We constructed gene co-expression networks using the WCGNA and coexpp packages in R (https://bitbucket.org/multiscale/coexpp), starting with the normalized expression data for 16,423 genes. To ensure a more robust correlation-based co-expression analysis, we first removed 5 samples as outliers based on IAC analysis, specifically those more than 4 standard deviations from the mean), leaving a final cohort for co-expression analysis consisting of 278 control samples and 254 cases with schizophrenia. We constructed gene co-expression networks separately in control individuals and SCZ cases117.
