compared to CommonMind eQTLs could be due to its smaller sample size. Also, the Braineac eQTLs were based on false discovery rate (FDR) correction whereas CommonMind eQTLs were defined using Bonferroni correction, and stronger associations captured by more stringent correction are more likely to replicate26. We also assessed the replication rate of our eQTLs in the CommonMind data, and estimated similar replication rate (π1=0.90). For the mQTL replication analysis, we explored restricting our mQTL analysis to a 100Kb window, and observed similar replication rate (π1=0.87) on the fetal brain mQTLs8, which suggests a 5Kb window already captures majority of the stronger associations between SNP and DNA methylation.
