Genomes to natural products PRediction Informatics for Secondary Metabolomes (PRISM)

Microbial natural products are an invaluable source of evolved bioactive small molecules and pharmaceutical agents. Next-generation and metagenomic sequencing indicates untapped genomic potential, yet high rediscovery rates of known metabolites increasingly frustrate conventional natural product screening programs. New methods to connect biosynthetic gene clusters to novel chemical scaffolds are therefore critical to enable the targeted discovery of genetically encoded natural products. Here, we present PRISM, a computational resource for the identification of biosynthetic gene clusters, prediction of genetically encoded nonribosomal peptides and type I and II polyketides, and bio- and cheminformatic dereplication of known natural products. PRISM implements novel algorithms which render it uniquely capable of predicting type II polyketides, deoxygenated sugars, and starter units, making it a comprehensive genome-guided chemical structure prediction engine. A library of 57 tailoring reactions is leveraged for combinatorial scaffold library generation when multiple potential substrates are consistent with biosynthetic logic. We compare the accuracy of PRISM to existing genomic analysis platforms. PRISM is an open-source, user-friendly web application available at http://magarveylab.ca/prism/.     

MLViS: A Web Tool for Machine Learning-Based Virtual Screening in Early-Phase of Drug Discovery and Development

Virtual screening is an important step in early-phase of drug discovery process. Since there are thousands of compounds, this step should be both fast and effective in order to distinguish drug-like and nondrug-like molecules. Statistical machine learning methods are widely used in drug discovery studies for classification purpose. Here, we aim to develop a new tool, which can classify molecules as drug-like and nondrug-like based on various machine learning methods, including discriminant, tree-based, kernel-based, ensemble and other algorithms. To construct this tool, first, performances of twenty-three different machine learning algorithms are compared by ten different measures, then, ten best performing algorithms have been selected based on principal component and hierarchical cluster analysis results. Besides classification, this application has also ability to create heat map and dendrogram for visual inspection of the molecules through hierarchical cluster analysis. Moreover, users can connect the PubChem database to download molecular information and to create two-dimensional structures of compounds. This application is freely available through www.biosoft.hacettepe.edu.tr/MLViS/.     

dSPRINT: predicting DNA,RNA, ion,peptide and small molecule interaction sites within protein domains

Domains are instrumental in facilitating protein interactions with DNA, RNA, small molecules, ions and peptides. Identifying ligand-binding domains within sequences is a critical step in protein function annotation, and the ligand-binding properties of proteins are frequently analyzed based upon whether they contain one of these domains. To date, however, knowledge of whether and how protein domains interact with ligands has been limited to domains that have been observed in co-crystal structures; this leaves approximately two-thirds of human protein domain families uncharacterized with respect to whether and how they bind DNA, RNA, small molecules, ions and peptides. To fill this gap, we introduce dSPRINT, a novel ensemble machine learning method for predicting whether a domain binds DNA, RNA, small molecules, ions or peptides, along with the positions within it that participate in these types of interactions. In stringent cross-validation testing, we demonstrate that dSPRINT has an excellent performance in uncovering ligand-binding positions and domains. We also apply dSPRINT to newly characterize the molecular functions of domains of unknown function. dSPRINT’s predictions can be transferred from domains to sequences, enabling predictions about the ligand-binding properties of 95% of human genes. The dSPRINT framework and its predictions for 6503 human protein domains are freely available at http://protdomain.princeton.edu/dsprint.     

Protein domain-based prediction of drug/compound–target interactions and experimental validation on LIM kinases

Predictive approaches such as virtual screening have been used in drug discovery with the objective of reducing developmental time and costs. Current machine learning and network-based approaches have issues related to generalization, usability, or model interpretability, especially due to the complexity of target proteins’ structure/function, and bias in system training datasets. Here, we propose a new method “DRUIDom” (DRUg Interacting Domain prediction) to identify bio-interactions between drug candidate compounds and targets by utilizing the domain modularity of proteins, to overcome problems associated with current approaches. DRUIDom is composed of two methodological steps. First, ligands/compounds are statistically mapped to structural domains of their target proteins, with the aim of identifying their interactions. As such, other proteins containing the same mapped domain or domain pair become new candidate targets for the corresponding compounds. Next, a million-scale dataset of small molecule compounds, including those mapped to domains in the previous step, are clustered based on their molecular similarities, and their domain associations are propagated to other compounds within the same clusters. Experimentally verified bioactivity data points, obtained from public databases, are meticulously filtered to construct datasets of active/interacting and inactive/non-interacting drug/compound–target pairs (~2.9M data points), and used as training data for calculating parameters of compound–domain mappings, which led to 27,032 high-confidence associations between 250 domains and 8,165 compounds, and a finalized output of ~5 million new compound–protein interactions. DRUIDom is experimentally validated by syntheses and bioactivity analyses of compounds predicted to target LIM-kinase proteins, which play critical roles in the regulation of cell motility, cell cycle progression, and differentiation through actin filament dynamics. We showed that LIMK-inhibitor-2 and its derivatives significantly block the cancer cell migration through inhibition of LIMK phosphorylation and the downstream protein cofilin. One of the derivative compounds (LIMKi-2d) was identified as a promising candidate due to its action on resistant Mahlavu liver cancer cells. The results demonstrated that DRUIDom can be exploited to identify drug candidate compounds for intended targets and to predict new target proteins based on the defined compound–domain relationships. Datasets, results, and the source code of DRUIDom are fully-available at: https://github.com/cansyl/DRUIDom.     

LAMP: A Database Linking Antimicrobial Peptides

The frequent emergence of drug-resistant bacteria has created an urgent demand for new antimicrobial agents. Traditional methods of novel antibiotic development are almost obsolete. Antimicrobial peptides (AMPs) are now regarded as a potential solution to revive the traditional methods of antibiotic development, although, until now, many AMPs have failed in clinical trials. A comprehensive database of AMPs with information about their antimicrobial activity and cytotoxicity will help promote the process of finding novel AMPs with improved antimicrobial activity and reduced cytotoxicity and eventually accelerate the speed of translating the discovery of new AMPs into clinical or preclinical trials. LAMP, a database linking AMPs, serves as a tool to aid the discovery and design of AMPs as new antimicrobial agents. The current version of LAMP has 5,547 entries, comprising 3,904 natural AMPs and 1,643 synthetic peptides. The database can be queried using either simply keywords or combinatorial conditions searches. Equipped with the detailed antimicrobial activity and cytotoxicity data, the cross-linking and top similar AMPs functions implemented in LAMP will help enhance our current understanding of AMPs and this may speed up the development of new AMPs for medical applications. LAMP is freely available at: http://biotechlab.fudan.edu.cn/database/lamp.     

Fast and Accurate Discovery of Degenerate Linear Motifs in Protein Sequences

Linear motifs mediate a wide variety of cellular functions, which makes their characterization in protein sequences crucial to understanding cellular systems. However, the short length and degenerate nature of linear motifs make their discovery a difficult problem. Here, we introduce MotifHound, an algorithm particularly suited for the discovery of small and degenerate linear motifs. MotifHound performs an exact and exhaustive enumeration of all motifs present in proteins of interest, including all of their degenerate forms, and scores the overrepresentation of each motif based on its occurrence in proteins of interest relative to a background (e.g., proteome) using the hypergeometric distribution. To assess MotifHound, we benchmarked it together with state-of-the-art algorithms. The benchmark consists of 11,880 sets of proteins from S. cerevisiae; in each set, we artificially spiked-in one motif varying in terms of three key parameters, (i) number of occurrences, (ii) length and (iii) the number of degenerate or “wildcard” positions. The benchmark enabled the evaluation of the impact of these three properties on the performance of the different algorithms. The results showed that MotifHound and SLiMFinder were the most accurate in detecting degenerate linear motifs. Interestingly, MotifHound was 15 to 20 times faster at comparable accuracy and performed best in the discovery of highly degenerate motifs. We complemented the benchmark by an analysis of proteins experimentally shown to bind the FUS1 SH3 domain from S. cerevisiae. Using the full-length protein partners as sole information, MotifHound recapitulated most experimentally determined motifs binding to the FUS1 SH3 domain. Moreover, these motifs exhibited properties typical of SH3 binding peptides, e.g., high intrinsic disorder and evolutionary conservation, despite the fact that none of these properties were used as prior information. MotifHound is available (http://michnick.bcm.umontreal.ca or http://tinyurl.com/motifhound) together with the benchmark that can be used as a reference to assess future developments in motif discovery.     

Epitope Predictions Indicate the Presence of Two Distinct Types of Epitope-Antibody-Reactivities Determined by Epitope Profiling of Intravenous Immunoglobulins

Epitope-antibody-reactivities (EAR) of intravenous immunoglobulins (IVIGs) determined for 75,534 peptides by microarray analysis demonstrate that roughly 9% of peptides derived from 870 different human protein sequences react with antibodies present in IVIG. Computational prediction of linear B cell epitopes was conducted using machine learning with an ensemble of classifiers in combination with position weight matrix (PWM) analysis. Machine learning slightly outperformed PWM with area under the curve (AUC) of 0.884 vs. 0.849. Two different types of epitope-antibody recognition-modes (Type I EAR and Type II EAR) were found. Peptides of Type I EAR are high in tyrosine, tryptophan and phenylalanine, and low in asparagine, glutamine and glutamic acid residues, whereas for peptides of Type II EAR it is the other way around. Representative crystal structures present in the Protein Data Bank (PDB) of Type I EAR are PDB 1TZI and PDB 2DD8, while PDB 2FD6 and 2J4W are typical for Type II EAR. Type I EAR peptides share predicted propensities for being presented by MHC class I and class II complexes. The latter interaction possibly favors T cell-dependent antibody responses including IgG class switching. Peptides of Type II EAR are predicted not to be preferentially presented by MHC complexes, thus implying the involvement of T cell-independent IgG class switch mechanisms. The high extent of IgG immunoglobulin reactivity with human peptides implies that circulating IgG molecules are prone to bind to human protein/peptide structures under non-pathological, non-inflammatory conditions. A webserver for predicting EAR of peptide sequences is available at www.sysmed-immun.eu/EAR.     

Computational Approaches for Predicting Biomedical Research Collaborations

Biomedical research is increasingly collaborative, and successful collaborations often produce high impact work. Computational approaches can be developed for automatically predicting biomedical research collaborations. Previous works of collaboration prediction mainly explored the topological structures of research collaboration networks, leaving out rich semantic information from the publications themselves. In this paper, we propose supervised machine learning approaches to predict research collaborations in the biomedical field. We explored both the semantic features extracted from author research interest profile and the author network topological features. We found that the most informative semantic features for author collaborations are related to research interest, including similarity of out-citing citations, similarity of abstracts. Of the four supervised machine learning models (naïve Bayes, naïve Bayes multinomial, SVMs, and logistic regression), the best performing model is logistic regression with an ROC ranging from 0.766 to 0.980 on different datasets. To our knowledge we are the first to study in depth how research interest and productivities can be used for collaboration prediction. Our approach is computationally efficient, scalable and yet simple to implement. The datasets of this study are available at https://github.com/qingzhanggithub/medline-collaboration-datasets.     

Cell-Type-Specific Predictive Network Yields Novel Insights into Mouse Embryonic Stem Cell Self-Renewal and Cell Fate

Self-renewal, the ability of a stem cell to divide repeatedly while maintaining an undifferentiated state, is a defining characteristic of all stem cells. Here, we clarify the molecular foundations of mouse embryonic stem cell (mESC) self-renewal by applying a proven Bayesian network machine learning approach to integrate high-throughput data for protein function discovery. By focusing on a single stem-cell system, at a specific developmental stage, within the context of well-defined biological processes known to be active in that cell type, we produce a consensus predictive network that reflects biological reality more closely than those made by prior efforts using more generalized, context-independent methods. In addition, we show how machine learning efforts may be misled if the tissue specific role of mammalian proteins is not defined in the training set and circumscribed in the evidential data. For this study, we assembled an extensive compendium of mESC data: ∼2.2 million data points, collected from 60 different studies, under 992 conditions. We then integrated these data into a consensus mESC functional relationship network focused on biological processes associated with embryonic stem cell self-renewal and cell fate determination. Computational evaluations, literature validation, and analyses of predicted functional linkages show that our results are highly accurate and biologically relevant. Our mESC network predicts many novel players involved in self-renewal and serves as the foundation for future pluripotent stem cell studies. This network can be used by stem cell researchers (at http://StemSight.org) to explore hypotheses about gene function in the context of self-renewal and to prioritize genes of interest for experimental validation.     

Properties of MHC Class I Presented Peptides That Enhance Immunogenicity

T-cells have to recognize peptides presented on MHC molecules to be activated and elicit their effector functions. Several studies demonstrate that some peptides are more immunogenic than others and therefore more likely to be T-cell epitopes. We set out to determine which properties cause such differences in immunogenicity. To this end, we collected and analyzed a large set of data describing the immunogenicity of peptides presented on various MHC-I molecules. Two main conclusions could be drawn from this analysis: First, in line with previous observations, we showed that positions P4–6 of a presented peptide are more important for immunogenicity. Second, some amino acids, especially those with large and aromatic side chains, are associated with immunogenicity. This information was combined into a simple model that was used to demonstrate that immunogenicity is, to a certain extent, predictable. This model (made available at http://tools.iedb.org/immunogenicity/) was validated with data from two independent epitope discovery studies. Interestingly, with this model we could show that T-cells are equipped to better recognize viral than human (self) peptides. After the past successful elucidation of different steps in the MHC-I presentation pathway, the identification of variables that influence immunogenicity will be an important next step in the investigation of T-cell epitopes and our understanding of cellular immune responses.     

cnvHiTSeq: integrative models for high-resolution copy number variation detection and genotyping using population sequencing data

Recent advances in sequencing technologies provide the means for identifying copy number variation (CNV) at an unprecedented resolution. A single next-generation sequencing experiment offers several features that can be used to detect CNV, yet current methods do not incorporate all available signatures into a unified model. cnvHiTSeq is an integrative probabilistic method for CNV discovery and genotyping that jointly analyzes multiple features at the population level. By combining evidence from complementary sources, cnvHiTSeq achieves high genotyping accuracy and a substantial improvement in CNV detection sensitivity over existing methods, while maintaining a low false discovery rate. cnvHiTSeq is available at http://sourceforge.net/projects/cnvhitseq     

MUSiCC: a marker genes based framework for metagenomic normalization and accurate profiling of gene abundances in the microbiome

Functional metagenomic analyses commonly involve a normalization step, where measured levels of genes or pathways are converted into relative abundances. Here, we demonstrate that this normalization scheme introduces marked biases both across and within human microbiome samples, and identify sample- and gene-specific properties that contribute to these biases. We introduce an alternative normalization paradigm, MUSiCC, which combines universal single-copy genes with machine learning methods to correct these biases and to obtain an accurate and biologically meaningful measure of gene abundances. Finally, we demonstrate that MUSiCC significantly improves downstream discovery of functional shifts in the microbiome.MUSiCC is available at http://elbo.gs.washington.edu/software.html.

Electronic supplementary material

The online version of this article (doi:10.1186/s13059-015-0610-8) contains supplementary material, which is available to authorized users.     

Human T Lymphotropic Virus Type 1 protein Tax reduces histone levels

Background

Human immunodeficiency virus type 1 (HIV-1) infects cells by means of ligand-receptor interactions. This lentivirus uses the CD4 receptor in conjunction with a chemokine coreceptor, either CXCR4 or CCR5, to enter a target cell. HIV-1 is characterized by high sequence variability. Nonetheless, within this extensive variability, certain features must be conserved to define functions and phenotypes. The determination of coreceptor usage of HIV-1, from its protein envelope sequence, falls into a well-studied machine learning problem known as classification. The support vector machine (SVM), with string kernels, has proven to be very efficient for dealing with a wide class of classification problems ranging from text categorization to protein homology detection. In this paper, we investigate how the SVM can predict HIV-1 coreceptor usage when it is equipped with an appropriate string kernel.

Results

Three string kernels were compared. Accuracies of 96.35% (CCR5) 94.80% (CXCR4) and 95.15% (CCR5 and CXCR4) were achieved with the SVM equipped with the distant segments kernel on a test set of 1425 examples with a classifier built on a training set of 1425 examples. Our datasets are built with Los Alamos National Laboratory HIV Databases sequences. A web server is available at http://genome.ulaval.ca/hiv-dskernel.

Conclusion

We examined string kernels that have been used successfully for protein homology detection and propose a new one that we call the distant segments kernel. We also show how to extract the most relevant features for HIV-1 coreceptor usage. The SVM with the distant segments kernel is currently the best method described.     

Identification of Multi-Functional Enzyme with Multi-Label Classifier

Enzymes are important and effective biological catalyst proteins participating in almost all active cell processes. Identification of multi-functional enzymes is essential in understanding the function of enzymes. Machine learning methods perform better in protein structure and function prediction than traditional biological wet experiments. Thus, in this study, we explore an efficient and effective machine learning method to categorize enzymes according to their function. Multi-functional enzymes are predicted with a special machine learning strategy, namely, multi-label classifier. Sequence features are extracted from a position-specific scoring matrix with autocross-covariance transformation. Experiment results show that the proposed method obtains an accuracy rate of 94.1% in classifying six main functional classes through five cross-validation tests and outperforms state-of-the-art methods. In addition, 91.25% accuracy is achieved in multi-functional enzyme prediction, which is often ignored in other enzyme function prediction studies. The online prediction server and datasets can be accessed from the link http://server.malab.cn/MEC/.     

AntiAngioPred: A Server for Prediction of Anti-Angiogenic Peptides

The process of angiogenesis is a vital step towards the formation of malignant tumors. Anti-angiogenic peptides are therefore promising candidates in the treatment of cancer. In this study, we have collected anti-angiogenic peptides from the literature and analyzed the residue preference in these peptides. Residues like Cys, Pro, Ser, Arg, Trp, Thr and Gly are preferred while Ala, Asp, Ile, Leu, Val and Phe are not preferred in these peptides. There is a positional preference of Ser, Pro, Trp and Cys in the N terminal region and Cys, Gly and Arg in the C terminal region of anti-angiogenic peptides. Motif analysis suggests the motifs “CG-G”, “TC”, “SC”, “SP-S”, etc., which are highly prominent in anti-angiogenic peptides. Based on the primary analysis, we developed prediction models using different machine learning based methods. The maximum accuracy and MCC for amino acid composition based model is 80.9% and 0.62 respectively. The performance of the models on independent dataset is also reasonable. Based on the above study, we have developed a user-friendly web server named “AntiAngioPred” for the prediction of anti-angiogenic peptides. AntiAngioPred web server is freely accessible at http://clri.res.in/subramanian/tools/antiangiopred/index.html (mirror site: http://crdd.osdd.net/raghava/antiangiopred/).     

A database of natural products and chemical entities from marine habitat

Marine compound database consists of marine natural products and chemical entities, collected from various literature sources, which are known to possess bioactivity against human diseases. The database is constructed using html code. The 12 categories of 182 compounds are provided with the source, compound name, 2-dimensional structure, bioactivity and clinical trial information. The database is freely available online and can be accessed at http://www.progenebio.in/mcdb/index.htm     

ThioFinder: A Web-Based Tool for the Identification of Thiopeptide Gene Clusters in DNA Sequences

Thiopeptides are a growing class of sulfur-rich, highly modified heterocyclic peptides that are mainly active against Gram-positive bacteria including various drug-resistant pathogens. Recent studies also reveal that many thiopeptides inhibit the proliferation of human cancer cells, further expanding their application potentials for clinical use. Thiopeptide biosynthesis shares a common paradigm, featuring a ribosomally synthesized precursor peptide and conserved posttranslational modifications, to afford a characteristic core system, but differs in tailoring to furnish individual members. Identification of new thiopeptide gene clusters, by taking advantage of increasing information of DNA sequences from bacteria, may facilitate new thiopeptide discovery and enrichment of the unique biosynthetic elements to produce novel drug leads by applying the principle of combinatorial biosynthesis. In this study, we have developed a web-based tool ThioFinder to rapidly identify thiopeptide biosynthetic gene cluster from DNA sequence using a profile Hidden Markov Model approach. Fifty-four new putative thiopeptide biosynthetic gene clusters were found in the sequenced bacterial genomes of previously unknown producing microorganisms. ThioFinder is fully supported by an open-access database ThioBase, which contains the sufficient information of the 99 known thiopeptides regarding the chemical structure, biological activity, producing organism, and biosynthetic gene (cluster) along with the associated genome if available. The ThioFinder website offers researchers a unique resource and great flexibility for sequence analysis of thiopeptide biosynthetic gene clusters. ThioFinder is freely available at http://db-mml.sjtu.edu.cn/ThioFinder/.