We also found a series of significant associations in brain and not blood. Several of these associations were probes on chromosome 12 associated with rs11171739 or rs1701704, two SNPs nominated for Type-I diabetes. Additional associations included a series of SNPs on chromosome 17. Previous studies have noted an effect of chromosome17 SNPs on expression of the MAPT gene that is associated with risk of PD and PSP (2011; Hoglinger et al., 2011). We therefore examined expression of ILM_1710903, which maps within the coding sequence of MAPT, with association of SNPs across Chr17 and saw robust signals in both the frontal cortex and cerebellum (Fig. 5B). This effect was driven by the H1/H2 haplotype across the MAPT locus as conditioning the analysis on a proxy SNP decreased the apparent eQTL signal.
