2.3. S-LDSC

LDSC is a computational framework that extracts multiple aspects of genetic architecture from GWAS. S-LDSC can be used to measure the heritability explained by a given genomic feature by partitioning SNPs into the regions of interest accounting for linkage disequilibrium (LD). Therefore, we performed S-LDSC (v1.0.1) using the generated LDSC annotation files [15]. The baseline model (v1.1.0) was used to assess heritability enrichment compared with the basic annotation of the genome. The S-LDSC results in heritability enrichment values and p-values that mark the significance of enrichment. Since genes have been previously shown to be enriched for heritability [15], high enrichment values are expected with the gene-centric approach that we undertook (e.g., DEGs). Thus, we mainly focused on the significance of enrichment by measuring the statistical significance of heritability enrichment (LDSC p-values) to confirm that the heritability enrichment explained by a given feature was not due to chance. When multiple gene sets were evaluated (e.g., kME/kTotal groups), we calculated FDR values using Benjamini and Hochberg (BH) to control for multiple tests. Gene sets with an enrichment value > 1 and an FDR < 0.05 were highlighted as enriched for schizophrenia heritability.