Abstract:
Computational drug design has gained an increasing interest in the last decade, revealing promising results in the discovery of novel therapeutic candidates. Both ab initio and de novo approaches enable the study and screening of a large number of molecules in silico, thus restraining the use of expensive and time-consuming experimental techniques. This thesis regards the attempt to use different computational methods (mainly Machine Learning and Direct Coupling Analysis approach) to predict novel peptide-based inhibitors of human urokinase-type plasminogen activator (uPA), an enzyme involved in tumour growth and invasion. The introduction of the thesis provides a general overview of peptides and computer science, while the discussion covers the analysis of different sets of peptides and describes the advantages and the disadvantages of the computational models used.
