Directly comparing expression datasets derived from brain and whole blood in human samples is difficult because brain samples are taken post-mortem whereas blood samples are routinely taken during life. Therefore, we used two large, well-powered series from different sets of individuals to maximize our ability to find eQTLs in each tissue type. For brain, we expanded our previous dataset (Gibbs et al., 2010) in frontal cortex and cerebellum and obtained whole blood from 712 individuals from the InCHIANTI study (Wood et al., 2011). For consistency, we used the same expression array platform (Illumina HT-12 beadchips containing 48,000 probes) for all samples. After quality control, the brain mRNA dataset included 399 samples with data at 9000 probes. The blood dataset included 501 samples containing expression data from 5094 probes. Following imputation and quality control, ~ 2.2 million SNPs were available for analysis in all sample sets.
