Dmitry Filimonov

Collaborative development of predictive toxicology applications

OpenTox provides an interoperable, standards-based Framework for the support of predictive toxicology data management, algorithms, modelling, validation and reporting. It is relevant to satisfying the chemical safety assessment requirements of the REACH legislation as it supports access to experimental data, (Quantitative) Structure-Activity Relationship models, and toxicological information through an integrating platform that adheres to regulatory requirements and OECD validation principles. Initial research defined the essential components of the Framework including the approach to data access, schema and management, use of controlled vocabularies and ontologies, architecture, web service and communications protocols, and selection and integration of algorithms for predictive modelling. OpenTox provides end-user oriented tools to non-computational specialists, risk assessors, and toxicological experts in addition to Application Programming Interfaces (APIs) for developers of new applications. OpenTox actively supports public standards for data representation, interfaces, vocabularies and ontologies, Open Source approaches to core platform components, and community-based collaboration approaches, so as to progress system interoperability goals.The OpenTox Framework includes APIs and services for compounds, datasets, features, algorithms, models, ontologies, tasks, validation, and reporting which may be combined into multiple applications satisfying a variety of different user needs. OpenTox applications are based on a set of distributed, interoperable OpenTox API-compliant REST web services. The OpenTox approach to ontology allows for efficient mapping of complementary data coming from different datasets into a unifying structure having a shared terminology and representation.Two initial OpenTox applications are presented as an illustration of the potential impact of OpenTox for high-quality and consistent structure-activity relationship modelling of REACH-relevant endpoints: ToxPredict which predicts and reports on toxicities for endpoints for an input chemical structure, and ToxCreate which builds and validates a predictive toxicity model based on an input toxicology dataset. Because of the extensible nature of the standardised Framework design, barriers of interoperability between applications and content are removed, as the user may combine data, models and validation from multiple sources in a dependable and time-effective way.

Multi-Targeted Natural Products Evaluation Based on Biological Activity Prediction with PASS

Natural products found a wide use in folk medicine. Presently, when routine development of new drugs faced a considerable challenge, they become an inspiration and valuable source in drugs discovery. Rather complex and diverse chemical structures of natural compounds provide a basis for modulation of different biological targets. Natural compounds exhibit a multitargeted action that may lead to additive/synergistic or antagonistic effects. Rational design of more safe and potent pharmaceuticals requires an estimation of probable multiple actions of natural products. Our software PASS can perform such estimation. It predicts with reasonable accuracy over 3500 pharmacotherapeutic effects, mechanisms of action, interaction with the metabolic system, and specific toxicity for drug-like molecules on the basis of their structural formulae. We analyzed PASS predictions utilizing PharmaExpert, which performs selection of compounds with multiple mechanisms of action, analysis of activity-activity relationships and drug-drug interactions. The paper describes an application of PASS and PharmaExpert to the evaluation of biological activity of natural compounds including marine sponge alkaloids, triterpenoids of lupane group, and their derivatives. Proposed computer-aided methods can generate combinatorial libraries of macrolides. They help to select the most promising pharmaceutical leads with the required properties. Case study, based on the analysis of biological activity spectra predicted for St Johns Wort constituents, clearly demonstrates capabilities of computational methods in the evaluation of multitargeted actions, additive/synergistic and/or antagonistic effects of natural products.

Quantitative prediction of antitarget interaction profiles for chemical compounds.

The evaluation of possible interactions between chemical compounds and antitarget proteins is an important task of the research and development process. Here, we describe the development and validation of QSAR models for the prediction of antitarget end-points, created on the basis of multilevel and quantitative neighborhoods of atom descriptors and self-consistent regression. Data on 4000 chemical compounds interacting with 18 antitarget proteins (13 receptors, 2 enzymes, and 3 transporters) were used to model 32 sets of end-points (IC(50), K(i), and K(act)). Each set was randomly divided into training and test sets in a ratio of 80% to 20%, respectively. The test sets were used for external validation of QSAR models created on the basis of the training sets. The coverage of prediction for all test sets exceeded 95%, and for half of the test sets, it was 100%. The accuracy of prediction for 29 of the end-points, based on the external test sets, was typically in the range of R(2)(test) = 0.6-0.9; three tests sets had lower R(2)(test) values, specifically 0.55-0.6. The proposed approach showed a reasonable accuracy of prediction for 91% of the antitarget end-points and high coverage for all external test sets. On the basis of the created models, we have developed a freely available online service for in silico prediction of 32 antitarget end-points: http://www.pharmaexpert.ru/GUSAR/antitargets.html.

Chapter 6:Probabilistic Approaches in Activity Prediction

Biological activity has a probabilistic nature, and the most appropriate approaches in activity prediction are based on the theory of probability. The statistical nature of the maximum likelihood method and the Bayesian approach is well recognized, but many other methods (multiple regression, factor...

Metabolism Site Prediction Based on Xenobiotic Structural Formulas and PASS Prediction Algorithm

A new ligand-based method for the prediction of sites of metabolism (SOMs) for xenobiotics has been developed on the basis of the LMNA (labeled multilevel neighborhoods of atom) descriptors and the PASS (prediction of activity spectra for substances) algorithm and applied to predict the SOMs of the 1A2, 2C9, 2C19, 2D6, and 3A4 isoforms of cytochrome P450. An average IAP (invariant accuracy of prediction) of SOMs calculated by the leave-one-out cross-validation procedure was 0.89 for the developed method. The external validation was made with evaluation sets containing data on biotransformations for 57 cardiovascular drugs. An average IAP of regioselectivity for evaluation sets was 0.83. It was shown that the proposed method exceeds accuracy of SOM prediction by RS-Predictor for CYP 1A2, 2D6, 2C9, 2C19, and 3A4 and is comparable to or better than SMARTCyp for CYP 2C9 and 2D6.

Why relevant chemical information cannot be exchanged without disclosing structures

SummaryBoth society and industry are interested in increasing the safety of pharmaceuticals. Potentially dangerous compounds could be filtered out at early stages of R&D by computer prediction of biological activity and ADMET characteristics. Accuracy of such predictions strongly depends on the quality & quantity of information contained in a training set. Suggestion that some relevant chemical information can be added to such training sets without disclosing chemical structures was generated at the recent ACS Symposium. We presented arguments that such safety exchange of relevant chemical information is impossible. Any relevant information about chemical structures can be used for search of either a particular compound itself or its close analogues. Risk of identifying such structures is enough to prevent pharma industry from relevant chemical information exchange.

Identification of Drug-Induced Myocardial Infarction-Related Protein Targets through the Prediction of Drug–Target Interactions and Analysis of Biological Processes

Drug-induced myocardial infarction (DIMI) is one of the most serious adverse drug effects that often lead to death. Therefore, the identification of DIMI at the early stages of drug development is essential. For this purpose, the in vitro testing and in silico prediction of interactions between drug-like substances and various off-target proteins associated with serious adverse drug reactions are performed. However, only a few DIMI-related protein targets are currently known. We developed a novel in silico approach for the identification of DIMI-related protein targets. This approach is based on the computational prediction of drug-target interaction profiles based on information from approximately 1738 human targets and 828 drugs, including 254 drugs that cause myocardial infarction. Through a statistical analysis, we revealed the 155 most significant associations between protein targets and DIMI. Because not all of the identified associations may lead to DIMI, an analysis of the biological functions of these proteins was performed. The Random Walk with Restart algorithm based on a functional linkage gene network was used to prioritize the revealed DIMI-related protein targets according to the functional similarity between their genes and known genes associated with myocardial infarction. The biological processes associated with the 155 selected protein targets were determined by gene ontology and pathway enrichment analysis. This analysis indicated that most of the processes leading to DIMI are associated with atherosclerosis. The revealed proteins were manually annotated with biological processes using functional and disease-related data extracted from the literature. Finally, the 155 protein targets were classified into three categories of confidence: (1) high (the protein targets are known to be involved in DIMI via atherosclerotic progression; 50 targets), (2) medium (the proteins are known to participate in biological processes related with DIMI; 65 targets), and (3) low (the proteins are indirectly involved in DIMI pathogenesis; 40 proteins).

Prediction of reacting atoms for the major biotransformation reactions of organic xenobiotics

BackgroundThe knowledge of drug metabolite structures is essential at the early stage of drug discovery to understand the potential liabilities and risks connected with biotransformation. The determination of the site of a molecule at which a particular metabolic reaction occurs could be used as a starting point for metabolite identification. The prediction of the site of metabolism does not always correspond to the particular atom that is modified by the enzyme but rather is often associated with a group of atoms. To overcome this problem, we propose to operate with the term “reacting atom”, corresponding to a single atom in the substrate that is modified during the biotransformation reaction. The prediction of the reacting atom(s) in a molecule for the major classes of biotransformation reactions is necessary to generate drug metabolites.ResultsSubstrates of the major human cytochromes P450 and UDP-glucuronosyltransferases from the Biovia Metabolite database were divided into nine groups according to their reaction classes, which are aliphatic and aromatic hydroxylation, N- and O-glucuronidation, N-, S- and C-oxidation, and N- and O-dealkylation. Each training set consists of positive and negative examples of structures with one labelled atom. In the positive examples, the labelled atom is the reacting atom of a particular reaction that changed adjacency. Negative examples represent non-reacting atoms of a particular reaction. We used Labelled Multilevel Neighbourhoods of Atoms descriptors for the designation of reacting atoms. A Bayesian-like algorithm was applied to estimate the structure–activity relationships. The average invariant accuracy of prediction obtained in leave-one-out and 20-fold cross-validation procedures for five human isoforms of cytochrome P450 and all isoforms of UDP-glucuronosyltransferase varies from 0.86 to 0.99 (0.96 on average).ConclusionsWe report that reacting atoms may be predicted with reasonable accuracy for the major classes of metabolic reactions—aliphatic and aromatic hydroxylation, N- and O-glucuronidation, N-, S- and C-oxidation, and N- and O-dealkylation. The proposed method is implemented as a freely available web service at http://www.way2drug.com/RA and may be used for the prediction of the most probable biotransformation reaction(s) and the appropriate reacting atoms in drug-like compounds.Graphical abstract.

In Silico Identification of Proteins Associated with Drug-Induced Liver Injury Based on the Prediction of Drug-Target Interactions

Drug‐induced liver injury (DILI) is the leading cause of acute liver failure as well as one of the major reasons for drug withdrawal from clinical trials and the market. Elucidation of molecular interactions associated with DILI may help to detect potentially hazardous pharmacological agents at the early stages of drug development. The purpose of our study is to investigate which interactions with specific human protein targets may cause DILI. Prediction of interactions with 1534 human proteins was performed for the dataset with information about 699 drugs, which were divided into three categories of DILI: severe (178 drugs), moderate (310 drugs) and without DILI (211 drugs). Based on the comparison of drug‐target interactions predicted for different drugs’ categories and interpretation of those results using clustering, Gene Ontology, pathway and gene expression analysis, we identified 61 protein targets associated with DILI. Most of the revealed proteins were linked with hepatocytes’ death caused by disruption of vital cellular processes, as well as the emergence of inflammation in the liver. It was found that interaction of a drug with the identified targets is the essential molecular mechanism of the severe DILI for the most of the considered pharmaceuticals. Thus, pharmaceutical agents interacting with many of the identified targets may be considered as candidates for filtering out at the early stages of drug research.

Recognition of protein function using the local similarity.

The functional annotation of amino acid sequences is one of the most important problems in bioinformatics. Different programs have been successfully applied for recognition of some functional classes; nevertheless, many functional groups still cannot be predicted with the required accuracy. We developed a new method for protein function recognition using the original approach of sequence description. Each sequence of the training set is compared with the query sequence, and the local similarity scores are calculated for the query sequence positions and used as input data for the original classifier. The method was tested using leave-one-out cross-validation for three data sets covering 58 enzyme classes. Two tested sets including noncrossing functional classes were recognized with high accuracy at various levels of classification hierarchy. The majority of these classes were predicted with 100% accuracy, showing a prediction ability comparable with the HMMer method and an accuracy superior to the SVM-Prot program. When the tested set was composed of intersected classes of ligand specificity, the prediction accuracy was less; however, the accuracy increased as the size of the predicted class expanded. The proposed method can be used for both predicting protein functional class and selecting the functionally significant sites in a sequence.