[DSC Croatia 2021] Artificial Intelligence Based Design of Short Peptides - Goran Mausa
The discovery of new peptides with desired properties based on their sequence is a challenging task, as they are part of a very large search space and the principles responsible for the target properties are not yet fully understood. Our approach to overcome these obstacles is the application of search-based algorithms coupled with machine learning in search of de novo peptide sequences that exhibit antimicrobial activity. We also use soft computing techniques to find peptides with catalytic activity, to predict their predisposition towards self-assembly and to estimate the antiviral activity of selected candidates