Peer Review File Manuscript Title: Whole genome sequencing identifies multiple loci for critical illness caused by COVID-19 Reviewer Comments & Author Rebuttals

This manuscript is an extension of work completed by this group previously evaluating COVID-19 severe outcomes in critical respiratory patients with population and mild/asymptomatic controls using sequencing. The team identifies 22 loci in Europeans and an additional 3 loci with a meta-analysis across ancestries. They also use transcriptomics and HLA classical alleles to support their findings. Overall, it is a great contribution to the field. However, there are several reservations about the approach that do deserve some attention. We thank the reviewer for their generous comments and have endeavoured to respond to each point below. the authors of HIBAG who note that increasing the CT comes with improvement in accuracy but at a cost of lower genotype call rate. We set genotypes per HLA allele as 0, 1, 2 with calls less than the CT set to missing. To verify the robustness to the choice of call threshold for posterior probability, we performed the association tests without any call threshold (CT=0), with CT=0.5, and with CT=0.7. We found that DRB1*04:01 remained the only significant locus across each of the three CT values tested with the Odds Ratios and P-values being extremely consistent across all alleles (Supplementary Figure 31, also pasted below for convenience).