We carried out a replication analysis by predicting the PSI in an independent, previously reported RNA-seq dataset from superior frontal gyrus of subjects from the New South Wales Brain Tissue Resource Center (NSWBTRC) [31]. PSI of each SE was predicted based on the genotypes of the NSWBTRC subjects using the models established from CMC. Using the same criteria as for the CMC cohort resulted in 570 SE, which were used to compare the model predictions from the CMC and the NSWBTRC cohorts. We found that the predictions from the NSWBTRC dataset was consistent with that of CMC (Fig. 1F). The majority (75.4%) of the results showed positive correlation between the genetically imputed and the RNA-seq PSI. Similar to the CMC result (Fig. 1B), the p value distribution deviated from the null hypothesis of randomness (Fig. 1G).
