Contrary to the reductionist approach aimed at understanding individual components, the new data revolution allows the understanding of complicated interactions and pathways through the use of statistical Data Mining (DM). This is an active area of research in mathematic/computer science with the increasing availability of big data collections of all sorts. In the last years, we have been applying a variety of DM algorithms for the classification of structural interactions at the protein level. In particular, we have developed  and optimized  new ways of detecting Hot-Spots, crucial residues at a protein-protein (PP) binding interface. A new web-server, SpotON, is freely available online to perform the identification and classification of these residues at a soluble PP interface (http://milou.science.uu.nl/cgi/services/SPOTON/spoton/).
1. Melo R, Fieldhouse R, Melo A, Correia JDG, Cordeiro MNDS, Gümüş ZG, Costa J, Bonvin AMJJ, Moreira IS, A Machine-Learning Approach for Hot-Spot Detection at Protein-Protein Interfaces, 2016, Int J Mol Sci, 27;17(8).
2. Moreira IS, Koukos P, Melo R, Almeida JG, Preto AJ, Schaarschmidt J, Trellet M, Gumus ZH, Costa J, Bonvin, AMJJ, SpotOne: a web server for prediction of protein-protein binding hot-spots, 2017, Sci Rep, 7(1):8007.