Fig. 4
From: Inter-protein residue covariation information unravels physically interacting protein dimers
Overview of our workflow. a Sequence of each protein has been searched against reference proteome for finding homologous sequences. MSA is built for both protein chains of each pair. Then, ASA thresholds are determined to select exposed residues. Finally, inter-protein residue interaction matrix is built for each pair of interacting site. b Each coevolution matrix is unified into a single "norm" quantity. Norm values are calculated for 166 dimers (83 positive and 83 negative putative pairs of proteins) in each ASA threshold which are fed into SVM as input features for further prediction. comp: protein complex (Heterodimers in our analysis)