For traits in which the exposure–outcome associations were nominally significant in all four estimators and reached test-wide significance ( P < 0.05/9) in two out of four (Supplementary Table 20 ), we applied the outlier-robust and pleiotropy-robust estimators MR-RAPS 94 and MR-PRESSO 95 .
2
—
—
The sentences
Highly significant associations were found for an elevated HLA MHC-I GRS in multiple sclerosis and an elevated HLA MHC-II GRS in rheumatoid arthritis (Fig. 4e ).