Fig. 2From: A large cohort study identifying a novel prognosis prediction model for lung adenocarcinoma through machine learning strategiesSurvival-related genes selected by the three algorithms from the discovery set and functional annotation. a 1000 genes from sigFeature algorithm, 892 genes from random forest algorithm and 1373 genes from univariate Cox algorithm. There are 2472 genes in total, and 49 genes that are in the overlapping region of the three algorithms. b Top 18 biological processes enriched significantly with |Moonlight-score| > =1 and FDR < 0.05 using above 2472 genes. Increased activities highlighted in yellow and decreased in purple, green indicates the -Log10FDR. A negative z-score indicates the activity of this biological process is decreased and a positive z-score indicates the oppositeBack to article page