Abstract
BackgroundThe efficacy of antibiotics against bacterial infections is decreasing due to the development of resistance in bacteria, and thus, there is a need to search for potential alternatives to antibiotics. In this scenario, peptidoglycan hydrolases can be used as alternate antibacterial agents due to their unique property of cleaving peptidoglycan cell wall present in both gram-positive and gram-negative bacteria. Along with a role in maintaining overall peptidoglycan turnover in a cell and in daughter cell separation, peptidoglycan hydrolases also play crucial role in bacterial pathophysiology requiring development of a computational tool for the identification and classification of novel peptidoglycan hydrolases from genomic and metagenomic data.ResultsIn this study, the known peptidoglycan hydrolases were divided into multiple classes based on their site of action and were used for the development of a computational tool ‘HyPe’ for identification and classification of novel peptidoglycan hydrolases from genomic and metagenomic data. Various classification models were developed using amino acid and dipeptide composition features by training and optimization of Random Forest and Support Vector Machines. Random Forest multiclass model was selected for the development of HyPe tool as it showed up to 71.12 % sensitivity, 99.98 % specificity, 99.55 % accuracy and 0.80 MCC in four different classes of peptidoglycan hydrolases. The tool was validated on 24 independent genomic datasets and showed up to 100 % sensitivity and 0.94 MCC. The ability of HyPe to identify novel peptidoglycan hydrolases was also demonstrated on 24 metagenomic datasets.ConclusionsThe present tool helps in the identification and classification of novel peptidoglycan hydrolases from complete genomic or metagenomic ORFs. To our knowledge, this is the only tool available for the prediction of peptidoglycan hydrolases from genomic and metagenomic data.Availability: http://metagenomics.iiserb.ac.in/hype/ and http://metabiosys.iiserb.ac.in/hype/.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-016-2753-8) contains supplementary material, which is available to authorized users.
Highlights
The efficacy of antibiotics against bacterial infections is decreasing due to the development of resistance in bacteria, and there is a need to search for potential alternatives to antibiotics
Amino Acid Composition (AAC) features and Dipeptide Composition (DPC) features of randomly selected 10 % data from the total dataset, where each sequence was tagged with its respective category, were used for evaluation using ten-fold crossvalidation using WEKA
These results suggest that both LibSVM and Random Forest (RF), using AAC and DPC features and on using 10 % of the total data, performed better than the other machine learning methods, and can be further evaluated and optimized on the complete dataset
Summary
The efficacy of antibiotics against bacterial infections is decreasing due to the development of resistance in bacteria, and there is a need to search for potential alternatives to antibiotics. The compounds which act against bacterial infection either by suppressing its growth or by killing the bacterium are mainly considered as antibacterial agents such as sulfonamide derivatives and tetracycline antibiotic [1]. The bacterial cell wall is made up of glycan strands which are cross-linked by flexible peptide side chains, Sharma et al BMC Genomics (2016) 17:411 providing strength and rigidity to the bacterial cell wall [8]. The peptidoglycan of both gram-positive and gramnegative bacteria comprises of repeating units of Nacetylglucosamine (NAG) and β-(1–4)-N-acetylmuramic acid (NAM) cross-linked by peptide stem chains attached to the NAM residues [9]. The peptidoglycan layer is highly dynamic during cell growth and reshapes on division
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
