACTN3 rs1815739 SPATA17 rs1832544 EPAS1 rs1867785 PTPRT rs1885831 MTHFD1 rs1950902 LIPG rs2000813 FTHFD rs2002287 VEGFA rs2010963 PPARD rs2016520 p300 (EP300) rs20551 GCH1 rs2057368 NOS3 rs2070744 RAB6A rs2140892 IL15RA rs2228059 VDR rs2228570 ESR1 rs2234693 H19 rs2251375 REEP3 rs2252578 CREB1 rs2253206 DOCK2 rs2270895 FTHFD rs2276731 TFAM rs2306604 LRP5 rs2306862 HYDIN rs235987 CREB1 rs2360969 NRF1 rs2402970 SLC22A3 rs2457571 CDH2 | DSC3 rs2542729 BTAF1 rs2792022 NFKB1 rs28362491 MYLK rs28497577 ACVR1B rs2854464 (intergenic) rs2854744 RBPMS rs2979481 CREBBP rs3025684 TOM1L2 rs3183702
67 | P a g e
LPL rs320 CHRM2 rs324640 LPL rs328 RYR2 rs34967813 NUAK1 rs3741886 ATP1B1 rs3766031 ATP1B1 rs3766044 NRP2 rs3770991 PDK4 rs3779478 GRHL1 rs3791749 EPAS1 rs4035887 TUSC1 rs4246861 PPARA rs4253778 CPVL rs4257918 APOE rs429358 PHF2 rs4498613 (intergenic) rs4522375 COMT rs4680 PIWIL1 rs4759659 ADD1 rs4961 LRP5 rs4988300 ADRB3 rs4994 TYMS rs502396 AGT rs5050 (intergenic) rs5082 EDN1 rs5370 GNB3 rs5443 NTRK2 rs615189 LRP5 rs634008 YWHAQ rs6432018 RTP1 rs6444210 SVIL rs6481619 GABARAPL2 rs6564267 UST rs6570913 (intergenic) rs659366 UCP2 rs660339 PON1 rs662 APOA1-AS rs670 p300 (EP300) rs67744457 MIR538 rs6865159 NRF1 rs6949152 AGT rs699 (intergenic) rs699947
68 | P a g e
CETP rs708272 GNAS rs7121 IGF1 rs7136446 TMX1 | FRMD6 rs7154161 GABPB1 rs7181866 CYP19A1 rs727479 MIPEP rs7324557 DEPDC6 rs7386139 APOE rs7412 C7orf74 rs7792872 TCERG1l rs7893895 NHLRC2 rs7901769 PDE3A rs7964046 GABPB1 rs8031031 LOC100131241 rs8069419 CKMM rs8111989 PRLHR rs8192524 PPARGC1A rs8192678 DRG2 rs854762 DRG2 rs854813 OR6N2 rs857838 ALX4 rs903514 MYLIP rs909562 ESR1 rs9340799 FOXQ1 rs9378283 SLFN11 rs938298 HDAC2 rs9488289 RTP1 rs9869355 RTP1 rs9872701 FTO rs9930506 ACE rs4646994
69 | P a g e
Table S4: SNPs Associated with Aerobic Capacity
SNP ID Gene(s) Hetero Beneficial Reference rs10497520 TTN C/T T (Stebbings GK, 2018) rs1053049 PPARD C/T C (Ildus I. Ahmetov, 2015) rs10768683 HBB C/G G (Maffulli, 2013) rs11051548 AMN1 G/A A (Yoo J, 2016) rs11096663 PUM2 | RHOB A/G A (Yoo J, 2016) rs11574 ID3 C/T T (Timmons, 2009) rs12594956 GABPB1 (NRF2) A/C A (He Z, 2007) rs12613181 KCNH7 C/T C (Yoo J, 2016) rs13060995 SRGAP3 | LOC100288831 A/G A (Yoo J, 2016) rs1451462 LOC730100 | ASB3 C/T C (Yoo J, 2016) rs1546570 DIS3L A/C A (Timmons, 2009) rs1570360 (intergenic) A/G A or G* (Prior SJ, 2006) rs1799945 HFE C/G C/G (Grealy, 2015) rs1800629 (intergenic) A/G A (Del Coso J, 2017) rs1800795 LOC541472 C/G G (Del Coso J, 2017) rs2010963 VEGFA C/G C or G* (Prior SJ, 2006) rs2016520 PPARD C/T C (Maciejewska-karlowska, 2013) rs2251375 H19 A/C A (Timmons, 2009) rs2402970 NRF1 C/T C (He Z, 2007) rs2457571 SLC22A3 C/T T (Timmons, 2009) rs2542729 CDH2 | DSC3 C/T C (Yoo J, 2016) rs2792022 BTAF1 C/T C (Timmons, 2009) rs28497577 MYLK G/T T (Del Coso J, 2017) rs3770991 NRP2 A/G A (Timmons, 2009) rs4257918 CPVL A/G A (Timmons, 2009) rs6481619 SVIL A/C C (Timmons, 2009) rs6570913 UST A/G A (Yoo J, 2016) rs6949152 NRF1 A/G A (He Z, 2007) rs699947 (intergenic) A/C A or C* (Prior SJ, 2006) rs7181866 GABPB1 (NRF2) A/G G (He Z, 2007) rs7324557 MIPEP A/G A (Timmons, 2009) rs7386139 DEPDC6 A/G G (Timmons, 2009) rs8031031 GABPB1 (NRF2) C/T T (He Z, 2007) rs8111989 CKMM C/T T (Chen, 2017) rs8192678 PPARGC1A C/T C (Calvo JA, 2008) * Either allele within a specific haplotype with other SNPs can be beneficial
70 | P a g e
Table S5: SNPs Associated with Physical Strength/Response to Training
SNP ID Gene(s) Hetero Beneficial Reference rs10072841 FAT2 C/T C (Yoo J, 2016) rs1050587 PPP1CC C/T C (Windelinckx A, 2011) rs10861553 (intergenic) G/T T (Windelinckx A, 2011) rs1137101 LEPR A/G G (Quinton, 2001) rs11549465 HIF1A C/T C (Eynon N, 2010) rs12095080 DIO1 A/G T (Peeters RP1, 2005) rs1341439 SLC10A2 | DAOA C/T C (Yoo J, 2016) rs16139 NPY C/T C (Kallio J, 2001) rs17044554 LOC100129278 | KLHL29 G/T G (Yoo J, 2016) rs1800012 COL1A1 A/C C (Baumert, 2018) rs1800169 CNTF A/G A (Conwit RA, 2005) rs1805086 MYOSTAIN (GDF-8) C/T C (Catalina Santiago, 2011) rs1867785 EPAS1 A/G G (Henderson J., 2005) rs2228059 IL15RA G/T G (Bruneau Jr M, 2018) rs2306604 TFAM A/G A (Maruszak A1, 2014) rs3741886 NUAK1 A/C C (Windelinckx A, 2011) rs4035887 EPAS1 A/G G (Sarah Voisin, 2016) rs4253778 PPARA C/G C (Stastny P, 2019) rs4522375 NDN | PWRN2 C/T C (Yoo J, 2016) rs4988300 LRP5 G/T G (Kim HJ, 2018) rs5443 GNB3 C/T T (Grove ML, 2007) rs6564267 GABARAPL2 G/T G (Yoo J, 2016) rs660339 UCP2 A/G A (Kim HJ, 2018) rs7136446 IGF1 C/T C (Kostek MC, 2005) rs7154161 TMX1 | FRMD6 C/T C (Yoo J, 2016) rs7901769 NHLRC2 (PPP1CC) A/G A (Windelinckx A, 2011) rs11689011 EPAS1 C/T C (Henderson J., 2005) rs2854464 ACVR1B A/G A (Sarah Voisin, 2016)
71 | P a g e
Table S6: SNPs Associated with Cardiovascular Health
SNP ID Gene(s) Hetero Beneficial Reference rs10099863 MATN2 A/C A (Rankinen T S. Y., 2012) rs10248479 TFEC G/T G (Rankinen T S. Y., 2012) rs1065852 CYP2D6 A/G G (Zateyshchikov DA, 2007) rs10732279 PLA2G2A C/T C (Rankinen T S. Y., 2012) rs1138486 NME7 C/T T (Chang YP1, 2007) rs11622895 RNASE1 C/T T (Rankinen T S. Y., 2012) rs12079745 (intergenic) A/G A (Prasad MK, 8(10): e76290) rs12672644 TMEM195 C/T T (Rankinen T S. Y., 2012) rs12692388 YWHAQ C/T T (Rankinen T S. Y., 2012) rs12789205 API5 G/T T (Rankinen T S. Y., 2012) rs1560488 GPRIN3 C/T T (Rankinen T S. Y., 2012) rs17508783 TTC17 A/G G (Rankinen T S. Y., 2012) rs17602729 AMPD1 A/G G (Rico-Sanz J, 2003) rs1800470 TGFB1 A/G A (Rivera MA, 2001) rs1800562 HFE G/A A (Sørensen E, 2019) rs1800775 CETP A/C C (Wang, 2014) rs1801252 ADRB1 A/G G (Pacanowski MA, 2008) rs1801253 ADRB1 C/G C (Pacanowski MA, 2008) rs1805087 MTR A/G G (Deng, 2019) rs1832544 SPATA17 A/G G (Rankinen T S. Y., 2012) rs1885831 PTPRT A/G G (Rankinen T S. Y., 2012) rs1950902 MTHFD1 A/G G (Wernimont, 2011) rs2057368 GCH1 A/G A (Rankinen T S. Y., 2012) rs2070744 NOS3 C/T T (Gamil, 2017) rs2140892 RAB6A A/G A (Rankinen T S. Y., 2012) rs2252578 REEP3 C/T C (Rankinen T S. Y., 2012) rs2253206 CREB1 A/G G (Rankinen T S. Y., 2012) rs2270895 DOCK2 A/G G (Rankinen T S. Y., 2012) rs235987 HYDIN C/T T (Rankinen T S. Y., 2012) rs2360969 CREB1 C/T T (Rankinen T S. Y., 2012) rs28362491 NFKB1 */ATTG ATTG (Park JY, 2007) rs2979481 RBPMS C/T C (Rankinen T S. Y., 2012) rs3183702 TOM1L2 A/G G (Rankinen T S. Y., 2012) rs320 LPL G/T G (Shatwan, 2018) rs324640 CHRM2 A/G A (Hautala AJ, 2006) rs34967813 RYR2 A/G G (Galati F, 2016) rs3766031 ATP1B1 C/T T (Chang YP1, 2007) rs3766044 ATP1B1 A/G G (Chang YP1, 2007) rs3791749 GRHL1 C/T C (Rankinen T S. Y., 2012) rs4246861 TUSC1 C/T C (Rankinen T S. Y., 2012)
72 | P a g e
rs4498613 PHF2 A/C C (Rankinen T S. Y., 2012) rs4680 COMT A/G A (Ge L, 2015) rs4759659 PIWIL1 A/G A (Rankinen T S. Y., 2012) rs4961 ADD1 G/T G (Sousa AC, 2017) rs502396 TYMS C/T T (Wernimont SM, 2012) rs5050 AGT G/T T (Purkait P, 2017) rs5370 EDN1 G/T G (Rankinen T C. T., 2007) rs615189 NTRK2 C/T C (Rankinen T S. Y., 2012) rs6432018 YWHAQ A/C A (Rankinen T S. Y., 2012) rs6444210 RTP1 C/T C (Rankinen T S. Y., 2012) rs659366 (intergenic) C/T T (Sun H, 2018) rs6865159 MIR538 C/T C (Rankinen T S. Y., 2012) rs699 AGT A/G G (Takeuchi F, 2012) rs7121 GNAS C/T C (Nieminen T, 2006) rs7412 APOE C/T C (Hagberg JM, 1999) rs7792872 C7orf74 C/T T (Rankinen T S. Y., 2012) rs7893895 TCERG1l C/T C (Rankinen T S. Y., 2012) rs7964046 PDE3A C/T T (Rankinen T S. Y., 2012) rs8069419 LOC100131241 A/C A (Rankinen T S. Y., 2012) rs8192524 PRLHR A/G A (Franks PW, 2004) rs854762 DRG2 A/G A (Rankinen T S. Y., 2012) rs854813 DRG2 T/C T (Rankinen T S. Y., 2012) rs857838 OR6N2 A/C A (Rankinen T S. Y., 2012) rs903514 ALX4 A/G A (Rankinen T S. Y., 2012) rs909562 MYLIP A/G G (Rankinen T S. Y., 2012) rs9378283 FOXQ1 A/G G (Rankinen T S. Y., 2012) rs938298 SLFN11 C/T T (Rankinen T S. Y., 2012) rs9869355 RTP1 C/T C (Rankinen T S. Y., 2012) rs9872701 RTP1 A/G G (Rankinen T S. Y., 2012)
73 | P a g e
Table S7: SNPs Associated with Increased BMI/Obesity
SNP ID Gene(s) Hetero Beneficial Reference rs10247649 PDK4 A/G A (Strachan DP, 2007) rs1042713 ADRB2 A/G A (Wolfarth B., 2000) rs1042714 ADRB2 C/G C (Gjesing AP1, 2007) rs1044498 ENPP1 A/C A (Pizzuti, 2008) rs11770725 (intergenic) C/T T (Jiang Y1, 2004) rs12535708 (intergenic) A/C A (Jiang Y1, 2004) rs12535747 (intergenic) A/C A (Jiang Y1, 2004) rs13245201 LEP A/G G (Jiang Y1, 2004) rs1349419 (intergenic) A/G A (Jiang Y1, 2004) rs13648 STS A/G G (Riechman SE, 2004) rs17231506 (intergenic) C/T C (Spielmann N, 2007) rs17782313 (intergenic) C/T T (Loos, 2008) rs1799883 FABP2 C/T C (Hancock AM, 2018) rs1800234 PPARA C/T T (Sabo, 2017) rs1800588 (intergenic) C/T C (Teran-Garcia M, 2005) rs1800849 UCP3 A/G G (Longo UG, 2010) rs1801394 MTRR A/G G (Bokor S, 2011) rs2000813 LIPG C/T C (Halverstadt A, 2003) rs2228570 VDR A/G G (Micheli ML, 2011) rs2854744 (intergenic) G/T G (Huuskonen A, 2011) rs328 LPL C/G C (Zhang S, 2015) rs3779478 PDK4 A/G G (Moon SS, 2012) rs429358 APOE C/T C or T* (Hagberg JM, 1999) rs4994 ADRB3 A/G A (Sakane N, 1997) rs5082 (intergenic) A/G T (Pisciotta L, 2003) rs634008 LRP5 C/T C (Guo, 2006) rs662 PON1 C/T C (Senti M, 2000) rs670 APOA1-AS G/A G (Ruano G, 2006) rs708272 CETP A/G G (Campos Y, 2002) rs9930506 FTO A/G A (Scuteri A, 2007) * Either allele within a specific haplotype with other SNPs can be beneficial
74 | P a g e
Table S8: SNPs Associated with Endurance and Strength
SNP ID Gene(s) Hetero Beneficial Reference rs1815739 ACTN3 T/C T or C (Yang N., 2003) rs4646994 ACE D/I I or D (Bigham, 2008)
Table S9: SNPs associated with Other Aspects of Physical Fitness
SNP ID Gene(s) Hetero Beneficial Reference rs1107946 (intergenic) A/C A (Stewart TL, 2006) rs1532723 ANO4 C/T T (Tian Y, 2012) rs17039192 EPS1 C/T T (Saito, 2010) rs1799768 (intergenic) -/G G (Yang, 2019) rs2234693 ESR1 C/T C (Tobias, 2007) rs2306862 LRP5 C/T C (Kiel, 2007) rs727479 CYP19A1 A/C C (Ellis, 2001) rs9340799 ESR1 A/G G (Tobias, 2007)
Table S10: SNPs Associated with Epigenetic Modifiers
SNP ID Gene(s) Hetero Beneficial Reference rs130021 CREBBP A/G n/a (Kumar D, 2011) rs13212283 HDAC2 A/G n/a (Jennifer Jurkin, 2011) rs1801131 MTHFR G/T n/a (Castro, 2004) rs1801133 MTHFR G/A n/a (Castro, 2004) rs2002287 FTHFD (ALDH1L1) A/G n/a (Stover, 2012) rs20551 p300 (EP300) A/G n/a (Jiao Li, 2017) rs2276731 FTHFD (ALDH1L1) C/T n/a (Min-Ae Song, 2016) rs3025684 CREBBP A/G n/a (Kumar D, 2011) rs67744457 p300 (EP300) GAA/* n/a (Jiao Li, 2017) rs9488289 HDAC2 C/G n/a (Jennifer Jurkin, 2011)
75 | P a g e
Table S11: ANOVA Analysis predictive SNPs for Anaerobic Peak Power % Change 0-12 Weeks
SNP ID Performance Significance rs10248479 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0002 rs1053049 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0500 rs1138486 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0073 rs11689011 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0364 rs12079745 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0073 rs1800562 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0009 rs1801131 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0193 rs1867785 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0257 rs2057368 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0111 rs28362491 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0478 rs4961 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0023 rs6432018 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0239 rs7181866 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0062 rs7154161 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0350 rs7792872 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0355 rs903514 AP (Anaerobic Peak Power) % Change 0-12 weeks p= 0.0424
Table S12: ANOVA Analysis predictive SNPs for Relative Peak Power % Change 0-12 Weeks
SNP ID Performance Significance rs10248479 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0002 rs1138486 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0083 rs11689011 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0264 rs12079745 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0083 rs1800562 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0365 rs1801131 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0385 rs1867785 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0169 rs2057368 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0065 rs28362491 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0418 rs4961 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0028 rs6432018 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0236 rs7181866 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0073 rs7154161 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0413 rs7792872 RP (Relative Peak Power) % Change 0-12 weeks p= 0.0426
76 | P a g e
Table S13: ANOVA Analysis predictive SNPs for Chest Press % Change 0-12 Weeks
SNP ID Performance Significance rs12594956 Chest Press % Change 0-12 weeks p= 0.0375 rs17044554 Chest Press % Change 0-12 weeks p= 0.0442 rs1805087 Chest Press % Change 0-12 weeks p= 0.0474 rs1805086 Chest Press % Change 0-12 weeks p= 0.0455 rs1885831 Chest Press % Change 0-12 weeks p= 0.0474 rs2002287 Chest Press % Change 0-12 weeks p= 0.0233 rs2979481 Chest Press % Change 0-12 weeks p= 0.0029 rs2854464 Chest Press % Change 0-12 weeks p= 0.0124 rs320 Chest Press % Change 0-12 weeks p= 0.0341 rs3779478 Chest Press % Change 0-12 weeks p= 0.0294 rs4246861 Chest Press % Change 0-12 weeks p= 0.0084 rs4253778 Chest Press % Change 0-12 weeks p= 0.0264
Table S14: ANOVA Analysis predictive SNPs for Knee Extension % Change 0-12 Weeks
SNP ID Performance Significance rs10072841 Knee Ext. % Change 0-12 weeks p= 0.0373 rs12613181 Knee Ext. % Change 0-12 weeks p= 0.0407 rs1546570 Knee Ext. % Change 0-12 weeks p= 0.0020 rs1800629 Knee Ext. % Change 0-12 weeks p= 0.0263 rs1800470 Knee Ext. % Change 0-12 weeks p= 0.0226 rs1832544 Knee Ext. % Change 0-12 weeks p= 0.0025 rs324640 Knee Ext. % Change 0-12 weeks p= 0.0289 rs34967813 Knee Ext. % Change 0-12 weeks p= 0.0225 rs4257918 Knee Ext. % Change 0-12 weeks p= 0.0123 rs4646994 Knee Ext. % Change 0-12 weeks p= 0.0076 rs6570913 Knee Ext. % Change 0-12 weeks p= 0.0445 rs699 Knee Ext. % Change 0-12 weeks p= 0.0300 rs7901769 Knee Ext. % Change 0-12 weeks p= 0.001
77 | P a g e
Table S15: ANOVA Analysis predictive SNPs for Overhead Press % Change 0-12 Weeks
SNP ID Performance Significance rs10732279 Overhead Press % Change 0-12 weeks p= 0.0038 rs1107946 Overhead Press % Change 0-12 weeks p= 0.0034 rs130021 Overhead Press % Change 0-12 weeks p= 0.0110 rs1560488 Overhead Press % Change 0-12 weeks p= 0.0290 rs17602729 Overhead Press % Change 0-12 weeks p= 0.0172 rs2070744 Overhead Press % Change 0-12 weeks p= 0.0031 rs2360969 Overhead Press % Change 0-12 weeks p= 0.0074
Table S16: ANOVA Analysis predictive SNPs for Knee Peak Power % Change 0-12 Weeks
SNP ID Performance Significance rs12095080 Knee Peak Power PP % Change 0-12 weeks p= 0.0437 rs130021 Knee Peak Power PP % Change 0-12 weeks p= 0.0471 rs1560488 Knee Peak Power PP % Change 0-12 weeks p= 0.0321 rs1799768 Knee Peak Power PP % Change 0-12 weeks p= 0.0132 rs2070744 Knee Peak Power PP % Change 0-12 weeks p= 0.0295 rs2140892 Knee Peak Power PP % Change 0-12 weeks p= 0.0027 rs28362491 Knee Peak Power PP % Change 0-12 weeks p= 0.0103 rs2457571 Knee Peak Power PP % Change 0-12 weeks p= 0.0248 rs660339 Knee Peak Power PP % Change 0-12 weeks p= 0.0019 rs659366 Knee Peak Power PP % Change 0-12 weeks p= 0.0003 rs7412 Knee Peak Power PP % Change 0-12 weeks p= 0.0054
Table S17: ANOVA Analysis predictive SNPs for Lat Pull % Change 0-12 Weeks
SNP ID Performance Significance rs10099863 Lat Pull % Change 0-12 weeks p= 0.0260 rs11549465 Lat Pull % Change 0-12 weeks p= 0.0299 rs1349419 Lat Pull % Change 0-12 weeks p= 0.0463 rs1570360 Lat Pull % Change 0-12 weeks p= 0.0035 rs2002287 Lat Pull % Change 0-12 weeks p= 0.0075 rs2979481 Lat Pull % Change 0-12 weeks p= 0.0054 rs320 Lat Pull % Change 0-12 weeks p= 0.0362 rs502396 Lat Pull % Change 0-12 weeks p= 0.0163 rs7121 Lat Pull % Change 0-12 weeks p= 0.0356 rs7792872 Lat Pull % Change 0-12 weeks p= 0.0011
78 | P a g e
Table S18: ANOVA Analysis predictive SNPs for Vertical Jump % Change 0-12 Weeks
SNP ID Performance Significance rs10732279 Vertical Jump VJ % Change 0-12 weeks p= 0.0070 rs1801253 Vertical Jump VJ % Change 0-12 weeks p= 0.0064 rs1805087 Vertical Jump VJ % Change 0-12 weeks p= 0.0126 rs2228059 Vertical Jump VJ % Change 0-12 weeks p= 0.0362 rs4994 Vertical Jump VJ % Change 0-12 weeks p= 0.0001 rs5082 Vertical Jump VJ % Change 0-12 weeks p= 0.0027 rs615189 Vertical Jump VJ % Change 0-12 weeks p= 0.0196 rs6444210 Vertical Jump VJ % Change 0-12 weeks p= 0.0261
Table S19: ANOVA Analysis predictive SNPs for Squat % Change 0-12 Weeks
SNP ID Performance Significance rs11622895 Squat % Change 0-12 weeks p= 0.0056 rs1532723 Squat % Change 0-12 weeks p= ˂ 0.001 rs1832544 Squat % Change 0-12 weeks p= 0.0476 rs1885831 Squat % Change 0-12 weeks p= 0.0433 rs2234693 Squat % Change 0-12 weeks p= 0.0057 rs28362491 Squat % Change 0-12 weeks p= 0.0222 rs324640 Squat % Change 0-12 weeks p= 0.0112 rs4246861 Squat % Change 0-12 weeks p= 0.0143 rs708272 Squat % Change 0-12 weeks p= 0.0410 rs7412 Squat % Change 0-12 weeks p= 0.0311 rs7964046 Squat % Change 0-12 weeks p= 0.0068 rs8069419 Squat % Change 0-12 weeks p= 0.0427 rs9930506 Squat % Change 0-12 weeks p= 0.0457
79 | P a g e
Table S20: ANOVA Analysis predictive SNPs for VO2max (ml/kg/min) % Change 0-12 Weeks
SNP ID Performance Significance rs17044554 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0111 rs1799768 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0406 rs1805087 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0279 rs1805086 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0079 rs2402970 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0006 rs2854744 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0121 rs3770991 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0298 rs4961 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0072 rs502396 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0077 rs660339 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0274 rs6949152 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0087 rs699947 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0463 rs7792872 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0429 rs857838 VO2 Peak (ml/kg/min) % Change 0-12 weeks p= 0.0015
Table S21: ANOVA Analysis predictive SNPs for Arm Lean Mass LM Change 0-12 Weeks (grams)
SNP ID Performance Significance rs1560488 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0067 rs1801133 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0425 rs2276731 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0101 rs2457571 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0149 rs2854744 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0234 rs4759659 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0424 rs7792872 Arm Lean Mass LM Change 0-12 weeks (grams) p= 0.0075
Table S22: ANOVA Analysis predictive SNPs for Max Voluntary Contraction % Change 0-12 Weeks
SNP ID Performance Significance rs17602729 MVC % Change 0-12 weeks p= 0.0124 rs2542729 MVC % Change 0-12 weeks p= 0.0432 rs324640 MVC % Change 0-12 weeks p= 0.0431 rs3741886 MVC % Change 0-12 weeks p= 0.0141 rs708272 MVC % Change 0-12 weeks p= 0.0150 rs727479 MVC % Change 0-12 weeks p= 0.0074
80 | P a g e
Table S23: ANOVA Analysis predictive SNPs for Bodyfat % Change 0-12 Weeks
SNP ID Performance Significance rs12672644 Bodyfat Change 0-12 weeks p= 0.0316 rs1799883 Bodyfat Change 0-12 weeks p= 0.0223 rs1800012 Bodyfat Change 0-12 weeks p= 0.0072 rs1801131 Bodyfat Change 0-12 weeks p= 0.0089 rs2228570 Bodyfat Change 0-12 weeks p= 0.0387 rs2792022 Bodyfat Change 0-12 weeks p= 0.0402 rs328 Bodyfat Change 0-12 weeks p= 0.0008 rs3770991 Bodyfat Change 0-12 weeks p= 0.0364 rs3791749 Bodyfat Change 0-12 weeks p= 0.0489 rs4257918 Bodyfat Change 0-12 weeks p= 0.0427 rs615189 Bodyfat Change 0-12 weeks p= 0.0218 rs7324557 Bodyfat Change 0-12 weeks p= 0.0340 rs8069419 Bodyfat Change 0-12 weeks p= 0.0037 rs903514 Bodyfat Change 0-12 weeks p= 0.0117
Table S24: ANOVA Analysis predictive SNPs for Thigh Lean Mass, Both Legs, Change 0-12 Weeks (grams)
SNP ID Performance Significance rs10072841 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0380 rs11689011 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0299 rs1532723 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0038 rs1546570 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0149 rs1801133 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0175 rs1867785 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0311 rs3025684 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0077 rs3791749 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0115 rs4680 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0459 rs5370 Thigh Lean Mass; Both Legs TLM Change 0-12 weeks (grams) p= 0.0316
81 | P a g e
Table S25: ANOVA Analysis predictive SNPs for Visceral Fat Mass Change 0-12 Weeks (grams)
SNP ID Performance Significance rs1042713 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0374 rs130021 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0452 rs17231506 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0203 rs17602729 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0224 rs1800775 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0275 rs1801252 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0420 rs34967813 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0285 rs3791749 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0385 rs4646994 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0273 rs4522375 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0008 rs4759659 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0345 rs708272 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0200 rs7324557 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0236 rs7181866 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0092 rs7901769 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0005 rs938298 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0029 rs9488289 Visceral Fat Mass Change 0-12 weeks (grams) p= 0.0287
Table S26: ANOVA Analysis predictive SNPs for Total Lean Mass Change 0-12 Weeks (grams)
SNP ID Performance Significance rs11689011 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0347 rs11622895 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0161 rs1799945 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0221 rs1801133 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0025 rs1867785 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0315 rs2252578 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0460 rs3025684 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0409 rs3791749 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0464 rs660339 Total Lean Mass LM Change 0-12 weeks (grams) p= 0.0167
82 | P a g e
VII. BIBLIOGRAPHY Baumert, P. e. (2018). Variations of collagen-encoding genes are associated with exercise-induced muscle damage. Physiological genomics, 50,9: 691-693.
Bigham, A. W.-V.-C. (2008). Angiotensin-converting enzyme genotype and arterial oxygen saturation at high altitude in Peruvian Quechua. High altitude medicine & biology, 9(2): 167-178.
Bokor S, M. A.-G. (2011). Common polymorphisms in six genes of the methyl group metabolism pathway and obesity in European adolescents. Int J Pediatr Obes., 6(2-2):e336-344.
Bruneau Jr M, W. S. (2018). A genetic variant in IL-15Ra correlates with physical activity amoung European-American Adults. Mol Genet Genomic Med., 6: 401-408.
Calvo JA, D. T. (2008). Muscle-specific expression of PPARgamma coactivator-1alpha improves exercise performance and increases peak oxygen uptake. J Appl Physiol, 104: 1304-1312.
Campos Y, G. A. (2002). Cosegregation of the mitochondrial DNA A1555G and G4309A mutations results in deafness and mitochondrial myopathy. Muscle Nerve, 25:185-188.
Castro, R. R. (2004). 5,10-Methylenetetrahydrofolate reductase (MTHFR) 677C-T and 1298A-C mutations are associated with DNA hypomethylation. . J. Med. Genet, 41:454-458.
Catalina Santiago, J. R.-R.-L.-F.-G. (2011). The K153R Polymorphism in the Myostatin Gene and Muscle Power Phenotypes in Young, Non-Athletic Men. PLOS One, 6(1): 16323.
Chang YP1, L. X. (2007). Multiple genes for essential-hypertension susceptibility on chromosome 1q. Am J Hum Genet, 80(2): 253-264.
Chappelle, R. (2015). USAD Combat Rescue Officer Course Washout Rate. Aerospace Medicine Summit.
Chen, C. S. (2017). A meta-analysis of the association of CKM gene rs8111989 polymorphism with sport performance. Biology of sport, 34(4): 323-330.
Conwit RA, L. S. (2005). The relationship between ciliary neurotrophic factor (CNTF) genotype and motor unit physiology: preliminary studies. BMC Physiol., 5:15.
Del Coso J, V. M. (2017). Optimum polygenic profile to resist exertional rhabdomyolysis during a marathon. PLoS ONE, 12(3).
Deng, C. e. (2019). Genetic polymorphisms in MTR are associated with non-syndromic congenital heart disease from a family-based case-control study in the Chinese population. Scientific Reports, 9,1 5065.
Ellis, J. A. (2001). Significant population variation in adult male height associated with the Y chromosome and the aromatase gene. J. Clin. Endocr. Metab., 86:4147-4150.
83 | P a g e
Eynon N, A. A. (2010). Is the interaction between HIF1A P582S and ACTN3 R577X determinant for power/sprint performance? Metabolism, 59(6): 861-865.
Flück M, K. M. (2019). Cellular Aspects of Muscle Specialization Demonstrate Genotype - Phenotype Interaction Effects in Athletes. Front Physiol, 10:526.
Franks PW, B. S. (2004). Association between physical activity and blood pressure is modified by variants in the G-protein coupled receptor 10. Hypertension, 43:224-228.
Galati F, G. A. (2016). RyR2 QQ2958 Genotype and Risk of Malignant Ventricular Arrhythmias. Cardiol Res Pract.
Gamil, S. e. (2017). Association of NOS3 gene polymorphisms with essential hypertension in Sudanese patients: a case control study. BMC medical genetics , 18,1: 128.
Ge L, W. H. (2015). COMT Val158Met polymorphism is associated with blood pressure and lipid levels in general families of Bama longevous area in China. Int J Clin Exp Pathol, 8(11): 15055-15064.
Gjesing AP1, A. G.-J. (2007). Studies of the associations between functional beta2-adrenergic receptor variants and obesity, hypertension and type 2 diabetes in 7,808 white subjects. Diabetologia, 50(3):563-568.
Goldman, M. (2008). Statistics for Bioinformatics. Stat C141, 1-5.
Grealy, R. H. (2015). Evaluation of a 7-Gene Genetic Profile for Athletic Endurance Phenotype in Ironman Championship Triathletes. PloS one, 10(12).
Grove ML, M. A. (2007). Gene–environment interaction and the GNB3gene in the Atherosclerosis Risk in Communities study. Int J Obes (Lond), 31: 919-926.
Guo, Y. X. (2006). Polymorphisms of the low-density lipoprotein receptor-related protein 5 (LRP5) gene are associated with obesity phenotypes in a large family-based association study. J. Med. Genet., 43: 798-803.
Hagberg JM, F. R. (1999). Exercise training-induced blood pressure and plasma lipid improvements in hypertensives may be genotype dependent. Hypertension, 34:18-23.
Halverstadt A, P. D. (2003). High-density lipoprotein-cholesterol, its subfractions, and responses to exercise training are dependent on endothelial lipase genotype. Metabolism, 52:1505-1511.
Hancock AM, W. D. (2018). Non-Synonymous Single-Nucleotide Polymorphisms and Physical Activity Interactions on Adiposity Parameters in Malaysian Adolescents. Frontiers in endocrinology, 9:209.
Hautala AJ, R. T. (2006). Heart rate recovery after maximal exercise is associated with acetylcholine receptor M2 (CHRM2) gene polymorphism. Am J Physiol Heart Circ Physiol, 291:H459-466.
84 | P a g e
He Z, H. Y. (2007). NRF2 genotype improves endurance capacity in response to training. Int J Sports Med, 28:717-721.
Henderson J., W.-C. J. (2005). EPAS1 gene infl uences the aerobic-anaerobic contribution in elite endurance athletes. Hum Genet., 118(3-4): 416-423.
Huuskonen A, L. J. (2011). Common genetic variation in the IGF1 associates with maximal force output. Med Sci Sports Exerc., 43(12):2368-2374.
Ildus I. Ahmetov, O. N. (2015). Current Progress in Sports Genomics. In O. N. Ildus I. Ahmetov, Advances in Clinical Chemistry (pp. 247-314).
Ioannis D. Papadimitriou, A. L. (2016). ACTN3 R577X and ACE I/D gene variants influence performance in elite sprinters: a multi-cohort study. BMC Genomics, 17:285.
Jennifer Jurkin, G. Z. (2011). Distinct and redundant functions of histone deacetylases HDAC1 and HDAC2 in proliferation and tumorigenesis. Cell Cycle, 10(3):406-412.
Jiang Y1, W. J. (2004). Common variants in the 5' region of the leptin gene are associated with body mass index in men from the National Heart, Lung, and Blood Institute Family Heart Study. Am J Hum Genet., 75(2):220-230.
Jiao Li, N. D. (2017). EP300 single nucleotide polymorphism rs20551 correlates with prolonged overall survival in diffuse large B cell lymphoma patients treated with R-CHOP. Cancer Cell Int., 17:70.
Kallio J, P. U. (2001). Altered intracellular processing and release of neuropeptide Y due to leucine 7 toproline 7 polymorphism in the signal peptide of preproneuropeptide Y in humans. FASEB J., 15: 1242-1244.
Kiel, D. (2007). Genetic Variation at the Low Density Lipoprotein Receptor-Related Protein 5 (LRP5) Locus Modulates Wnt Signaling and the relationship of Physical Activity with Bone Mineral Density in Men. Bone, 40(3):587-596.
Kim HJ, L. S. (2018). Association between gene polymorphisms and obesity and physical fitness in Korean children. Biol Sport, 35(1): 21-27.
Kostek MC, D. M. (2005). Muscle strength response to strength training is influenced by insulin-like growth factor 1 genotype in older adults. J Appl Physiol, 98: 2147-2154.
Kumar D, D. I. (2011). A polymorphism of the CREB binding protein (CREBBP) gene is a risk factor for addiction. Brain Res., 1406: 59-64.
Longo UG, F. V. (2010). Bilateral consecutive rupture of the quadriceps tendon in a man with BstUI polymorphism of the COL5A1 gene. Knee Surg Sports Traumatol Arthrosc., 18:514-518.
85 | P a g e
Loos, R. J. (2008). Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, and 96 others. Common variants near MC4R are associated with fat mass, weight and risk of obesity. . Nature Genet., 40:768-775.
Maciejewska-karlowska, A. (2013). Polymorphic variants of the PPAR (Peroxisome Proliferator-Activated Receptor) genes: relevance for athletic performance. TRENDS in Sport Science, 1(20): 5-15.
Maffulli, N. M. (2013). The genetics of sports injuries and athletic performance. . Muscles, ligaments and tendons journal, 3(3): 173-189.
Maruszak A1, A. J. (2014). Mitochondrial DNA variation is associated with elite athletic status in the Polish population. Scand J Med Sci Sports, 24(2): 311-318.
McDonald, J. H. (2014). Handbook of Biological Statistics. Baltimore: Sparky House Publishing .
Megan Hawkins, P. R.-G. (2007). Maximal Oxygen Uptake as a Parametric Measure of Cardiorespiratory Capacity. Medicine & Science in Sports & Exercise, 39(1):103-107.
Micheli ML, G. M. (2011). Angiotensin-converting enzyme/vitamin D receptor gene polymorphisms and bioelectrical impedance analysis in predicting athletic performances of Italian young soccer players. J Strength Cond Res., 25(8):2084-2091.
Min-Ae Song, T. M. (2016). Genetic variation in one-carbon metabolism in relation to genome-wide DNA methylation in breast tissue from healthy women. Carcinogenesis, 37(5):471-480.
Moon SS, L. J. (2012). Association of pyruvate dehydrogenase kinase 4 gene polymorphisms with type 2 diabetes and metabolic syndrome. Diabetes Res Clin Pract., 95(2):230-236.
Nieminen T, L. T. (2006). Effects of polymorphisms in beta1-adrenoceptor and alpha-subunit of G protein on heart rate and blood pressure during exercise test. The Finnish Cardiovascular Study. J Appl Physiol, 100:507-511.
Olteanu H, M. T. (2002). Differences in the efficiency of reductive activation of methionine synthase and exogenous electron acceptors between the common polymorphic variants of human methionine synthase reductase. Biochemistry, 12;41(45):13378-13385.
Pacanowski MA, Z. I.-D. (2008). Adrenergic gene polymorphisms and cardiovascular risk in the NHLBI- sponsored Women's Ischemia Syndrome Evaluation. J Transl Med., 6:11.
Park JY, F. I. (2007). NFKB1 promoter variation implicates shear-induced NOS3 gene expression and endothelial function in prehypertensives and stage I hypertensives. Am J Physiol Heart Circ Physiol, 293: H2320-2327.
Peeters RP1, v. d. (2005). A polymorphism in type I deiodinase is associated with circulating free insulin- like growth factor I levels and body composition in humans. J Clin Endocrinol Metab, 90(1): 256- 263.
86 | P a g e
Piotr Gronek, J. H. (2013). Genes and physical fi tness. TRENDS in Sport Sciences, 1(20):16-29.
Pisciotta L, C. A. (2003). Physical activitymodulates effects of some genetic polymorphisms affectingcardiovascular risk in men aged over 40 years. Nutr MetabCardiovasc Dis, 13:202-210.
Pizzuti, A. F. (2008). A polymorphism (K121Q) of the human glycoprotein PC-1 gene coding region is strongly associated with insulin resistance. Diabetes, 57(7):1971-1977.
Prasad MK, B. K. (8(10): e76290). A Polymorphic 3’UTR Element in ATP1B1 Regulates Alternative Polyadenylation and is associated with Blood Pressure. PLoS ONE.
Prior SJ, H. J. (2006). DNA sequence variation in the promoter region of the VEGF gene impacts VEGF gene expression and maximal oxygen consumption. Am J Physiol Heart Circ Physiol, 290:H1848- 1855.
Purkait P, H. K. (2017). Association of angiotensinogen gene SNPs and haplotypes with risk of hypertension in eastern Indian population. Clin Hypertens., 23:12.
Quinton, N. D. (2001). A single nucleotide polymorphism (SNP) in the leptin receptor is associated with BMI, fat mass and leptin levels in postmenopausal Caucasian women. Hum. Genet., 108: 233- 236.
Rankinen T, C. T. (2007). Effect of endothelin 1genotype on blood pressure is dependent on physical activity or fitness levels. Hypertension, 50:1120-1125.
Rankinen T, S. Y. (2012). Heritability of submaximal exercise heart rate response to exercise training is accounted for by nine SNPs. J Appl Physiol (1985), 112(5): 892-897.
Rico-Sanz J, R. T. (2003). Associations between cardiorespiratory responses to exercise and the C34T AMPD1 gene polymorphism in the HERITAGE Family Study. Physiol Genomics, 14: 161-166.
Riechman SE, F. T. (2004). Steroid sulfatase gene variation and DHEA responsiveness to resistance exercise in MERET. Physiol Genomics, 17:300-306.
Rivera MA, E. M. (2001). TGF-beta(1) gene–race interactions for resting and exercise blood pressure in the HERITAGE Family Study. J Appl Physiol, 91: 1808-1813.
Ruano G, S. R. (2006). Apolipoprotein A1genotype affects the change in high density lipoprotein cholesterolsubfractions with exercise training. Atherosclerosis, 185:65-69.
Sabo, A. e. (2017). Exome sequencing reveals novel genetic loci influencing obesity-related traits in Hispanic children. Obesity (Silver Spring, Md.), 25(7):1270-1276.
Saito, T. F. (2010). Transcriptional regulation of endochondral ossification by HIF-2-alpha during skeletal growth and osteoarthritis development. Nature Med., 16:678-686.
87 | P a g e
Sakane N, Y. T. (1997). Effects of Trp64Arg mutation in the beta 3-adrenergicreceptor gene on weight loss, body fat distribution, glycemiccontrol, and insulin resistance in obese type 2 diabetic patients. Diabetes Care, 20:1887-1890.
Sarah Voisin, J. P. (2016). ACVR1B rs2854464 Is Associated with Sprint/Power Athletic Status in a Large Cohort of Europeans but Not Brazilians. PLOS One, 11(6): e0156316.
Scuteri A, S. S. (2007). Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet., 3(7):e115.
Senti M, A. C. (2000). Effect of physical activity on lipid levels in a population-based sample of men with and without the Arg192 variant of the human paraoxonase gene. Genet Epidemiol, 18:276-286.
Shatwan, I. (2018). Association of Apolipoprotein E Gene Polymorphisms with Blood Lipids and their Interaction with Dietary Factors. Lipids in Health and Disease, 17:98.
Sørensen E, R. A. (2019). Genetic factors influencing hemoglobin levels in 15,567 blood donors: results from the Danish Blood Donor Study. Transfusion, 59(1): 226-231.
Sousa AC, M. M. (2017). Synergistic Association of Genetic Variants with Environmental Risk Factors in Susceptibility to Essential Hypertension. Genet Test Mol Biomarkers, 21(10):625-631.
Spielmann N, L. A. (2007). CETP genotypes and HDL-cholesterol phenotypes in the HERITAGE Family Study. Physiol Genomics, 31:25-31.
Stastny P, L. M. (2019). Effect of COL5A1, GDF5, and PPARA Genes on a Movement Screen and Neuromuscular Performance in Adolescent Team Sport Athletes. J Strength Cond Res.
Stebbings GK, W. A. (2018). TTN genotype is associated with fascicle length and marathon running performance. Scandinavian Journal of Medicine & Science in Sports, 28(2): 400-406.
Stewart TL, J. H.-G. (2006). Haplotypes defined by promoter and intron 1 polymorphisms of the COLIA1 gene regulate bone mineral density in women. J Clin Endocrinol Metab., 91(9):3575-3583.
Stover, P. J. (2012). Polymorphisms in 1-Carbon Metabolism, Epigenetics and Folate-Related Pathologies. J Nutrigenet Nutrigenomics, 4)5_:293-305.
Strachan DP, R. A. (2007). Lifecourse influences on health among British adults: effects of region of residence in childhood and adulthood. Int J Epidemiol, 36(3):522-531.
Stranger B, S. E. (2011). Progress and Promise of Genome-Wide Association Studies for Human Complex Trait Genetics. Genetics, 187(2):367-383.
Sun H, Z. J. (2018). Association of uncoupling protein gene polymorphisms with essential hypertension in a northeastern Han Chinese population. J Hum Hypertens.
88 | P a g e
Takeuchi F, Y. K. (2012). Reevaluation of the association of seven candidate genes with blood pressure and hypertension: a replication study and meta-analysis with a larger sample size. Hypertens Res, 35(8):825-831.
Teran-Garcia M, S. N. (2005). Hepatic lipase gene variant j514C 9 T is associated with lipoprotein and insulin sensitivity response to regular exercise: the HERITAGE Family Study. Diabetes, 54:2251- 2255.
Tian Y, S. R. (2012). Anoctamins are a family of Ca2+-activated Cl- channels. J Cell Sci. , 125:4991-4998.
Timmons, J. A. (2009). Using molecular classification to predict gains in maximal aerobic capacity following endurance exercise training in humans. Journal of applied physiology (Bethesda, Md. : 1985), 108(6): 1487-1496.
Tobias, J. H.-G. (2007). Estrogen receptor alpha regulates area-adjusted bone mineral content in late pubertal girls. J. Clin. Endocr. Metab., 92:641-647.
Wang, Q. e. (2014). Seven functional polymorphisms in the CETP gene and myocardial infarction risk: a meta-analysis and meta-regression. PloS One, 9,2 e88118.
Wernimont SM, C. A. (2012). Folate network genetic variation predicts cardiovascular disease risk in non-Hispanic white males. J Nutr., 142(7):1272-1279.
Wernimont, S. M. (2011). Folate network genetic variation, plasma homocysteine, and global genomic methylation content: a genetic association study. BMC medical genetics, 12:150.
Windelinckx A, D. M. (2011). Identification and prioritization of NUAK1 and PPP1CC as positional candidate loci for skeletal muscle strength phenotypes. Physiol Genomics, 43: 981-992.
Wolfarth B., R. M. (2000). A polymorphism in the alpha2a-adrenoceptor gene and endurance athletes status. Med. Sci. Sports Exerc., 32:1709-1712.
Yang N., M. D. (2003). ACTN3 genotype is associated with human elite athletic performance. American journal of human genetics, 73(3): 627-631.
Yang, J. e. (2019). Association between genetic polymorphisms and osteonecrosis in steroid treatment populations: a detailed stratified and dose-response meta-analysis. Bioscience reports, 39:5.
Yoo J, K. B. (2016). Genetic polymorphisms to predict gains in maximal O2 uptake and knee peak torque after a high intensity training program in humans. European Journal of Applied Physiology, 116(5): 947-957.
Zateyshchikov DA, M. L. (2007). Association of CYP2D6 and ADRB1 genes with hypotensive and antichronotropic action of betaxolol in patients with arterial hypertension. Fundam Clin Pharmacol, 21: 437-443.
89 | P a g e
Zhang S, M. Y. (2015). Diets high in carbohydrate may not be appropriate for rs328 G carriers with the metabolic syndrome. Asia Pac J Clin Nutr, 24(3):546-554.
90 | P a g e