A prediction tool for classification of AMPs


Antimicrobial peptides (AMPs) are being explored as substitutes to antibiotics. The broad range of activity of AMPs, multiple cellular targets and their ability to inhibit multi-drug resistant microbes make them ideal candidates as future drugs.

To explore AMPs as drugs it is essential to understand sequence-specificity relationship of AMPs. ClassAMP has been developed using Random Forests method to understand the role of sequence of AMPs in its specificity towards bacteria, fungii and viruses.

Two algorithms have been developed for prediction of antibacterial, antifungal and antiviral peptides based on their sequence features. The models were trained and tested on sequences from the CAMP database. The antibacterial, antifungal and antiviral models gave an MCC of 0.92, 0.83 and 0.96 respectively.

A Help option explaining the input format for prediction and the results has been provided.

This work has been funded by grants from Department of Science & Technology, India and Indian Council of Medical Research.


People involved:

Ms. Shaini Joseph
NIRRH, Mumbai
Mr. Shreyas Karnik
School of Informatics, Indiana University
Mr. Ram Shankar Barai
NIRRH, Mumbai
Mr. Pravin Nilawe
NIRRH, Mumbai
Dr. V. K. Jayaraman
CDAC, Pune
Dr. Susan Idicula-Thomas
NIRRH, Mumbai

Contact us: