It is conceivable, indeed probable, that certain cells or cell types (e.g., pyramidal neurons) are more salient for risk than the heterogeneous tissue evaluated here. Depending on the pattern of cell-specific gene expression, this scenario could have little or no impact on differential expression or it could diminish it somewhat. The same is true for detection of eQTLs. We do not expect, however, that the scenario will compromise the bulk of our results, all of which complement the genomic studies of this disease. Alterations of the cellular composition in SCZ versus controls might also introduce a systemic bias in the analysis of differential expression; e.g., if the proportion of neurons were reduced by 2% in SCZ versus controls, multiple neuronal genes might appear to be downregulated in SCZ. Analyses of cell composition, however, do not support global differences in the cellular composition in DLPFC tissue from SCZ versus control subjects.
