Here, we demonstrated the relevance of PC1 to a range of psychiatric disorders in relation to neural development with the transcriptional analysis. Although these analyses focused on PC1, future studies may explore the significance of the remaining PCs as well as other dimensionality-reduction techniques. Also, several gene set enrichment analysis methods are available to annotate the transcriptional associations with PC1 but there is no consensus about the best practice [62, 63]. Among the three approaches tested here, GCEA identified more PC1-related GO categories than ORA, suggesting the score-based approach can be more sensitive than ORA in detecting subtle but coordinated associations within the prior gene sets. But GSEA did not discover any over-represented gene sets, suggesting that how gene scores are aggregated within gene sets matters and the rank-based score could be a conservative test statistic [64, 65]. Together with limitations in the Allen Human Brain Atlas that have been noted previously [66], the associations with molecular profiles reported in the current study need to be interpreted with caution. Apart from this, other limitations in the present study should be
