Pietro Bongini (DIISM, University of Siena)
Jan 09, 2019 – 11:00 AM
DIISM, Artificial Intelligence laboratory (room 201), Siena SI
Drug discovery is a flourishing field of research with the goal of improving human health. Even though, in the last two decades, a massive amount of informatics has been introduced in this field, providing many new research tools, the development of a new drug becomes more expensive every year. CRISPR technology adds new alternatives to cure diseases by removing DNA defects responsible of genome–related pathologies. In principle, the same technology may be exploited to induce protein mutations, that can open new surface pockets, each of which could correspond to a new active site. The effects of these mutations should be controlled by the presence of suitable ligands. We propose a machine learning approach to evaluate the druggability score of new surface pockets created by replacing large amino acids with glycine, Gly, the smallest natural amino acid. Preliminary experimental results are very promising, showing that 10\% of new created pockets are correctly predicted as druggable. A further development of this technique, which consists of classifying the pockets as druggable or not, without the help of the druggability score, is also introduced.