Analysis of whole genome sequenced cases and controls shows that the association of variants in TOMM40, BCAM, NECTIN2 and APOC1 with late onset Alzheimer's disease is driven by linkage disequilibrium with APOE ε2/ε3/ε4 alleles

Variants in APOE are associated with risk of late onset Alzheimer's disease (LOAD) but the magnitude of the effect has been reported to vary across ancestries. Also, other variants in the region have been reported to show association though it has been unclear whether this was secondary to their linkage disequilibrium with the APOE variants rs429358 and rs7412. Previous analyses of exome-sequenced samples have identified other genes in which rare variants impact risk of disease. In this study 2000 whole genome sequenced cases and controls with different ancestries were subjected to gene-based weighted burden analysis to identify risk genes. Additionally, individual variants in the APOE region were tested for association with LOAD. When using the APOE variants as covariates no individual genes showed statistically significant evidence for association after Bonferroni correction for multiple testing, which may well be a consequence of the modest sample size. Likewise, for those variants initially showing evidence of association with LOAD incorporating the APOE variants as covariates dramatically reduced the strength of association. These results demonstrate that the differential association of APOE across ancestries does not appear to be driven by another variant in the region. It seems likely that no other genes in the region have a direct effect on LOAD risk.