. 2018 Jul 17;3(4):e00053-18. doi: 10.1128/mSystems.00053-18

TABLE 1 .

Comparison of model complexity and discrimination accuracy of microbial signatures for Crohn’s disease status^a

Method	Median no. of taxa	Mean cv-AUC
selbal	12	0.8196
DESeq2	33	0.7752
edgeR	34	0.7721
ANCOM	5	0.7125
ALDEx2	31	0.8156

For each method, the table indicates the median number of taxa included in the model and the mean cv-AUC for 10 iterations of a 5-fold cross-validation process.