Extended Data Fig. 3: FUGAsseM achieves high accuracy and generalizability for GO term prediction.
From: Predicting functions of uncharacterized gene products from microbial communities
(a-c) FUGAsseM achieved AUROCs comparable to state-of-the-art methods across GO categories: BP (n = 1,010 term-species pairs), MF (n = 1,282 term-species pairs), and CC (n = 92 term-species pairs). (d-e) When predicting BP annotations with newly accumulated experimental evidence, FUGAsseM-MTX (d) and FUGAsseM-full (e) maintain high accuracy with community-wide data compared to those with isolate-based data (n = 34 total terms for temporal hold-out evaluation). ‘Accumulated evidence’: the annotations that lacked experimental evidence at T0 but gained accumulated experimental validation from T0 to T1; ‘New evidence’: totally unseen annotations at T0 with accumulated experimental validation at T1. 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).
