esohkevin/gwas-tips

tips and tricks in genome-wide association studies - a tutorial

UNDER DEVELOPMENT

• I have removed all scripts to redevelop and redeploy this repository as a one-stop, organized, updated, and simple-to-follow GWAS tutorial.
• It will contain the tips and tricks that I have accumulated throughout my experience in analyzing highly diverse African populations.

WATCH OUT FOR UPDATES!!!

Pipeline (Workflow)

Pre-QC Association analysis

• Model: Logistic (95% confidence interval), 1df Chi square allelic test (adjusted to assess the genomic control inflation factor - λ).
• Mode of inheritance (MOI): Additive, Recessive, HetHom, Allelic and Genotypic,
• Tools: PLINK1.9, SNPTEST, R

Sample (per individual) QC

• Identification of individuals with discordant sex information.
• Identification of individuals with high missing values or outlying heterozygosities.
• Identification of duplicate or related individuals or individuals of divergent ancestry
• Tools: QCTOOL, PLINK1.9, R

SNP (per marker) QC

• Identification of SNPs with excessive missing genotype
• Exclusion of rare SNPs (MAF < 1%)
• Identification of SNPs demonstrating significant deviation from HWE
• Identification of SNPs with significant differential genotyping call rate between cases and controls
• Tools: PLINK1.9, R

Population Structure Determination

• Multidimensional scaling (eliminate population outliers)
• Principal component analysis with 10 axes of genetic variation (principal components)
• Fst and Haplotype based fine structure analysis
• Tools: Plink1.9, EIGENSOFT, fsStructure, ChromoPainter, GLOBETROTTER, R

Haplotype Estimation (phasing)

• SHAPEIT2, Eagle2

Genotype Imputation

• IMPUTE2, PBWT, MINIMACH

Post-Imputation Association analyses

• Models: Logistic regression, Linear mixed models (LMM)
• Modes of inheritance: dominant, recessive, heterozygous, additive, allelic
• Tools: PLINK1.9, SNPTEST2, R

Follow-up Imputation of putative associations

• Phasing with IMPUTE2 MCMC approach
• Imputation with IMPUTE2

Association analysis

• Models: Logistic regression, Linear mixed models (LMM)
• Modes of inheritance (MOI): dominant, recessive, heterozygous, additive, allelic
• Tools: PLINK1.9, SNPTEST2, R

License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.