🧬Computational Genomics Unit 8 – Population Genomics and GWAS

Key Concepts in Population Genomics

• Population genomics studies genetic variation within and between populations to understand evolutionary processes and population structure
• Genetic diversity refers to the total number of genetic characteristics in the genetic makeup of a species
• Genetic drift is the change in allele frequencies in a population due to random sampling of organisms
• Natural selection is the process whereby organisms better adapted to their environment tend to survive and produce more offspring
• Gene flow is the transfer of genetic variation from one population to another through migration or admixture
• Linkage disequilibrium (LD) is the non-random association of alleles at different loci in a given population
  • LD can be influenced by factors such as population structure, selection, and recombination rates
• Hardy-Weinberg equilibrium (HWE) is a state in which allele and genotype frequencies remain constant from generation to generation in the absence of evolutionary influences

Genetic Variation and Population Structure

• Single nucleotide polymorphisms (SNPs) are the most common type of genetic variation used in population genomics studies
• Copy number variations (CNVs) and insertions/deletions (indels) also contribute to genetic variation within populations
• Population structure refers to the presence of genetically distinct subgroups within a population
• Principal component analysis (PCA) is a statistical method used to visualize and assess population structure
  • PCA reduces high-dimensional genetic data into a smaller number of principal components that capture the majority of the variation
• Admixture analysis estimates the proportions of an individual's genome that originate from different ancestral populations
• F-statistics (Fst) measure the degree of genetic differentiation between populations
  • Fst values range from 0 (no differentiation) to 1 (complete differentiation)
• Isolation by distance (IBD) is a pattern where genetic differences between populations increase with geographic distance due to limited gene flow

GWAS Fundamentals

• Genome-wide association studies (GWAS) aim to identify genetic variants associated with traits or diseases in a population
• GWAS typically use a case-control design, comparing allele frequencies between individuals with (cases) and without (controls) a specific phenotype
• The common disease-common variant (CDCV) hypothesis suggests that common diseases are influenced by common genetic variants with small effect sizes
• Genotyping arrays are used to simultaneously genotype hundreds of thousands to millions of SNPs across the genome
• Imputation is the process of inferring unobserved genotypes based on reference panels and linkage disequilibrium patterns
• Multiple testing correction is essential in GWAS to control for false positives due to the large number of statistical tests performed
  • Bonferroni correction and false discovery rate (FDR) are commonly used methods for multiple testing correction
• Manhattan plots visualize GWAS results, with the negative logarithm of the p-value plotted against the genomic position of each SNP

Data Collection and Quality Control

• Study design considerations for GWAS include sample size, case-control ratio, and population stratification
• Genotyping quality control (QC) steps are crucial to ensure the accuracy and reliability of GWAS results
• SNP QC measures include call rate, minor allele frequency (MAF), and Hardy-Weinberg equilibrium (HWE) testing
  • SNPs with low call rates, low MAF, or deviations from HWE are often excluded from analysis
• Sample QC measures include individual call rate, heterozygosity, and relatedness checks
  • Samples with low call rates, extreme heterozygosity, or cryptic relatedness may be removed
• Population stratification can lead to spurious associations and is often addressed using principal component analysis (PCA) or mixed models
• Batch effects can arise from technical factors (genotyping platform, lab, or processing date) and should be identified and corrected
• Phenotype data quality is equally important, with considerations for phenotype definition, measurement, and harmonization across studies

Statistical Methods in GWAS

• Single-SNP association tests, such as the chi-square test or logistic regression, are used to assess the association between each SNP and the phenotype of interest
• Linear regression is used for quantitative traits, while logistic regression is used for binary traits (case-control studies)
• Covariates, such as age, sex, and principal components, can be included in the regression models to adjust for potential confounding factors
• Mixed linear models (MLMs) are used to account for population structure and cryptic relatedness by incorporating a kinship matrix
• Meta-analysis combines GWAS results from multiple studies to increase statistical power and identify robust associations
  • Fixed-effect and random-effect models are used depending on the heterogeneity of effect sizes across studies
• Bayesian methods, such as Bayesian variable selection regression (BVSR), can be used to prioritize SNPs and estimate their effect sizes
• Polygenic risk scores (PRS) aggregate the effects of multiple SNPs to predict an individual's risk for a specific trait or disease

Interpreting GWAS Results

• Genome-wide significance threshold is typically set at $p < 5 \times 10^{-8}$ to account for multiple testing in GWAS
• Locus zoom plots visualize the association signals and linkage disequilibrium patterns in a specific genomic region
• Functional annotation of GWAS hits involves integrating information from various sources (e.g., gene expression, epigenetics, and biological pathways) to understand their potential functional impact
• Heritability estimates the proportion of phenotypic variance explained by genetic factors and can be calculated using GWAS summary statistics
• Genetic correlation analysis assesses the shared genetic basis between different traits or diseases using GWAS summary statistics
• Mendelian randomization uses genetic variants as instrumental variables to infer causal relationships between exposures and outcomes
• Replication of GWAS findings in independent cohorts is essential to validate the associations and assess their generalizability

Challenges and Limitations

• Missing heritability refers to the gap between the heritability estimates from family studies and the variance explained by GWAS-identified variants
• Rare variants (MAF < 1%) are not well captured by standard GWAS genotyping arrays and may require sequencing-based approaches
• Gene-environment interactions can modulate the effect of genetic variants on the phenotype but are often not accounted for in GWAS
• Phenotypic heterogeneity, where different genetic variants contribute to different subtypes of a disease, can reduce the power of GWAS
• Population-specific genetic effects may limit the transferability of GWAS findings across diverse populations
• Biological interpretation of GWAS results can be challenging, as associated variants may not be the causal variants and may affect genes or regulatory elements distant from the SNP
• Ethical considerations, such as informed consent, data privacy, and the potential for genetic discrimination, must be addressed in GWAS

Applications and Future Directions

• Drug target discovery and repositioning: GWAS can identify novel therapeutic targets and suggest potential drug repurposing opportunities
• Precision medicine: GWAS findings can inform personalized risk prediction, diagnosis, and treatment strategies
• Integration of multi-omics data (transcriptomics, epigenomics, and proteomics) can provide a more comprehensive understanding of the biological mechanisms underlying GWAS associations
• Fine-mapping and functional validation studies are necessary to pinpoint the causal variants and elucidate their functional consequences
• Transethnic GWAS and meta-analyses can improve the power to detect associations and assess the generalizability of findings across diverse populations
• Polygenic risk scores (PRS) have the potential to improve risk stratification and targeted interventions, but their clinical utility and ethical implications need to be carefully considered
• Machine learning and artificial intelligence approaches can be applied to GWAS data to improve risk prediction, identify novel associations, and uncover complex genetic architectures
• Collaboration and data sharing among researchers, institutions, and countries are crucial to accelerate progress in GWAS and translate the findings into tangible benefits for human health