Fig. 4
From: Detecting false positive sequence homology: a machine learning approach
Examples of a high quality homology (a) and false-positive homology (b) clusters (OD_S data set) classified by meta-classifier w/ logistic regression. All sequences within the homology cluster (a) belong to one protein family (FAM81A1-like protein). The sequence in the false-positive homology cluster indicated by the arrow represents Aprataxin and PNK-like factor whereas other sequences represent tyrosyl-DNA phosphodiesterase