paperKB
coga / coga-kb
Help
Sign in

Chunk #114 — ONLINE METHODS — Cross-validation of differential expression

Source
Gene expression elucidates functional impact of polygenic risk for schizophrenia.
Embedded
yes

Text

We performed cross-validation of the differential expression by randomly splitting the full cohort into an 80% “discovery” cohort and 20% “replication” cohort (with equal proportions of SCZ cases and controls into the two parts of the split). This splitting process was repeated 20 times. Each time, we chose the t-statistics of the genes considered to be differentially expressed at an FDR < 5% in the discovery cohort and looked up the corresponding statistics in the independent 20% replication cohort. Across the 20 samplings, the median number of FDR < 5% differentially expressed genes was 216 (mean = 315, sd = 261, 25th percentile = 92, 75th percentile = 562). For these FDR < 5% “discovery” differentially expressed genes, the median Pearson correlation of t-statistics with the “replication” cohort was 0.79 (mean = 0.75, sd = 0.16, 25th percentile = 0.67, 75th percentile = 0.88). This strongly supports the robustness of the differential expression results described herein.