Testing gene sets associated with normal neurocognitive variation for enrichment of association with bipolar disorder.

q-value, obtained from 3 GSEA runs with 1,000 permutations each). The maximum standard deviation from the average q-value was 0.07. Sets that passed the enrichment threshold (p-value≤0.05, FDR q-value≤0.25) were tested for validation using random mimic sets (see Table S4 in File S1). For each GWAS dataset, the 5 most enriched candidate sets are shown. For the German dataset, the 14 most enriched sets are presented to show the overlap with the other datasets. The rank position (R) of the gene set within the total number of gene sets tested is determined by the average false discovery rate (

% of the random sets (i.e. validated sets).^a indicates sets that were more enriched than 98

^b indicates sets that did not pass the enrichment threshold but were among the 5 most enriched in the corresponding sample.

“n.e.”. Visuospatial attention.1 – Visuospatial attention task with valid cue to the location of the visual target; Visuospatial attention.3 – Visuospatial attention task with neutral cue to the location of the visual target. The number after each gene set name represents the number of genes within that set (e.g. the Colour-word interference −25 set contains the top 25 genes within the colour-word interference ranking list of genes). Sets that did not pass the enrichment threshold and ranked outside the top 5 are indicated by

p-value of zero (0.0) indicates an actual p-value of less than 1/number-of-permutations. A reported