Extended Data Fig. 2: FUGAsseM accurately predicts MF and CC terms in microbial communities.
From: Predicting functions of uncharacterized gene products from microbial communities
(a-d) FUGAsseM predictions significantly correlated with STRING-based results (MF: n = 95 terms, 148 species; CC: n = 12 terms, 35 species) and identified additional terms in species lacking isolates (MF: n = 27 terms, 159 species; CC: n = 2 terms, 46 species). Pearson correlation coefficients (95% CI) and unadjusted P values shown. (e,f) FUGAsseM achieved comparable accuracy to DeepGOPlus and NetGO2.0; STRING data were available for 14 of 25 (MF) and 6 of 18 (CC) abundant species. (g,h) Hold-out validation confirmed high AUROC (MF: n = 12 terms; CC: n = 10 terms). (i,j) FUGAsseM predicted significantly higher scores (GSEA method; FDR-adjusted P < 0.002 for multiple comparisons) for annotations gaining experimental support over time. Box plots display the median (line at the 50th percentile), interquartile range (box spanning the 25th to 75th percentiles), whiskers (extending to 1.5× IQR), and mean values (dark points).
