Fig. 1.

CABS-dock pipeline. The CABS-dock modelling procedure consists of the following steps: CABS simulation (providing K models, default: 10 000); energy scoring (providing N models, default: 1000); structural clustering (providing n clusters, default: 10, default method: k-medoids); selecting cluster representatives (providing n models, default selection method: cluster medoids); reconstruction to all-atom (AA) representation [providing n models, default reconstruction method: MODELLER (Webb and Sali, 2014)]. At each modelling step, the number of models may be modified by the user. The default output contains sets of models (generated after each modelling step) in PDB format (in C-alpha representation, or all-atom after the reconstruction procedure), PNG and CSV files with residue–residue contact maps (MAP, for all models, sets of similar models found in clusters and top-scored models) and contact frequency histograms (HIST) for all models generated in CABS simulation