‘In-silico prediction of gene function’

Alfonso Valencia Herrera, David Alejandro Juan Sopeña, Fedrico Abascal Sebastían de Erice, José María Fernández, Manuel José Gómez, Osvaldo Graña Castro, Ramón Alonso-Allende Erhardt.

Here we propose a number of options for improving current methods, including a (better) treatment of the results of sequence searches in the context of the corresponding protein families.The promise of this approach is that the information from protein interactions is a completely independent avenue for the prediction of function. The present project addresses the key technical developments required to fulfil this expectation. Beyond the prediction of function we plan to combine a number of existing methodologies to provide clues about potential protein active sites (binding sites of other proteins or cofactors). This additional information provided by methods that others and the present research group have developed will be particularly useful as a guide for the design of experimental strategies to assess protein function.The overall goal of the proposal is to combine the information provided by five different levels of function prediction (sequence-based, family-based, prediction by the combination of sequence features, prediction based on protein interactions and the prediction of binding sites) in a single prediction schema. At the technical level, the methods and their application to genomes will be part of a database schema and a publicly accessible web platform.