Predicting target profiles with confidence as a service using docking scores
Ontology highlight
ABSTRACT: Background
Identifying and assessing ligand-target binding is a core component in early drug discovery as one or more unwanted interactions may be associated with safety issues. Contributions
We present an open-source, extendable web service for predicting target profiles with confidence using machine learning for a panel of 7 targets, where models are trained on molecular docking scores from a large virtual library. The method uses conformal prediction to produce valid measures of prediction efficiency for a particular confidence level. The service also offers the possibility to dock chemical structures to the panel of targets with QuickVina on individual compound basis. Results
The docking procedure and resulting models were validated by docking well-known inhibitors for each of the 7 targets using QuickVina. The model predictions showed comparable performance to molecular docking scores against an external validation set. The implementation as publicly available microservices on Kubernetes ensures resilience, scalability, and extensibility.
SUBMITTER: Ahmed L
PROVIDER: S-EPMC7566026 | biostudies-literature | 2020 Jan
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Ligand-based models can be used in drug discovery to obtain an early indication of potential off-target interactions that could be linked to adverse effects. Another application is to combine such models into a panel, allowing to compare and search for compounds with similar profiles. Most contemporary methods and implementations however lack valid measures of confidence in their predictions, and only provide point predictions. We here describe a methodology that uses Conformal Prediction for predicting off-target interactions, with models trained on data from 31 targets in the ExCAPE-DB dataset selected for their utility in broad early hazard assessment. Chemicals were represented by the signature molecular descriptor and support vector machines were used as the underlying machine learning method. By using conformal prediction, the results from predictions come in the form of confidence p-values for each class. The full pre-processing and model training process is openly available as scientific workflows on GitHub, rendering it fully reproducible. We illustrate the usefulness of the developed methodology on a set of compounds extracted from DrugBank. The resulting models are published online and are available via a graphical web interface and an OpenAPI interface for programmatic access.
| S-EPMC6233526 | biostudies-literature
Project description:BackgroundIdentification of novel drug targets and their inhibitors is a major challenge in the field of drug designing and development. Diaminopimelic acid (DAP) pathway is a unique lysine biosynthetic pathway present in bacteria, however absent in mammals. This pathway is vital for bacteria due to its critical role in cell wall biosynthesis. One of the essential enzymes of this pathway is dihydrodipicolinate synthase (DHDPS), considered to be crucial for the bacterial survival. In view of its importance, the development and prediction of potent inhibitors against DHDPS may be valuable to design effective drugs against bacteria, in general.ResultsThis paper describes a methodology for predicting novel/potent inhibitors against DHDPS. Here, quantitative structure activity relationship (QSAR) models were trained and tested on experimentally verified 23 enzyme's inhibitors having inhibitory value (Ki) in the range of 0.005-22(mM). These inhibitors were docked at the active site of DHDPS (1YXD) using AutoDock software, which resulted in 11 energy-based descriptors. For QSAR modeling, Multiple Linear Regression (MLR) model was engendered using best four energy-based descriptors yielding correlation values R/q2 of 0.82/0.67 and MAE of 2.43. Additionally, Support Vector Machine (SVM) based model was developed with three crucial descriptors selected using F-stepping remove-one approach, which enhanced the performance by attaining R/q2 values of 0.93/0.80 and MAE of 1.89. To validate the performance of QSAR models, external cross-validation procedure was adopted which accomplished high training/testing correlation values (q2/r2) in the range of 0.78-0.83/0.93-0.95.ConclusionsOur results suggests that ligand-receptor binding interactions for DHDPS employing QSAR modeling seems to be a promising approach for prediction of antibacterial agents. To serve the experimentalist to develop novel/potent inhibitors, a webserver "KiDoQ" has been developed http://crdd.osdd.net/raghava/kidoq, which allows the prediction of Ki value of a new ligand molecule against DHDPS.
| S-EPMC2841597 | biostudies-literature
Project description:In polyploid genomes, homoeologs are a specific subtype of homologs, and can be thought of as orthologs between subgenomes. In Orthologous MAtrix, we infer homoeologs in three polyploid plant species: upland cotton (Gossypium hirsutum), rapeseed (Brassica napus), and bread wheat (Triticum aestivum). While we can typically recognize the features of a "good" homoeolog prediction (a consistent evolutionary distance, high synteny, and a one-to-one relationship), none of them is a hard-fast criterion. We devised a novel fuzzy logic-based method to assign confidence scores to each pair of predicted homoeologs. We inferred homoeolog pairs and used the new and improved method to assign confidence scores, which ranged from 0 to 100. Most confidence scores were between 70 and 100, but the distribution varied between genomes. The new confidence scores show an improvement over our previous method and were manually evaluated using a subset from various confidence ranges.
| S-EPMC6330999 | biostudies-literature
Project description:Late-stage or post-market identification of adverse drug reactions (ADRs) is a significant public health issue and a source of major economic liability for drug development. Thus, reliable in silico screening of drug candidates for possible ADRs would be advantageous. In this work, we introduce a computational approach that predicts ADRs by combining the results of molecular docking and leverages known ADR information from DrugBank and SIDER. We employed a recently parallelized version of AutoDock Vina (VinaLC) to dock 906 small molecule drugs to a virtual panel of 409 DrugBank protein targets. L1-regularized logistic regression models were trained on the resulting docking scores of a 560 compound subset from the initial 906 compounds to predict 85 side effects, grouped into 10 ADR phenotype groups. Only 21% (87 out of 409) of the drug-protein binding features involve known targets of the drug subset, providing a significant probe of off-target effects. As a control, associations of this drug subset with the 555 annotated targets of these compounds, as reported in DrugBank, were used as features to train a separate group of models. The Vina off-target models and the DrugBank on-target models yielded comparable median area-under-the-receiver-operating-characteristic-curves (AUCs) during 10-fold cross-validation (0.60-0.69 and 0.61-0.74, respectively). Evidence was found in the PubMed literature to support several putative ADR-protein associations identified by our analysis. Among them, several associations between neoplasm-related ADRs and known tumor suppressor and tumor invasiveness marker proteins were found. A dual role for interstitial collagenase in both neoplasms and aneurysm formation was also identified. These associations all involve off-target proteins and could not have been found using available drug/on-target interaction data. This study illustrates a path forward to comprehensive ADR virtual screening that can potentially scale with increasing number of CPUs to tens of thousands of protein targets and millions of potential drug candidates.
| S-EPMC4156361 | biostudies-literature
Project description:A set of ninety-eight B-RafV600E inhibitors was used for the development of a molecular docking based QSAR model using linear and non-linear regression models. The integration of docking scores and key interaction profiles significantly improved the accuracy of the QSAR models, providing reasonable statistical parameters (Rtrain2 = 0.935, Rtest2 = 0.728 and QCV2 = 0.905). The established MD-SVR (molecular docking based SMV regression) model as well as model screening of a natural product database was carried out and two natural products (quercetin and myricetin) with good prediction activities were biologically evaluated. Both compounds exhibited promising B-RafV600E inhibitory activities (ICQuercetin50 = 7.59 μM and ICMyricetin50 = 1.56 μM), suggesting a high reliability and good applicability of the established MD-SVR model in the future development of B-RafV600E inhibitors with high efficacy.
| S-EPMC6084233 | biostudies-literature
Project description:We examined the performance of methylation scores (MS) and polygenic scores (PGS) for birth weight, BMI, prenatal maternal smoking exposure, and smoking status to assess the extent to which MS could predict these traits and exposures over and above the PGS in a multi-omics prediction model. MS may be seen as the epigenetic equivalent of PGS, but because of their dynamic nature and sensitivity of non-genetic exposures may add to complex trait prediction independently of PGS. MS and PGS were calculated based on genotype data and DNA-methylation data in blood samples from adults (Illumina 450 K; N = 2,431; mean age 35.6) and in buccal samples from children (Illumina EPIC; N = 1,128; mean age 9.6) from the Netherlands Twin Register. Weights to construct the scores were obtained from results of large epigenome-wide association studies (EWASs) based on whole blood or cord blood methylation data and genome-wide association studies (GWASs). In adults, MSs in blood predicted independently from PGSs, and outperformed PGSs for BMI, prenatal maternal smoking, and smoking status, but not for birth weight. The largest amount of variance explained by the multi-omics prediction model was for current vs. never smoking (54.6%) of which 54.4% was captured by the MS. The two predictors captured 16% of former vs. never smoking initiation variance (MS:15.5%, PGS: 0.5%), 17.7% of prenatal maternal smoking variance (MS:16.9%, PGS: 0.8%), 11.9% of BMI variance (MS: 6.4%, PGS 5.5%), and 1.9% of birth weight variance (MS: 0.4%, PGS: 1.5%). In children, MSs in buccal samples did not show independent predictive value. The largest amount of variance explained by the two predictors was for prenatal maternal smoking (2.6%), where the MSs contributed 1.5%. These results demonstrate that blood DNA MS in adults explain substantial variance in current smoking, large variance in former smoking, prenatal smoking, and BMI, but not in birth weight. Buccal cell DNA methylation scores have lower predictive value, which could be due to different tissues in the EWAS discovery studies and target sample, as well as to different ages. This study illustrates the value of combining polygenic scores with information from methylation data for complex traits and exposure prediction.
| S-EPMC8357987 | biostudies-literature
Project description:MotivationHigh-throughput single nucleotide polymorphism (SNP) arrays have become the standard platform for linkage and association analyses. The high SNP density of these platforms allows high-resolution identification of ancestral recombination events even for distant relatives many generations apart. However, such inference is sensitive to marker mistyping and current error detection methods rely on the genotyping of additional close relatives. Genotyping algorithms provide a confidence score for each marker call that is currently not integrated in existing methods. There is a need for a model that incorporates this prior information within the standard identical by descent (IBD) and association analyses.ResultsWe propose a novel model that incorporates marker confidence scores within IBD methods based on the Lander-Green Hidden Markov Model. The novel parameter of this model is the joint distribution of confidence scores and error status per array. We estimate this probability distribution by applying a modified expectation-maximization (EM) procedure on data from nuclear families genotyped with Affymetrix 250K SNP arrays. The converged tables from two different genotyping algorithms are shown for a wide range of error rates. We demonstrate the efficacy of our method in refining the detection of IBD signals using nuclear pedigrees and distant relatives.AvailabilityPlinke, a new version of Plink with an extended pairwise IBD inference model allowing per marker error probabilities is freely available at: http://bioinfo.bgu.ac.il/bsu/software/plinke.Contactobirk@bgu.ac.il; markusb@bgu.ac.ilSupplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC3187655 | biostudies-literature
Project description: Not available
| S-EPMC3319356 | biostudies-literature
Project description:BackgroundProtein-protein interactions (PPIs) play key roles in various cellular functions. In addition, some critical inter-species interactions such as host-pathogen interactions and pathogenicity occur through PPIs. Phytopathogenic bacteria infect hosts through attachment to host tissue, enzyme secretion, exopolysaccharides production, toxins release, iron acquisition, and effector proteins secretion. Many such mechanisms involve some kind of protein-protein interaction in hosts. Our first aim was to predict the whole protein interaction pairs (interactome) of Xanthomonas oryzae pathovar oryzae (Xoo) that is an important pathogenic bacterium that causes bacterial blight (BB) in rice. We developed a detection protocol to find possibly interacting proteins in its host using whole genome PPI prediction algorithms. The second aim was to build a DB server and a bioinformatic procedure for finding target proteins in Xoo for developing pesticides that block host-pathogen protein interactions within critical biochemical pathways.DescriptionA PPI network in Xoo proteome was predicted by bioinformatics algorithms: PSIMAP, PEIMAP, and iPfam. We present the resultant species specific interaction network and host-pathogen interaction, XooNET. It is a comprehensive predicted initial PPI data for Xoo. XooNET can be used by experimentalists to pick up protein targets for blocking pathological interactions. XooNET uses most of the major types of PPI algorithms. They are: 1) Protein Structural Interactome MAP (PSIMAP), a method using structural domain of SCOP, 2) Protein Experimental Interactome MAP (PEIMAP), a common method using public resources of experimental protein interaction information such as HPRD, BIND, DIP, MINT, IntAct, and BioGrid, and 3) Domain-domain interactions, a method using Pfam domains such as iPfam. Additionally, XooNET provides information on network properties of the Xoo interactome.ConclusionXooNET is an open and free public database server for protein interaction information for Xoo. It contains 4,538 proteins and 26,932 possible interactions consisting of 18,503 (PSIMAP), 3,118 (PEIMAP), and 8,938 (iPfam) pairs. In addition, XooNET provides 3,407 possible interaction pairs between two sets of proteins; 141 Xoo proteins that are predicted as membrane proteins and rice proteomes. The resultant interacting partners of a query protein can be easily retrieved by users as well as the interaction networks in graphical web interfaces. XooNET is freely available from http://bioportal.kobic.kr/XooNET/.
| S-EPMC2246157 | biostudies-literature
Project description:We aim to determine whether ischemic stroke(IS)-related PRSs are also associated with and further predict 3-year all-cause mortality. 1756 IS patients with European ancestry were randomly split into training (n = 1226) and testing (n = 530) groups with 3-year post-event observations. Univariate Cox proportional hazards regression model (CoxPH) was used for primary screening of individual prognostic PRSs. Only the significantly associated PRSs and clinical risk factors with the same direction for a causal relationship with IS were used to construct a multivariate CoxPH. Feature selection was conducted by the LASSO method. After feature selection, a prediction model with 11 disease-associated pathway-specific PRSs outperformed the base model, as demonstrated by a higher concordance index (0.751, 95%CI [0.693-0.809] versus 0.729, 95%CI [0.676-0.782]) in the testing sample. A PRS derived from endothelial cell apoptosis showed independent predictability in the multivariate CoxPH (Hazard Ratio = 1.193 [1.027-1.385], p = 0.021). These PRSs fine-tuned the model by better stratifying high, intermediate, and low-risk groups. Several pathway-specific PRSs were associated with clinical risk factors in an age-dependent manner and further confirmed some known etiologies of IS and all-cause mortality. In conclusion, Pathway-specific PRSs for IS are associated with all-cause mortality, and the integrated multivariate risk model provides prognostic value in this context.
| S-EPMC9296485 | biostudies-literature