pyDock performance in 5th CAPRI edition: from docking and scoring to binding affinity predictions and other challenges

Proteins form the executive machinery underlying all the biological processes that occur within and between cells, from DNA replication to protein degradation. Although genome-scale technologies enable to clarify their large, intricate and highly dynamics networks, they fail to elucidate the detailed molecular mechanism that underlies the protein association process. Therefore, one of the most challenging objectives in biological research is to functionally characterize protein interactions by solving 3D complex structures. This is, however, not a trivial task as confirmed by the large gap that exist between the number of complexes identified by large-scale proteomics efforts and those for which high-resolution 3D experimental structures are available. For these reasons, computational docking methods, aimed to predict the binding mode of two proteins starting from the coordinates of the individual subunits, are bound to become a complementary approach to solve the structural interactome. Given its importance, the field of protein docking has experienced an explosion in recent years partially propelled by CAPRI (http://www.ebi.ac.uk/msd-srv/capri/). CAPRI (Critical Assessment of PRedicted Interaction) is a community-wide blind experiment aimed at objectively assessing the performance of computational methods for modeling protein interactions by inviting developers to test their algorithms on the same target system and quantitatively evaluating the results. In order to test pyDock, a docking scoring algorithm developed in our group, the PID (Protein Interaction and Docking) group of the BSC Life Science Department, we have participated in all the 15 targets (T46 to T58) of the 5th CAPRI edition (2010-2012). Our automated protocol confirmed to be highly successful to provide correct models in easy-to-medium difficulty protein-protein docking cases placing among the Top5 ranked groups out of more than 60 participants.