Analysis pipeline for HIV whole genome sequence (WGS) genome-wide association study (GWAS) compared to a human study using a SNP chip.

<p>Step 1) Diploidy defined for both human and pathogen, to reflect ‘real’ heterozygosity and heterozygosity from within host viral diversity. 2) While missingness and Hardy-Weinberg Equilibrium are used to assess genotyping quality in human GWAS, in viral GWAS we used depth of sequencing to assess variant calls. As such, higher calling confidence is associated with higher missingness in viral SNPs, while the reverse is true in humans. Low minor allele frequency (MAF) is always used to remove variants that have low power to detect effects and may reflect errors. 3&4) Correction for ancestry and relatedness are key to human GWAS, however due to both more homogenous sampling and difficulty in applying conventional corrections in human data to viral, this was done as a sensitivity test in a smaller sample for top SNPs in HIV GWAS.