Our variant-level association tests detected rare variant associations that are not frequent enough to be captured by microarray-based studies (that is, as rare as MAF = 0.0012%). Our gene-level collapsing analyses evaluated the aggregate effect of private-to-rare functional variants, 83% of which were not detected in single-variant tests for binary traits. Among gene-level signals for which an individual variant also achieved significance, we found examples where both common and rare risk variants in these genes contributed to disease burden. This is consistent with previous work demonstrating that common and rare PTVs in FLG have similar effect sizes for the risk of early asthma43.
