ABDpred: Prediction of active antimicrobial compounds using supervised machine learning techniques.

BACKGROUND OBJECTIVES: Discovery of new antibiotics is the need of the hour to treat infectious diseases. An ever-increasing repertoire of multidrug-resistant pathogens poses an imminent threat to human lives across the globe. However, the low success rate of the existing approaches and technologies for antibiotic discovery remains a major bottleneck. In silico methods like machine learning (ML) deem more promising to meet the above challenges compared with the conventional experimental approaches. The goal of this study was to create ML models that may be used to successfully predict new antimicrobial compounds. METHODS: In this article, we employed eight different ML algorithms namely, extreme gradient boosting, random forest, gradient boosting classifier, deep neural network, support vector machine, multilayer perceptron, decision tree, and logistic regression. These models were trained using a dataset comprising 312 antibiotic drugs and a negative set of 936 non-antibiotic drugs in a five-fold cross validation approach. RESULTS: The top four ML classifiers (extreme gradient boosting, random forest, gradient boosting classifier and deep neural network) were able to achieve an accuracy of 80 per cent and above during the evaluation of testing and blind datasets. INTERPRETATION CONCLUSIONS: We aggregated the top performing four models through a soft-voting technique to develop an ensemble-based ML method and incorporated it into a freely accessible online prediction server named ABDpred ( http://clinicalmedicinessd.com.in/abdpred/ ).

MYCbase: a database of functional sites and biochemical properties of Myc in both normal and cancer cells.

BACKGROUND: Myc is an essential gene having multiple functions such as in cell growth, differentiation, apoptosis, genomic stability, angiogenesis, and disease biology. A large number of researchers dedicated to Myc biology are generating a substantial amount of data in normal and cancer cells/tissues including Burkitt's lymphoma and ovarian cancer. RESULTS: MYCbase ( http://bicresources.jcbose.ac.in/ssaha4/mycbase ) is a collection of experimentally supported functional sites in Myc that can influence the biological cellular processes. The functional sites were compiled according to their role which includes mutation, methylation pattern, post-translational modifications, protein-protein interactions (PPIs), and DNA interactions. In addition, biochemical properties of Myc are also compiled, which includes metabolism/pathway, protein abundance, and modulators of protein-protein interactions. The OMICS data related to Myc- like gene expression, proteomics expression using mass-spectrometry and miRNAs targeting Myc were also compiled in MYCbase. The mutation and pathway data from the MYCbase were analyzed to look at the patterns and distributions across different diseases. There were few proteins/genes found common in Myc-protein interactions and Myc-DNA binding, and these can play a significant role in transcriptional feedback loops. CONCLUSION: In this report, we present a comprehensive integration of relevant information regarding Myc in the form of MYCbase. The data compiled in MYCbase provides a reliable data resource for functional sites at the residue level and biochemical properties of Myc in various cancers.

PSCRIdb: A database of regulatory interactions and networks of pluripotent stem cell lines.

Pluripotency in stem cells is regulated by a complex network between the transcription factors, signaling molecules, mRNAs, and epigenetic regulators like non-coding RNAs. Different pluripotent stem cell (PSC) lines were isolated and characterized to study the regulatory network topology to understand the mechanism that control developmental potential of pluripotent cells. PSCRIdb is a manually curated database of regulatory interactions including protein-protein, protein-DNA, gene-gene, and miRNA-mRNA interactions in mouse and human pluripotent stem cells including embryonic stem cells and embryonic carcinoma cells. At present, 22 different mouse and human pluripotent stem-cell-line-specific regulatory interactions are compiled in the database. Detailed information of the four types of interaction data are presented in tabular format and graphical network view in Cytoscape layout. The database is available at http://bicresources.jcbose.ac.in/ ssaha4/pscridb. The database contains 3037 entries of experimentally validated molecular interactions that can be useful for systematic study of pluripotency integrating multi-omics data. In summary, the database can be a useful resource for identification of regulatory networks present in different pluripotent stem cell lines.

LMDIPred: A web-server for prediction of linear peptide sequences binding to SH3, WW and PDZ domains.

Protein-peptide interactions form an important subset of the total protein interaction network in the cell and play key roles in signaling and regulatory networks, and in major biological processes like cellular localization, protein degradation, and immune response. In this work, we have described the LMDIPred web server, an online resource for generalized prediction of linear peptide sequences that may bind to three most prevalent and well-studied peptide recognition modules (PRMs)-SH3, WW and PDZ. We have developed support vector machine (SVM)-based prediction models that achieved maximum Matthews Correlation Coefficient (MCC) of 0.85 with an accuracy of 94.55% for SH3, MCC of 0.90 with an accuracy of 95.82% for WW, and MCC of 0.83 with an accuracy of 92.29% for PDZ binding peptides. LMDIPred output combines predictions from these SVM models with predictions using Position-Specific Scoring Matrices (PSSMs) and string-matching methods using known domain-binding motif instances and regular expressions. All of these methods were evaluated using a five-fold cross-validation technique on both balanced and unbalanced datasets, and also validated on independent datasets. LMDIPred aims to provide a preliminary bioinformatics platform for sequence-based prediction of probable binding sites for SH3, WW or PDZ domains.

PPIMpred: a web server for high-throughput screening of small molecules targeting protein-protein interaction.

PPIMpred is a web server that allows high-throughput screening of small molecules for targeting specific protein-protein interactions, namely Mdm2/P53, Bcl2/Bak and c-Myc/Max. Three different kernels of support vector machine (SVM), namely, linear, polynomial and radial basis function (RBF), and two other machine learning techniques including Naive Bayes and Random Forest were used to train the models. A fivefold cross-validation technique was used to measure the performance of these classifiers. The RBF kernel of SVM outperformed and/or was comparable with all other methods with accuracy values of 83%, 79% and 90% for Mdm2/P53, Bcl2/Bak and c-Myc/Max, respectively. About 80% of the predicted SVM scores of training/testing datasets from Mdm2/P53 and Bcl2/Bak have significant IC50 values and docking scores. The proposed models achieved an accuracy of 66-90% with blind sets. The three mentioned (Mdm2/P53, Bcl2/Bak and c-Myc/Max) proposed models were screened in a large dataset of 265 242 small chemicals from National Cancer Institute open database. To further realize the robustness of this approach, hits with high and random SVM scores were used for molecular docking in AutoDock Vina wherein the molecules with high and random predicted SVM scores yielded moderately significant docking scores (p-values < 0.1). In addition to the above-mentioned classification scheme, this web server also allows users to get the structural and chemical similarities with known chemical modulators or drug-like molecules based on Tanimoto coefficient similarity search algorithm. PPIMpred is freely available at http://bicresources.jcbose.ac.in/ssaha4/PPIMpred/.

LMPID: a manually curated database of linear motifs mediating protein-protein interactions.

Linear motifs (LMs), used by a subset of all protein-protein interactions (PPIs), bind to globular receptors or domains and play an important role in signaling networks. LMPID (Linear Motif mediated Protein Interaction Database) is a manually curated database which provides comprehensive experimentally validated information about the LMs mediating PPIs from all organisms on a single platform. About 2200 entries have been compiled by detailed manual curation of PubMed abstracts, of which about 1000 LM entries were being annotated for the first time, as compared with the Eukaryotic LM resource. The users can submit their query through a user-friendly search page and browse the data in the alphabetical order of the bait gene names and according to the domains interacting with the LM. LMPID is freely accessible at http://bicresources.jcbose. ac.in/ssaha4/lmpid and contains 1750 unique LM instances found within 1181 baits interacting with 552 prey proteins. In summary, LMPID is an attempt to enrich the existing repertoire of resources available for studying the LMs implicated in PPIs and may help in understanding the patterns of LMs binding to a specific domain and develop prediction model to identify novel LMs specific to a domain and further able to predict inhibitors/modulators of PPI of interest.

Prediction of Intra-Species Protein-Protein Interactions in Enteropathogens Facilitating Systems Biology Study.

Protein-protein interactions in Escherichia coli (E. coli) has been studied extensively using high throughput methods such as tandem affinity purification followed by mass spectrometry and yeast two-hybrid method. This can in turn be used to understand the mechanisms of bacterial cellular processes. However, experimental characterization of such huge amount of interactions data is not available for other important enteropathogens. Here, we propose a support vector machine (SVM)-based prediction model using the known PPIs data of E. coli that can be used to predict PPIs in other enteropathogens, such as Vibrio cholerae, Salmonella Typhi, Shigella flexneri and Yersinia entrocolitica. Different features such as domain-domain association (DDA), network topology, and sequence information were used in developing the SVM model. The proposed model using DDA, degree and amino acid composition features has achieved an accuracy of 82% and 62% on 5-fold cross validation and blind E. coli datasets, respectively. The predicted interactions were validated by Gene Ontology (GO) semantic similarity measure and String PPIs database (experimental PPIs only). Finally, we have developed a user-friendly webserver named EnPPIpred to predict intra-species PPIs in enteropathogens, which will be of great help for the experimental biologists. The webserver EnPPIpred is freely available at http://bicresources.jcbose.ac.in/ssaha4/EnPPIpred/.

Ring-Hydroxylating Oxygenase database: a database of bacterial aromatic ring-hydroxylating oxygenases in the management of bioremediation and biocatalysis of aromatic compounds.

Bacterial Rieske-type aromatic ring-hydroxylating oxygenases (RHOs) constitute a large family of enzymes, primarily involved in bioremediation of diverse aromatic compounds in the environment. In the present study, we have designed a manually curated database, Ring-Hydroxylating Oxygenase database (RHObase), which provides comprehensive information on all biochemically characterized bacterial RHOs. It consists of ∼ 1000 entries including 196 oxygenase α-subunits, 153 oxygenase ß-subunits, 92 ferredoxins and 110 reductases, distributed among 131 different bacterial strains implementing a total of 318 oxygenation reactions. For each protein, users can get detailed information about its structure and conserved domain(s) with motif signature. RHObase allows users to search a query, based on organism, oxygenase, substrate, or protein structure. In addition, this resource provides analysis tools to perform blast search against RHObase for prediction of putative substrate(s) for the query oxygenase and its phylogenetic affiliation. Furthermore, there is an integrated cheminformatics tool to search for structurally similar compound(s) in the database vis-a-vis RHO(s) capable of transforming those compound(s). Resources in the RHObase and multiple search/display options therein are intended to provide oxygenase-related requisite information to researchers, especially working in the field of environmental microbiology and biocatalysis to attain difficult chemistry of biotechnological importance.