Despite these interesting findings and observations, there are a few caveats and limitations in our study that should be mentioned and need to be carefully overcome in future studies. First, we did not have a group of samples from 22q11.2 carriers without a SZ or SAD diagnosis. An inclusion of them may help us to disentangle the factors that depend only on the 22q11.2 deletions but do not necessarily contribute to SZ pathogenesis directly. Secondly, the heterogeneity between samples was relatively large, even in those from the same individuals. Here we applied advanced analytic approaches to reduce this confounder, but it will be much better to generate more samples and then select the homogenous ones for gene expression comparison. With an increased number of samples, we will also be able to apply FDR rather than nominal p-values to select DEGs for the downstream analysis, which will enhance the validity of our findings. Thirdly, it will be valuable to perform similar systematic gene expression analyses using samples at different stages of neuronal differentiation, including mature neurons.
