Fig. 7
GO term and cancer gene set enrichment. a Heatmap showing enrichment −log10(p-values) of top GO terms (rows) for six gene groups (columns): all “inactive genes” (< 1.0 RPKM), and subsets of those genes assigned to each of five ChromGene annotations associated with lower expression. Asterisks denote significant enrichments after multiple testing correction (“Methods”). Square brackets denote the number of genes in each GO term or gene group. The first row corresponds to the most enriched GO term based on expression only. The second through sixth rows correspond to the most enriched GO term for each of the subsets corresponding to the ChromGene annotations listed in the second through sixth columns, respectively. “Sensory perception of smell” is listed twice since it was most enriched based on both expression only and the subset with the “quiescent” ChromGene annotation. b As in a, but for highly expressed genes (> 100 RPKM) and using the three ChromGene annotations with the highest overall expression, “strong_trans_enh,” “strong_trans,” and “trans_enh.” c Count of how often each ChromGene annotation yielded significant enrichment p-values (adjusted p < 0.01, Bonferroni corrected for the number of combinations of 127 cell types, 12 ChromGene annotations, and gene sets) across 6036 “biological process” GO term gene sets. d As in c, but for 967 cancer gene sets