engleski

In silico search strategies for finding new AMPs

We have developed partially or fully automated knowledge-based methods: a)For finding natural peptide antibiotics exhibiting promise as lead compounds (by using an indirect search for antimicrobial peptides associated with conserved signal peptides in the EST databases in combination with the therapeutic index (TI) prediction), b)For suggesting amino acid substitutions expected to decrease toxicity, increase selectivity and maintain antibacterial activity (the Mutator algorithm which uses the TI prediction), c)For constructing novel classes of peptide antibiotics with very little homology (less than 50%) to any known natural or synthetic antimicrobial peptide (the Designer algorithm). In-silico designed peptides belong to the class of anuran-like helical peptides for which the TI-predictor http://split4.pmfst.hr/split/dserv1/ has been trained by using the DADP database http://split4.pmfst.hr/dadp/ and confirmed to work well [1-4]. Tests performed at the University of Trieste and at the University Pierre and Marie Curie (Paris) for predicted and measured TI with respect to several identified natural peptides, modified versions of known peptides and ab initio designed peptides have validated the method, with good correlation between predicted and experimental values. References: 1. D Juretić, D Vukičević, N Ilić, N Antcheva, A Tossi, J. Chem. Inf. Model. 49, 2873-2882 (2009). 2. D Juretić, D Vukičević, D Petrov, M Novković, V Bojović, B Lučić, N Ilić, A Tossi, Eur. Biophys. J. 40, 371-385 (2011). 3. V Tessera, F Guida, D Juretić, A Tossi, FEBS Journal 279, 724–736 (2012). 4. M Novković, J Simunić, V Bojović, A Tossi, D Juretić, Bioinformatics 28, 1406-1407 (2012).

antimicrobial peptides; Designer; therapeutic index

nije evidentirano