The differential expression observed here is smaller than that reported in earlier studies (Supplementary data file 1), but it is consistent with plausible models for average differential gene expression and the polygenic inheritance of SCZ (Supplementary Text, “Differential gene expression: expectation, variability, and power analyses”). Consider, for example, a gene for which the major determinant of differential expression is the case-control difference in allele frequency at an eQTL SNP. For that gene, the expected magnitude of differential expression fold change will be on the order of the allele frequency differences seen in the recent large Psychiatric Genomic Consortium SCZ genetic association study (~1–2%)3, precisely what is observed in the CMC data. Beyond case-control difference in allele frequency, our modeling can also explain the difference between earlier studies and CMC results (Supplementary Fig. 9); because earlier studies tend to be far smaller in sample size, their larger differential expression is consistent with either the well-known “Winner’s Curse”41 or false positives that may occur in smaller samples. Finally, our results imply a need for thousands of samples to ensure 80% statistical power
