13073_2016_290_MOESM1_ESM.pdf (19.11 kB)
Additional file 1: Figure S1. of Microbiota-based model improves the sensitivity of fecal immunochemical test for detecting colonic lesions
journal contribution
posted on 2016-04-06, 05:00 authored by Nielson Baxter, Mack Ruffin, Mary Rogers, Patrick SchlossRandom forest feature selection for detecting adenomas. (A) Change in AUC with varying number of variables in the random forest model. The model with the highest AUC contained 22 OTUs. (B) Importance of each OTU in the model as measured by mean decrease accuracy when the OTU is removed from the model. (C) Relative abundance of the most discriminatory OTUs in adenoma and normal samples. (PDF 19 kb)