ANTIMIC: A database of antimicrobial peptides
Manisha Brahmachary1, Judice L.Y.Koh, Mohammad Asif Khan,Seah Seng Hong Tin Wee Tan, Vladimir Bajic
1manisha@lit.org.sg, Institute of Infocomm Research
Anti-microbial peptides (AMPs) are important components of the innate immune response of many species. Many of these peptides are endowed with
(a) rapid lytic activity,
(b) activity against a broad spectrum of microbes, which include bacteria, fungi and enveloped viruses, and
(c) effective action against pathogens which are resistant to conventional antibiotics.
Such peptides have the potential to serve as natural templates for design the of novel anti-microbial peptide drugs. These new peptides can be designed with enhanced antimicrobial properties and thus there arises a need for bioinformatics-based tools to assist in such designs. Currently, there exists considerable amount of data for AMPs and it would be advantageous to have this collected and stored in a single repository. We have approached this need by creating a comprehensive database, ANTIMIC, of known and putative AMPs and integrated within it analysis tools for the convenience of wet-lab scientists. This database contains over 2000 AMP entries collected from public databases, manually checked for consistency and accuracy, and enriched in annotation. The ANTIMIC database has keyword search option and has integrated tools to aid the analysis at molecular level. The integrated tools include BLAST module for the analysis of similarity of query sequence to the ANTIMIC database content, and a peptide structure viewer module. We are currently developing two additional analysis modules, a FeatureTemplate module and an AMPpredict module. The FeatureTemplate module evaluates composition of a query peptide based on a set of physicochemical properties of amino acid residues that have structure-function relevance to AMPs. The AMPpredict module represents domains of individual AMP families as position weight matrices and calculates membership of a query peptide based on this matrix. All peptides, which have the matching score of domains above, the minimum score obtained for the family are classified as putatively belonging to that family. This set of tools is not provided in other databases dedicated to AMPs. By using FeatureTemplate module it is possible to putatively determine positions of ‘critical’ and ‘non-critical’ residues to be used as a guide in designing artificial AMPs. This module can also suggest the AMP family to which the query sequence is likely to belong and its potential function.
ANTIMIC database is available for academic and non-profit users at http://sdmc.lit.org.sg/Templar/DB/Antimic/.