Skip to main content
. 2008 Nov 27;25(2):211–217. doi: 10.1093/bioinformatics/btn592

Fig. 1.

Fig. 1.

Simulation results: ability to rank enriched GO terms log10-rankings of enriched GO terms were calculated to compare the ability of methods to correctly rank these categories at the top of the list. Thus, lower ranking scores are better. Methods are LRpath, FE with the following three criteria for detecting DEGs (P<0.001, P<0.01, P<0.05, P<0.10 and P<0.50), BayGO, sigPathway (sigPath), allez and ProbCD. Initial four parameter sets (A) used 90%, 75%, 50% and 25% enrichment with DEGs, 500 total DEGs, normally distributed fold changes, two enriched categories and three replicates for treated and control groups. Subsequent groups had the following differences: (B) 1000 DEGs, (C) DEGs with higher fold changes, (D) five enriched GO terms, (E) five replicates. Data shown are averages from 30 simulation runs for each parameter set. LRpath performed significantly better than the next best methods (P=2.2×10−4 compared with FE P<0.05 and P=1.5×10−4 compared with FE P<0.10) using a Wilcoxon rank test.