Statistical analysis was conducted using SPSS software v22 (IBM, Armonk, NY). Linear mixed-effects models were used to compare gene expression as a function of genotype or cluster membership to account for biological replicates for a subset of lines (biological replicates were available for 15 lines for qPCR data and 6 lines for RNA sequence data). Analysis was repeated using a single value per donor subject by averaging biological replicates and using t-tests for comparison of group means. Statistically significant contrasts identified using mixed-effects models were also seen using t-tests. Pearson correlations and two-step cluster analysis were used to examine the correlation in expression of GABAA genes and to group samples based on patterns of chromosome 4 GABAA gene expression. One-way ANOVA was used to compare adult brain RNA expression as a function of genotype. P values of ≤0.05 were considered significant.
