Use of Ligand Based Models for Protein Domains To Predict Novel Molecular Targets and Applications To Triage Affinity Chromatography Data

MOST: most-similar ligand based approach to target prediction

BACKGROUND

Many computational approaches have been used for target prediction, including machine learning, reverse docking, bioactivity spectra analysis, and chemical similarity searching. Recent studies have suggested that chemical similarity searching may be driven by the most-similar ligand. However, the extent of bioactivity of most-similar ligands has been oversimplified or even neglected in these studies, and this has impaired the prediction power.

RESULTS

Here we propose the MOst-Similar ligand-based Target inference approach, namely MOST, which uses fingerprint similarity and explicit bioactivity of the most-similar ligands to predict targets of the query compound. Performance of MOST was evaluated by using combinations of different fingerprint schemes, machine learning methods, and bioactivity representations. In sevenfold cross-validation with a benchmark Ki dataset from CHEMBL release 19 containing 61,937 bioactivity data of 173 human targets, MOST achieved high average prediction accuracy (0.95 for pKi ≥ 5, and 0.87 for pKi ≥ 6). Morgan fingerprint was shown to be slightly better than FP2. Logistic Regression and Random Forest methods performed better than Naïve Bayes. In a temporal validation, the Ki dataset from CHEMBL19 were used to train models and predict the bioactivity of newly deposited ligands in CHEMBL20. MOST also performed well with high accuracy (0.90 for pKi ≥ 5, and 0.76 for pKi ≥ 6), when Logistic Regression and Morgan fingerprint were employed. Furthermore, the p values associated with explicit bioactivity were found be a robust index for removing false positive predictions. Implicit bioactivity did not offer this capability. Finally, p values generated with Logistic Regression, Morgan fingerprint and explicit activity were integrated with a false discovery rate (FDR) control procedure to reduce false positives in multiple-target prediction scenario, and the success of this strategy it was demonstrated with a case of fluanisone. In the case of aloe-emodin's laxative effect, MOST predicted that acetylcholinesterase was the mechanism-of-action target; in vivo studies validated this prediction.

CONCLUSIONS

Using the MOST approach can result in highly accurate and robust target prediction. Integrated with a FDR control procedure, MOST provides a reliable framework for multiple-target inference. It has prospective applications in drug repurposing and mechanism-of-action target prediction.

Open Source Bayesian Models. 3. Composite Models for Prediction of Binned Responses

Bayesian models constructed from structure-derived fingerprints have been a popular and useful method for drug discovery research when applied to bioactivity measurements that can be effectively classified as active or inactive. The results can be used to rank candidate structures according to their probability of activity, and this ranking benefits from the high degree of interpretability when structure-based fingerprints are used, making the results chemically intuitive. Besides selecting an activity threshold, building a Bayesian model is fast and requires few or no parameters or user intervention. The method also does not suffer from such acute overtraining problems as quantitative structure–activity relationships or quantitative structure–property relationships (QSAR/QSPR). This makes it an approach highly suitable for automated workflows that are independent of user expertise or prior knowledge of the training data. We now describe a new method for creating a composite group of Bayesian models to extend the method to work with multiple states, rather than just binary. Incoming activities are divided into bins, each covering a mutually exclusive range of activities. For each of these bins, a Bayesian model is created to model whether or not the compound belongs in the bin. Analyzing putative molecules using the composite model involves making a prediction for each bin and examining the relative likelihood for each assignment, for example, highest value wins. The method has been evaluated on a collection of hundreds of data sets extracted from ChEMBL v20 and validated data sets for ADME/Tox and bioactivity.

Nano-MoO3-mediated synthesis of bioactive thiazolidin-4-ones acting as anti-bacterial agents and their mode-of-action analysis using in silico target prediction, docking and similarity searching

Thiazolidin-4-ones inhibit bacterial growth by potentially targeting the FtsK motor domain of DNA translocase ofSalmonella typhi.

Target prediction utilising negative bioactivity data covering large chemical space

BACKGROUND

In silico analyses are increasingly being used to support mode-of-action investigations; however many such approaches do not utilise the large amounts of inactive data held in chemogenomic repositories. The objective of this work is concerned with the integration of such bioactivity data in the target prediction of orphan compounds to produce the probability of activity and inactivity for a range of targets. To this end, a novel human bioactivity data set was constructed through the assimilation of over 195 million bioactivity data points deposited in the ChEMBL and PubChem repositories, and the subsequent application of a sphere-exclusion selection algorithm to oversample presumed inactive compounds.

RESULTS

A Bernoulli Naïve Bayes algorithm was trained using the data and evaluated using fivefold cross-validation, achieving a mean recall and precision of 67.7 and 63.8 % for active compounds and 99.6 and 99.7 % for inactive compounds, respectively. We show the performances of the models are considerably influenced by the underlying intraclass training similarity, the size of a given class of compounds, and the degree of additional oversampling. The method was also validated using compounds extracted from WOMBAT producing average precision-recall AUC and BEDROC scores of 0.56 and 0.85, respectively. Inactive data points used for this test are based on presumed inactivity, producing an approximated indication of the true extrapolative ability of the models. A distance-based applicability domain analysis was also conducted; indicating an average Tanimoto Coefficient distance of 0.3 or greater between a test and training set can be used to give a global measure of confidence in model predictions. A final comparison to a method trained solely on active data from ChEMBL performed with precision-recall AUC and BEDROC scores of 0.45 and 0.76.

CONCLUSIONS

The inclusion of inactive data for model training produces models with superior AUC and improved early recognition capabilities, although the results from internal and external validation of the models show differing performance between the breadth of models. The realised target prediction protocol is available at https://github.com/lhm30/PIDGIN.Graphical abstractThe inclusion of large scale negative training data for in silico target prediction improves the precision and recall AUC and BEDROC scores for target models.

Identifying compound efficacy targets in phenotypic drug discovery

The identification of the efficacy target(s) for hits from phenotypic compound screens remains a key step to progress compounds into drug development. In addition to efficacy targets, the characterization of epistatic proteins influencing compound activity often facilitates the elucidation of the underlying mechanism of action; and, further, early determination of off-targets that cause potentially unwanted secondary phenotypes helps in assessing potential liabilities. This short review discusses the most important technologies currently available for characterizing the direct and indirect target space of bioactive compounds following phenotypic screening. We present a comprehensive strategy employing complementary approaches to balance individual technology strengths and weaknesses.

Proteochemometric modelling coupled to in silico target prediction: an integrated approach for the simultaneous prediction of polypharmacology and binding affinity/potency of small molecules

The rampant increase of public bioactivity databases has fostered the development of computational chemogenomics methodologies to evaluate potential ligand-target interactions (polypharmacology) both in a qualitative and quantitative way. Bayesian target prediction algorithms predict the probability of an interaction between a compound and a panel of targets, thus assessing compound polypharmacology qualitatively, whereas structure-activity relationship techniques are able to provide quantitative bioactivity predictions. We propose an integrated drug discovery pipeline combining in silico target prediction and proteochemometric modelling (PCM) for the respective prediction of compound polypharmacology and potency/affinity. The proposed pipeline was evaluated on the retrospective discovery of Plasmodium falciparum DHFR inhibitors. The qualitative in silico target prediction model comprised 553,084 ligand-target associations (a total of 262,174 compounds), covering 3,481 protein targets and used protein domain annotations to extrapolate predictions across species. The prediction of bioactivities for plasmodial DHFR led to a recall value of 79% and a precision of 100%, where the latter high value arises from the structural similarity of plasmodial DHFR inhibitors and T. gondii DHFR inhibitors in the training set. Quantitative PCM models were then trained on a dataset comprising 20 eukaryotic, protozoan and bacterial DHFR sequences, and 1,505 distinct compounds (in total 3,099 data points). The most predictive PCM model exhibited R²_0test and RMSE_test values of 0.79 and 0.59 pIC₅₀ units respectively, which was shown to outperform models based exclusively on compound (R²_0test/RMSE_test = 0.63/0.78) and target information (R²_0test/RMSE_test = 0.09/1.22), as well as inductive transfer knowledge between targets, with respective R²_0test and RMSE_test values of 0.76 and 0.63 pIC₅₀ units. Finally, both methods were integrated to predict the protein targets and the potency on plasmodial DHFR for the GSK TCAMS dataset, which comprises 13,533 compounds displaying strong anti-malarial activity. 534 of those compounds were identified as DHFR inhibitors by the target prediction algorithm, while the PCM algorithm identified 25 compounds, and 23 compounds (predicted pIC₅₀ > 7) were identified by both methods. Overall, this integrated approach simultaneously provides target and potency/affinity predictions for small molecules.

Graphical abstract

In silico target predictions: defining a benchmarking data set and comparison of performance of the multiclass Naïve Bayes and Parzen-Rosenblatt window

In this study, two probabilistic machine-learning algorithms were compared for in silico target prediction of bioactive molecules, namely the well-established Laplacian-modified Naïve Bayes classifier (NB) and the more recently introduced (to Cheminformatics) Parzen-Rosenblatt Window. Both classifiers were trained in conjunction with circular fingerprints on a large data set of bioactive compounds extracted from ChEMBL, covering 894 human protein targets with more than 155,000 ligand-protein pairs. This data set is also provided as a benchmark data set for future target prediction methods due to its size as well as the number of bioactivity classes it contains. In addition to evaluating the methods, different performance measures were explored. This is not as straightforward as in binary classification settings, due to the number of classes, the possibility of multiple class memberships, and the need to translate model scores into "yes/no" predictions for assessing model performance. Both algorithms achieved a recall of correct targets that exceeds 80% in the top 1% of predictions. Performance depends significantly on the underlying diversity and size of a given class of bioactive compounds, with small classes and low structural similarity affecting both algorithms to different degrees. When tested on an external test set extracted from WOMBAT covering more than 500 targets by excluding all compounds with Tanimoto similarity above 0.8 to compounds from the ChEMBL data set, the current methodologies achieved a recall of 63.3% and 66.6% among the top 1% for Naïve Bayes and Parzen-Rosenblatt Window, respectively. While those numbers seem to indicate lower performance, they are also more realistic for settings where protein targets need to be established for novel chemical substances.

Mapping small molecule binding data to structural domains

BACKGROUND

Large-scale bioactivity/SAR Open Data has recently become available, and this has allowed new analyses and approaches to be developed to help address the productivity and translational gaps of current drug discovery. One of the current limitations of these data is the relative sparsity of reported interactions per protein target, and complexities in establishing clear relationships between bioactivity and targets using bioinformatics tools. We detail in this paper the indexing of targets by the structural domains that bind (or are likely to bind) the ligand within a full-length protein. Specifically, we present a simple heuristic to map small molecule binding to Pfam domains. This profiling can be applied to all proteins within a genome to give some indications of the potential pharmacological modulation and regulation of all proteins.

RESULTS

In this implementation of our heuristic, ligand binding to protein targets from the ChEMBL database was mapped to structural domains as defined by profiles contained within the Pfam-A database. Our mapping suggests that the majority of assay targets within the current version of the ChEMBL database bind ligands through a small number of highly prevalent domains, and conversely the majority of Pfam domains sampled by our data play no currently established role in ligand binding. Validation studies, carried out firstly against Uniprot entries with expert binding-site annotation and secondly against entries in the wwPDB repository of crystallographic protein structures, demonstrate that our simple heuristic maps ligand binding to the correct domain in about 90 percent of all assessed cases. Using the mappings obtained with our heuristic, we have assembled ligand sets associated with each Pfam domain.

CONCLUSIONS

Small molecule binding has been mapped to Pfam-A domains of protein targets in the ChEMBL bioactivity database. The result of this mapping is an enriched annotation of small molecule bioactivity data and a grouping of activity classes following the Pfam-A specifications of protein domains. This is valuable for data-focused approaches in drug discovery, for example when extrapolating potential targets of a small molecule with known activity against one or few targets, or in the assessment of a potential target for drug discovery or screening studies.

Characterizing protein domain associations by Small-molecule ligand binding

BACKGROUND

Protein domains are evolutionarily conserved building blocks for protein structure and function, which are conventionally identified based on protein sequence or structure similarity. Small molecule binding domains are of great importance for the recognition of small molecules in biological systems and drug development. Many small molecules, including drugs, have been increasingly identified to bind to multiple targets, leading to promiscuous interactions with protein domains. Thus, a large scale characterization of the protein domains and their associations with respect to small-molecule binding is of particular interest to system biology research, drug target identification, as well as drug repurposing.

METHODS

We compiled a collection of 13,822 physical interactions of small molecules and protein domains derived from the Protein Data Bank (PDB) structures. Based on the chemical similarity of these small molecules, we characterized pairwise associations of the protein domains and further investigated their global associations from a network point of view.

RESULTS

We found that protein domains, despite lack of similarity in sequence and structure, were comprehensively associated through binding the same or similar small-molecule ligands. Moreover, we identified modules in the domain network that consisted of closely related protein domains by sharing similar biochemical mechanisms, being involved in relevant biological pathways, or being regulated by the same cognate cofactors.

CONCLUSIONS

A novel protein domain relationship was identified in the context of small-molecule binding, which is complementary to those identified by traditional sequence-based or structure-based approaches. The protein domain network constructed in the present study provides a novel perspective for chemogenomic study and network pharmacology, as well as target identification for drug repurposing.

Large-Scale QSAR in Target Prediction and Phenotypic HTS Assessment

The advent of in silico compound target prediction offers a potential paradigm shift in how large compound collections are understood and used strategically in high-throughput screens (HTS). Specifically, phenotypic HTS hits may be annotated both with known targets and predicted targets using large-scale QSAR models, enabling a more sophisticated hit assessment. Efforts in massive bioactivity data integration and standardization is empowering such compound-target annotations. These approaches differ fundamentally from the traditional role of QSAR in lead optimization and binding affinity predictions to global, probabilistic target predictions for thousands of human proteins.

From in silico target prediction to multi-target drug design: current databases, methods and applications

Given the tremendous growth of bioactivity databases, the use of computational tools to predict protein targets of small molecules has been gaining importance in recent years. Applications span a wide range, from the 'designed polypharmacology' of compounds to mode-of-action analysis. In this review, we firstly survey databases that can be used for ligand-based target prediction and which have grown tremendously in size in the past. We furthermore outline methods for target prediction that exist, both based on the knowledge of bioactivities from the ligand side and methods that can be applied in situations when a protein structure is known. Applications of successful in silico target identification attempts are discussed in detail, which were based partly or in whole on computational target predictions in the first instance. This includes the authors' own experience using target prediction tools, in this case considering phenotypic antibacterial screens and the analysis of high-throughput screening data. Finally, we will conclude with the prospective application of databases to not only predict, retrospectively, the protein targets of a small molecule, but also how to design ligands with desired polypharmacology in a prospective manner.

Multi-parameter phenotypic profiling: using cellular effects to characterize small-molecule compounds

Multi-parameter phenotypic profiling of small molecules provides important insights into their mechanisms of action, as well as a systems level understanding of biological pathways and their responses to small molecule treatments. It therefore deserves more attention at an early step in the drug discovery pipeline. Here, we summarize the technologies that are currently in use for phenotypic profiling--including mRNA-, protein- and imaging-based multi-parameter profiling--in the drug discovery context. We think that an earlier integration of phenotypic profiling technologies, combined with effective experimental and in silico target identification approaches, can improve success rates of lead selection and optimization in the drug discovery process.