Software systems for automatic annotation and discovery of pharmacologically-relevant proteins at the genome level.

Nowadays it is urgent to annotate the proteome in different organisms in relation to protein function. We present a suite of programs based on machine learning, and trained on the presently available structures in order to cluster proteins according to their structural class. This makes it feasible to discriminate membrane proteins from globular ones, in order to develop specific drugs to target them. Our main products are integrated tools (BioDecoding Machines) for decoding and annotating from biological sequences, with special emphasis on trans-membrane protein detection and classification. When applicable, we build our products as web-based application servers, in order to ease the fruition of the tools by non-specialized users. We also strive to use technological standards and state-of-the-art algorithms, and we try to achieve this goal by an ongoing interaction with a world-class academic research group in bioinformatics.