Similarity methods in chemoinformatics

Artificial Intelligence and Machine Learning Models for Predicting Drug-Induced Kidney Injury in Small Molecules

BACKGROUND/OBJECTIVES

Drug-Induced Kidney Injury (DIKI) presents a significant challenge in drug development, often leading to clinical-stage failures. The early prediction of DIKI risk can improve drug safety and development efficiency. Existing models tend to focus on physicochemical properties alone, often overlooking drug-target interactions crucial for DIKI. This study introduces an AI/ML (artificial intelligence/machine learning) model that integrates both physicochemical properties and off-target interactions to enhance DIKI prediction.

METHODS

We compiled a dataset of 360 FDA-classified compounds (231 non-nephrotoxic and 129 nephrotoxic) and predicted 6064 off-target interactions, 59% of which were validated in vitro. We also calculated 55 physicochemical properties for these compounds. Machine learning (ML) models were developed using four algorithms: Ridge Logistic Regression (RLR), Support Vector Machine (SVM), Random Forest (RF), and Neural Network (NN). These models were then combined into an ensemble model for enhanced performance.

RESULTS

The ensemble model achieved an ROC-AUC of 0.86, with a sensitivity and specificity of 0.79 and 0.78, respectively. The key predictive features included 38 off-target interactions and physicochemical properties such as the number of metabolites, polar surface area (PSA), pKa, and fraction of Sp3-hybridized carbons (fsp3). These features effectively distinguished DIKI from non-DIKI compounds.

CONCLUSIONS

The integrated model, which combines both physicochemical properties and off-target interaction data, significantly improved DIKI prediction accuracy compared to models that rely on either data type alone. This AI/ML model provides a promising early screening tool for identifying compounds with lower DIKI risk, facilitating safer drug development.

Artificial Intelligence/Machine Learning-Driven Small Molecule Repurposing via Off-Target Prediction and Transcriptomics

The process of discovering small molecule drugs involves screening numerous compounds and optimizing the most promising ones, both in vitro and in vivo. However, approximately 90% of these optimized candidates fail during trials due to unexpected toxicity or insufficient efficacy. Current concepts with respect to drug-protein interactions suggest that each small molecule interacts with an average of 6-11 targets. This implies that approved drugs and even discontinued compounds could be repurposed by leveraging their interactions with unintended targets. Therefore, we developed a computational repurposing framework for small molecules, which combines artificial intelligence/machine learning (AI/ML)-based and chemical similarity-based target prediction methods with cross-species transcriptomics information. This repurposing methodology incorporates eight distinct target prediction methods, including three machine learning methods. By using multiple orthogonal methods for a "dataset" composed of 2766 FDA-approved drugs targeting multiple therapeutic target classes, we identified 27,371 off-target interactions involving 2013 protein targets (i.e., an average of around 10 interactions per drug). Relative to the drugs in the dataset, we identified 150,620 structurally similar compounds. The highest number of predicted interactions were for drugs targeting G protein-coupled receptors (GPCRs), enzymes, and kinases with 10,648, 4081, and 3678 interactions, respectively. Notably, 17,283 (63%) of the off-target interactions have been confirmed in vitro. Approximately 4000 interactions had an IC50 of <100 nM for 1105 FDA-approved drugs and 1661 interactions had an IC50 of <10 nM for 696 FDA-approved drugs. Together, the confirmation of numerous predicted interactions and the exploration of tissue-specific expression patterns in human and animal tissues offer insights into potential drug repurposing for new therapeutic applications.

Computational Strategy for Minimizing Mycotoxins in Cereal Crops: Assessment of the Biological Activity of Compounds Resulting from Virtual Screening

Cereal crops are frequently affected by toxigenic Fusarium species, among which the most common and worrying in Europe are Fusarium graminearum and Fusarium culmorum. These species are the causal agents of grain contamination with type B trichothecene (TCTB) mycotoxins. To help reduce the use of synthetic fungicides while guaranteeing low mycotoxin levels, there is an urgent need to develop new, efficient and environmentally-friendly plant protection solutions. Previously, F. graminearum proteins that could serve as putative targets to block the fungal spread and toxin production were identified and a virtual screening undertaken. Here, two selected compounds, M1 and M2, predicted, respectively, as the top compounds acting on the trichodiene synthase, a key enzyme of TCTB biosynthesis, and the 24-sterol-C-methyltransferase, a protein involved in ergosterol biosynthesis, were submitted for biological tests. Corroborating in silico predictions, M1 was shown to significantly inhibit TCTB yield by a panel of strains. Results were less obvious with M2 that induced only a slight reduction in fungal biomass. To go further, seven M1 analogs were assessed, which allowed evidencing of the physicochemical properties crucial for the anti-mycotoxin activity. Altogether, our results provide the first evidence of the promising potential of computational approaches to discover new anti-mycotoxin solutions

Extended continuous similarity indices: theory and application for QSAR descriptor selection

Extended (or n-ary) similarity indices have been recently proposed to extend the comparative analysis of binary strings. Going beyond the traditional notion of pairwise comparisons, these novel indices allow comparing any number of objects at the same time. This results in a remarkable efficiency gain with respect to other approaches, since now we can compare N molecules in O(N) instead of the common quadratic O(N²) timescale. This favorable scaling has motivated the application of these indices to diversity selection, clustering, phylogenetic analysis, chemical space visualization, and post-processing of molecular dynamics simulations. However, the current formulation of the n-ary indices is limited to vectors with binary or categorical inputs. Here, we present the further generalization of this formalism so it can be applied to numerical data, i.e. to vectors with continuous components. We discuss several ways to achieve this extension and present their analytical properties. As a practical example, we apply this formalism to the problem of feature selection in QSAR and prove that the extended continuous similarity indices provide a convenient way to discern between several sets of descriptors.

Improved Deep Learning Based Method for Molecular Similarity Searching Using Stack of Deep Belief Networks

Virtual screening (VS) is a computational practice applied in drug discovery research. VS is popularly applied in a computer-based search for new lead molecules based on molecular similarity searching. In chemical databases similarity searching is used to identify molecules that have similarities to a user-defined reference structure and is evaluated by quantitative measures of intermolecular structural similarity. Among existing approaches, 2D fingerprints are widely used. The similarity of a reference structure and a database structure is measured by the computation of association coefficients. In most classical similarity approaches, it is assumed that the molecular features in both biological and non-biologically-related activity carry the same weight. However, based on the chemical structure, it has been found that some distinguishable features are more important than others. Hence, this difference should be taken consideration by placing more weight on each important fragment. The main aim of this research is to enhance the performance of similarity searching by using multiple descriptors. In this paper, a deep learning method known as deep belief networks (DBN) has been used to reweight the molecule features. Several descriptors have been used for the MDL Drug Data Report (MDDR) dataset each of which represents different important features. The proposed method has been implemented with each descriptor individually to select the important features based on a new weight, with a lower error rate, and merging together all new features from all descriptors to produce a new descriptor for similarity searching. Based on the extensive experiments conducted, the results show that the proposed method outperformed several existing benchmark similarity methods, including Bayesian inference networks (BIN), the Tanimoto similarity method (TAN), adapted similarity measure of text processing (ASMTP) and the quantum-based similarity method (SQB). The results of this proposed multi-descriptor-based on Stack of deep belief networks method (SDBN) demonstrated a higher accuracy compared to existing methods on structurally heterogeneous datasets.

A Targeted Computational Screen of the SWEETLEAD Database Reveals FDA-Approved Compounds with Anti-Dengue Viral Activity

Affordable and effective antiviral therapies are needed worldwide, especially against agents such as dengue virus that are endemic in underserved regions. Many antiviral compounds have been studied in cultured cells but are unsuitable for clinical applications due to pharmacokinetic profiles, side effects, or inconsistent efficacy across dengue serotypes. Such tool compounds can, however, aid in identifying clinically useful treatments. Here, computational screening (Rapid Overlay of Chemical Structures) was used to identify entries in an in silico database of safe-in-human compounds (SWEETLEAD) that display high chemical similarities to known inhibitors of dengue virus. Inhibitors of the dengue proteinase NS2B/3, the dengue capsid, and the host autophagy pathway were used as query compounds. Three FDA-approved compounds that resemble the tool molecules structurally, cause little toxicity, and display strong antiviral activity in cultured cells were selected for further analysis. Pyrimethamine (50% inhibitory concentration [IC₅₀] = 1.2 μM), like the dengue proteinase inhibitor ARDP0006 to which it shows structural similarity, inhibited intramolecular NS2B/3 cleavage. Lack of toxicity early in infection allowed testing in mice, in which pyrimethamine also reduced viral loads. Niclosamide (IC₅₀ = 0.28 μM), like dengue core inhibitor ST-148, affected structural components of the virion and inhibited early processes during infection. Vandetanib (IC₅₀ = 1.6 μM), like cellular autophagy inhibitor spautin-1, blocked viral exit from cells and could be shown to extend survival in vivo Thus, three FDA-approved compounds with promising utility for repurposing to treat dengue virus infections and their potential mechanisms were identified using computational tools and minimal phenotypic screening.IMPORTANCE No antiviral therapeutics are currently available for dengue virus infections. By computationally overlaying the three-dimensional (3D) chemical structures of compounds known to inhibit dengue virus over those of compounds known to be safe in humans, we identified three FDA-approved compounds that are attractive candidates for repurposing as antivirals. We identified targets for two previously identified antiviral compounds and revealed a previously unknown potential anti-dengue compound, vandetanib. This computational approach to analyze a highly curated library of structures has the benefits of speed and cost efficiency. It also leverages mechanistic work with query compounds used in biomedical research to provide strong hypotheses for the antiviral mechanisms of the safer hit compounds. This workflow to identify compounds with known safety profiles can be expanded to any biological activity for which a small-molecule query compound has been identified, potentially expediting the translation of basic research to clinical interventions.

Do We Build Similar Molecules for Comorbid Diseases? Tevarud in Drug Design, an Analysis for Depression and Inflammation

Tevarud designates two poets coincidently writing a same verse in the Ottoman Divan literature. This study aims to analyze the structural similarity of molecules independently designed for inflammation and depression to determine if coincidentally we are building similar molecules for comorbid diseases. For this purpose, a molecule library was first constituted with structures that were developed as anti-inflammatory (AI) and antidepressant (AD) agents these last decades. Then, the similarity of the structures was determined by calculating the Tanimoto and Cosine similarity coefficients for each AD/AI pair. The highest scores were obtained for two theophylline derivatives: AD17 (for which some AI activity was found to be mentioned) and AI42. The study also pointed out the similarity of some AD coumarins with some AI flavonoids interestingly found to be highly similar to some AI coumarins and AD flavonoids, respectively. Thus, our investigation demonstrated that structures independently developed as AD and AI derivatives can present extremely high structural similarity, a finding that can suggest mechanistic interconnection for these comorbid diseases and also guide for the design of novel bioactive compounds.

ccbmlib - a Python package for modeling Tanimoto similarity value distributions

The ccbmlib Python package is a collection of modules for modeling similarity value distributions based on Tanimoto coefficients for fingerprints available in RDKit. It can be used to assess the statistical significance of Tanimoto coefficients and evaluate how molecular similarity is reflected when different fingerprint representations are used. Significance measures derived from p-values allow a quantitative comparison of similarity scores obtained from different fingerprint representations that might have very different value ranges. Furthermore, the package models conditional distributions of similarity coefficients for a given reference compound. The conditional significance score estimates where a test compound would be ranked in a similarity search. The models are based on the statistical analysis of feature distributions and feature correlations of fingerprints of a reference database. The resulting models have been evaluated for 11 RDKit fingerprints, taking a collection of ChEMBL compounds as a reference data set. For most fingerprints, highly accurate models were obtained, with differences of 1% or less for Tanimoto coefficients indicating high similarity.

ccbmlib - a Python package for modeling Tanimoto similarity value distributions

The ccbmlib Python package is a collection of modules for modeling similarity value distributions based on Tanimoto coefficients for fingerprints available in RDKit. It can be used to assess the statistical significance of Tanimoto coefficients and evaluate how molecular similarity is reflected when different fingerprint representations are used. Significance measures derived from p-values allow a quantitative comparison of similarity scores obtained from different fingerprint representations that might have very different value ranges. Furthermore, the package models conditional distributions of similarity coefficients for a given reference compound. The conditional significance score estimates where a test compound would be ranked in a similarity search. The models are based on the statistical analysis of feature distributions and feature correlations of fingerprints of a reference database. The resulting models have been evaluated for 11 RDKit fingerprints, taking a collection of ChEMBL compounds as a reference data set. For most fingerprints, highly accurate models were obtained, with differences of 1% or less for Tanimoto coefficients indicating high similarity.

The rcdk and cluster R packages applied to drug candidate selection

The aim of this article is to show how thevpower of statistics and cheminformatics can be combined, in R, using two packages: rcdk and cluster.We describe the role of clustering methods for identifying similar structures in a group of 23 molecules according to their fingerprints. The most commonly used method is to group the molecules using a "score" obtained by measuring the average distance between them. This score reflects the similarity/non-similarity between compounds and helps us identify active or potentially toxic substances through predictive studies.Clustering is the process by which the common characteristics of a particular class of compounds are identified. For clustering applications, we are generally measure the molecular fingerprint similarity with the Tanimoto coefficient. Based on the molecular fingerprints, we calculated the molecular distances between the methotrexate molecule and the other 23 molecules in the group, and organized them into a matrix. According to the molecular distances and Ward 's method, the molecules were grouped into 3 clusters. We can presume structural similarity between the compounds and their locations in the cluster map. Because only 5 molecules were included in the methotrexate cluster, we considered that they might have similar properties and might be further tested as potential drug candidates.

Novel Computational Approach to Predict Off-Target Interactions for Small Molecules

Most small molecule drugs interact with unintended, often unknown, biological targets and these off-target interactions may lead to both preclinical and clinical toxic events. Undesired off-target interactions are often not detected using current drug discovery assays, such as experimental polypharmacological screens. Thus, improvement in the early identification of off-target interactions represents an opportunity to reduce safety-related attrition rates during preclinical and clinical development. In order to better identify potential off-target interactions that could be linked to predictable safety issues, a novel computational approach to predict safety-relevant interactions currently not covered was designed and evaluated. These analyses, termed Off-Target Safety Assessment (OTSA), cover more than 7,000 targets (~35% of the proteome) and > 2,46,704 preclinical and clinical alerts (as of January 20, 2019). The approach described herein exploits a highly curated training set of >1 million compounds (tracking >20 million compound-structure activity relationship/SAR data points) with known in vitro activities derived from patents, journals, and publicly available databases. This computational process was used to predict both the primary and secondary pharmacological activities for a selection of 857 diverse small molecule drugs for which extensive secondary pharmacology data are readily available (456 discontinued and 401 FDA approved). The OTSA process predicted a total of 7,990 interactions for these 857 molecules. Of these, 3,923 and 4,067 possible high-scoring interactions were predicted for the discontinued and approved drugs, respectively, translating to an average of 9.3 interactions per drug. The OTSA process correctly identified the known pharmacological targets for >70% of these drugs, but also predicted a significant number of off-targets that may provide additional insight into observed in vivo effects. About 51.5% (2,025) and 22% (900) of these predicted high-scoring interactions have not previously been reported for the discontinued and approved drugs, respectively, and these may have a potential for repurposing efforts. Moreover, for both drug categories, higher promiscuity was observed for compounds with a MW range of 300 to 500, TPSA of ~200, and clogP ≥7. This computation also revealed significantly lower promiscuity (i.e., number of confirmed off-targets) for compounds with MW > 700 and MW<200 for both categories. In addition, 15 internal small molecules with known off-target interactions were evaluated. For these compounds, the OTSA framework not only captured about 56.8% of in vitro confirmed off-target interactions, but also identified the right pharmacological targets for 14 compounds as one of the top scoring targets. In conclusion, the OTSA process demonstrates good predictive performance characteristics and represents an additional tool with utility during the lead optimization stage of the drug discovery process. Additionally, the computed physiochemical properties such as clogP (i.e., lipophilicity), molecular weight, pKa and logS (i.e., solubility) were found to be statistically different between the approved and discontinued drugs, but the internal compounds were close to the approved drugs space in most part.

Discovering Highly Potent Molecules from an Initial Set of Inactives Using Iterative Screening

The versatility of similarity searching and quantitative structure-activity relationships to model the activity of compound sets within given bioactivity ranges (i.e., interpolation) is well established. However, their relative performance in the common scenario in early stage drug discovery where lots of inactive data but no active data points are available (i.e., extrapolation from the low-activity to the high-activity range) has not been thoroughly examined yet. To this aim, we have designed an iterative virtual screening strategy which was evaluated on 25 diverse bioactivity data sets from ChEMBL. We benchmark the efficiency of random forest (RF), multiple linear regression, ridge regression, similarity searching, and random selection of compounds to identify a highly active molecule in the test set among a large number of low-potency compounds. We use the number of iterations required to find this active molecule to evaluate the performance of each experimental setup. We show that linear and ridge regression often outperform RF and similarity searching, reducing the number of iterations to find an active compound by a factor of 2 or more. Even simple regression methods seem better able to extrapolate to high-bioactivity ranges than RF, which only provides output values in the range covered by the training set. In addition, examination of the scaffold diversity in the data sets used shows that in some cases similarity searching and RF require two times as many iterations as random selection depending on the chemical space covered in the initial training data. Lastly, we show using bioactivity data for COX-1 and COX-2 that our framework can be extended to multitarget drug discovery, where compounds are selected by concomitantly considering their activity against multiple targets. Overall, this study provides an approach for iterative screening where only inactive data are present in early stages of drug discovery in order to discover highly potent compounds and the best experimental set up in which to do so.

Discovery of Potential Inhibitors of Squalene Synthase from Traditional Chinese Medicine Based on Virtual Screening and In Vitro Evaluation of Lipid-Lowering Effect

Squalene synthase (SQS), a key downstream enzyme involved in the cholesterol biosynthetic pathway, plays an important role in treating hyperlipidemia. Compared to statins, SQS inhibitors have shown a very significant lipid-lowering effect and do not cause myotoxicity. Thus, the paper aims to discover potential SQS inhibitors from Traditional Chinese Medicine (TCM) by the combination of molecular modeling methods and biological assays. In this study, cynarin was selected as a potential SQS inhibitor candidate compound based on its pharmacophoric properties, molecular docking studies and molecular dynamics (MD) simulations. Cynarin could form hydrophobic interactions with PHE54, LEU211, LEU183 and PRO292, which are regarded as important interactions for the SQS inhibitors. In addition, the lipid-lowering effect of cynarin was tested in sodium oleate-induced HepG2 cells by decreasing the lipidemic parameter triglyceride (TG) level by 22.50%. Finally. cynarin was reversely screened against other anti-hyperlipidemia targets which existed in HepG2 cells and cynarin was unable to map with the pharmacophore of these targets, which indicated that the lipid-lowering effects of cynarin might be due to the inhibition of SQS. This study discovered cynarin is a potential SQS inhibitor from TCM, which could be further clinically explored for the treatment of hyperlipidemia.

Cross-Classified Multilevel Modelling of the Effectiveness of Similarity-Based Virtual Screening

The screening effectiveness of a chemical similarity search depends on a range of factors, including the bioactivity of interest, the types of similarity coefficient and fingerprint that comprise the similarity measure, and the nature of the reference structure that is being searched against a database. This study introduces the use of cross-classified multilevel modelling as a way to investigate the relative importance of these four factors when carrying out similarity searches on the ChEMBL database. Two principal conclusions can be drawn from the analyses: that the fingerprint plays a more important role than the similarity coefficient in determining the effectiveness of a similarity search, and that comparative studies of similarity measures should involve many more reference structures than has been the case in much previous work.

Chemometric-assisted method development in hydrophilic interaction liquid chromatography: A review

With an enormous growth in the application of hydrophilic interaction liquid chromatography (HILIC), there has also been significant progress in HILIC method development. HILIC is a chromatographic method that utilises hydro-organic mobile phases with a high organic content, and a hydrophilic stationary phase. It has been applied predominantly in the determination of small polar compounds. Theoretical studies in computer-aided modelling tools, most importantly the predictive, quantitative structure retention relationship (QSRR) modelling methods, have attracted the attention of researchers and these approaches greatly assist the method development process. This review focuses on the application of computer-aided modelling tools in understanding the retention mechanism, the classification of HILIC stationary phases, prediction of retention times in HILIC systems, optimisation of chromatographic conditions, and description of the interaction effects of the chromatographic factors in HILIC separations. Additionally, what has been achieved in the potential application of QSRR methodology in combination with experimental design philosophy in the optimisation of chromatographic separation conditions in the HILIC method development process is communicated. Developing robust predictive QSRR models will undoubtedly facilitate more application of this chromatographic mode in a broader variety of research areas, significantly minimising cost and time of the experimental work.

Rapid Method Development in Hydrophilic Interaction Liquid Chromatography for Pharmaceutical Analysis Using a Combination of Quantitative Structure-Retention Relationships and Design of Experiments

A design-of-experiment (DoE) model was developed, able to describe the retention times of a mixture of pharmaceutical compounds in hydrophilic interaction liquid chromatography (HILIC) under all possible combinations of acetonitrile content, salt concentration, and mobile-phase pH with R² > 0.95. Further, a quantitative structure-retention relationship (QSRR) model was developed to predict retention times for new analytes, based only on their chemical structures, with a root-mean-square error of prediction (RMSEP) as low as 0.81%. A compound classification based on the concept of similarity was applied prior to QSRR modeling. Finally, we utilized a combined QSRR-DoE approach to propose an optimal design space in a quality-by-design (QbD) workflow to facilitate the HILIC method development. The mathematical QSRR-DoE model was shown to be highly predictive when applied to an independent test set of unseen compounds in unseen conditions with a RMSEP value of 5.83%. The QSRR-DoE computed retention time of pharmaceutical test analytes and subsequently calculated separation selectivity was used to optimize the chromatographic conditions for efficient separation of targets. A Monte Carlo simulation was performed to evaluate the risk of uncertainty in the model's prediction, and to define the design space where the desired quality criterion was met. Experimental realization of peak selectivity between targets under the selected optimal working conditions confirmed the theoretical predictions. These results demonstrate how discovery of optimal conditions for the separation of new analytes can be accelerated by the use of appropriate theoretical tools.

ROCS-derived features for virtual screening

Rapid overlay of chemical structures (ROCS) is a standard tool for the calculation of 3D shape and chemical ("color") similarity. ROCS uses unweighted sums to combine many aspects of similarity, yielding parameter-free models for virtual screening. In this report, we decompose the ROCS color force field into color components and color atom overlaps, novel color similarity features that can be weighted in a system-specific manner by machine learning algorithms. In cross-validation experiments, these additional features significantly improve virtual screening performance relative to standard ROCS.

Chemoinformatics at the University of Sheffield 2002-2014

This paper summarises work in chemoinformatics carried out in the Information School of the University of Sheffield during the period 2002-2014. Research studies are described on fingerprint-based similarity searching, data fusion, applications of reduced graphs and pharmacophore mapping, and on the School's teaching in chemoinformatics.

USR-VS: a web server for large-scale prospective virtual screening using ultrafast shape recognition techniques

Ligand-based Virtual Screening (VS) methods aim at identifying molecules with a similar activity profile across phenotypic and macromolecular targets to that of a query molecule used as search template. VS using 3D similarity methods have the advantage of biasing this search toward active molecules with innovative chemical scaffolds, which are highly sought after in drug design to provide novel leads with improved properties over the query molecule (e.g. patentable, of lower toxicity or increased potency). Ultrafast Shape Recognition (USR) has demonstrated excellent performance in the discovery of molecules with previously-unknown phenotypic or target activity, with retrospective studies suggesting that its pharmacophoric extension (USRCAT) should obtain even better hit rates once it is used prospectively. Here we present USR-VS (http://usr.marseille.inserm.fr/), the first web server using these two validated ligand-based 3D methods for large-scale prospective VS. In about 2 s, 93.9 million 3D conformers, expanded from 23.1 million purchasable molecules, are screened and the 100 most similar molecules among them in terms of 3D shape and pharmacophoric properties are shown. USR-VS functionality also provides interactive visualization of the similarity of the query molecule against the hit molecules as well as vendor information to purchase selected hits in order to be experimentally tested.

Adapting Document Similarity Measures for Ligand-Based Virtual Screening

Quantifying the similarity of molecules is considered one of the major tasks in virtual screening. There are many similarity measures that have been proposed for this purpose, some of which have been derived from document and text retrieving areas as most often these similarity methods give good results in document retrieval and can achieve good results in virtual screening. In this work, we propose a similarity measure for ligand-based virtual screening, which has been derived from a text processing similarity measure. It has been adopted to be suitable for virtual screening; we called this proposed measure the Adapted Similarity Measure of Text Processing (ASMTP). For evaluating and testing the proposed ASMTP we conducted several experiments on two different benchmark datasets: the Maximum Unbiased Validation (MUV) and the MDL Drug Data Report (MDDR). The experiments have been conducted by choosing 10 reference structures from each class randomly as queries and evaluate them in the recall of cut-offs at 1% and 5%. The overall obtained results are compared with some similarity methods including the Tanimoto coefficient, which are considered to be the conventional and standard similarity coefficients for fingerprint-based similarity calculations. The achieved results show that the performance of ligand-based virtual screening is better and outperforms the Tanimoto coefficients and other methods.

Computational methods to identify metabolic sub-networks based on metabolomic profiles

Untargeted metabolomics makes it possible to identify compounds that undergo significant changes in concentration in different experimental conditions. The resulting metabolomic profile characterizes the perturbation concerned, but does not explain the underlying biochemical mechanisms. Bioinformatics methods make it possible to interpret results in light of the whole metabolism. This knowledge is modelled into a network, which can be mined using algorithms that originate in graph theory. These algorithms can extract sub-networks related to the compounds identified. Several attempts have been made to adapt them to obtain more biologically meaningful results. However, there is still no consensus on this kind of analysis of metabolic networks. This review presents the main graph approaches used to interpret metabolomic data using metabolic networks. Their advantages and drawbacks are discussed, and the impacts of their parameters are emphasized. We also provide some guidelines for relevant sub-network extraction and also suggest a range of applications for most methods.

In silico methods to address polypharmacology: current status, applications and future perspectives

Polypharmacology, a new paradigm in drug discovery that focuses on multi-target drugs (MTDs), has potential application for drug repurposing, the process of finding new uses for existing approved drugs, prediction of off-target toxicities and rational design of MTDs. In this scenario, computational strategies have demonstrated great potential in predicting polypharmacology and in facilitating drug repurposing. Here, we provide a comprehensive overview of various computational approaches that enable the prediction and analysis of in vitro and in vivo drug-response phenotypes and outline their potential for drug discovery. We give an outlook on the latest advances in rational design of MTDs and discuss possible future directions of algorithm development in this field.

Modelling ligand selectivity of serine proteases using integrative proteochemometric approaches improves model performance and allows the multi-target dependent interpretation of features

Serine proteases, implicated in important physiological functions, have a high intra-family similarity, which leads to unwanted off-target effects of inhibitors with insufficient selectivity. However, the availability of sequence and structure data has now made it possible to develop approaches to design pharmacological agents that can discriminate successfully between their related binding sites. In this study, we have quantified the relationship between 12,625 distinct protease inhibitors and their bioactivity against 67 targets of the serine protease family (20,213 data points) in an integrative manner, using proteochemometric modelling (PCM). The benchmarking of 21 different target descriptors motivated the usage of specific binding pocket amino acid descriptors, which helped in the identification of active site residues and selective compound chemotypes affecting compound affinity and selectivity. PCM models performed better than alternative approaches (models trained using exclusively compound descriptors on all available data, QSAR) employed for comparison with R(2)/RMSE values of 0.64 ± 0.23/0.66 ± 0.20 vs. 0.35 ± 0.27/1.05 ± 0.27 log units, respectively. Moreover, the interpretation of the PCM model singled out various chemical substructures responsible for bioactivity and selectivity towards particular proteases (thrombin, trypsin and coagulation factor 10) in agreement with the literature. For instance, absence of a tertiary sulphonamide was identified to be responsible for decreased selective activity (by on average 0.27 ± 0.65 pChEMBL units) on FA10. Among the binding pocket residues, the amino acids (arginine, leucine and tyrosine) at positions 35, 39, 60, 93, 140 and 207 were observed as key contributing residues for selective affinity on these three targets.

Efficient heuristics for maximum common substructure search

Maximum common substructure search is a computationally hard optimization problem with diverse applications in the field of cheminformatics, including similarity search, lead optimization, molecule alignment, and clustering. Most of these applications have strict constraints on running time, so heuristic methods are often preferred. However, the development of an algorithm that is both fast enough and accurate enough for most practical purposes is still a challenge. Moreover, in some applications, the quality of a common substructure depends not only on its size but also on various topological features of the one-to-one atom correspondence it defines. Two state-of-the-art heuristic algorithms for finding maximum common substructures have been implemented at ChemAxon Ltd., and effective heuristics have been developed to improve both their efficiency and the relevance of the atom mappings they provide. The implementations have been thoroughly evaluated and compared with existing solutions (KCOMBU and Indigo). The heuristics have been found to greatly improve the performance and applicability of the algorithms. The purpose of this paper is to introduce the applied methods and present the experimental results.

Polypharmacology modelling using proteochemometrics (PCM): recent methodological developments, applications to target families, and future prospects

Proteochemometric (PCM) modelling is a computational method to model the bioactivity of multiple ligands against multiple related protein targets simultaneously.

Condorcet and borda count fusion method for ligand-based virtual screening

Background

It is known that any individual similarity measure will not always give the best recall of active molecule structure for all types of activity classes. Recently, the effectiveness of ligand-based virtual screening approaches can be enhanced by using data fusion. Data fusion can be implemented using two different approaches: group fusion and similarity fusion. Similarity fusion involves searching using multiple similarity measures. The similarity scores, or ranking, for each similarity measure are combined to obtain the final ranking of the compounds in the database.

Results

The Condorcet fusion method was examined. This approach combines the outputs of similarity searches from eleven association and distance similarity coefficients, and then the winner measure for each class of molecules, based on Condorcet fusion, was chosen to be the best method of searching. The recall of retrieved active molecules at top 5% and significant test are used to evaluate our proposed method. The MDL drug data report (MDDR), maximum unbiased validation (MUV) and Directory of Useful Decoys (DUD) data sets were used for experiments and were represented by 2D fingerprints.

Conclusions

Simulated virtual screening experiments with the standard two data sets show that the use of Condorcet fusion provides a very simple way of improving the ligand-based virtual screening, especially when the active molecules being sought have a lowest degree of structural heterogeneity. However, the effectiveness of the Condorcet fusion was increased slightly when structural sets of high diversity activities were being sought.

Large-scale SAR analysis

The analysis of structure–activity relationships (SARs) is a central task in medicinal chemistry. Traditionally, SAR exploration has concentrated on individual compound series. This conventional approach is complemented by large-scale SAR analysis, which puts strong emphasis on data mining and SAR visualization. This contribution reviews recent concepts for large-scale SAR analysis including numerical functions to characterize global and local SAR information content of compound data sets, alternative activity landscape representations and data mining strategies.

Retrospective group fusion similarity search based on eROCE evaluation metric

In this study, a simple evaluation metric, denoted as eROCE was proposed to measure the early enrichment of predictive methods. We demonstrated the superior robustness of eROCE compared to other known metrics throughout several active to inactive ratios ranging from 1:10 to 1:1000. Group fusion similarity search was investigated by varying 16 similarity coefficients, five molecular representations (binary and non-binary) and two group fusion rules using two reference structure set sizes. We used a dataset of 3478 actives and 43,938 inactive molecules and the enrichment was analyzed by means of eROCE. This retrospective study provides optimal similarity search parameters in the case of ALDH1A1 inhibitors.

Comparison of chemical similarity measures using different numbers of query structures

Many different similarity measures have been described for searching chemical databases. Drawing on previous work in textual information retrieval, this paper investigates the numbers of queries that are required to make robust statements as to the relative retrieval effectiveness of different similarity measures. Experiments with the MDL Drug Data Report database suggest that much larger numbers of queries are ideally required for this purpose than has been the case in previous comparative studies in chemoinformatics.

PubChem3D: Biologically relevant 3-D similarity

Background

The use of 3-D similarity techniques in the analysis of biological data and virtual screening is pervasive, but what is a biologically meaningful 3-D similarity value? Can one find statistically significant separation between "active/active" and "active/inactive" spaces? These questions are explored using 734,486 biologically tested chemical structures, 1,389 biological assay data sets, and six different 3-D similarity types utilized by PubChem analysis tools.

Results

The similarity value distributions of 269.7 billion unique conformer pairs from 734,486 biologically tested compounds (all-against-all) from PubChem were utilized to help work towards an answer to the question: what is a biologically meaningful 3-D similarity score? The average and standard deviation for the six similarity measures ST^ST-opt, CT^ST-opt, ComboT^ST-opt, ST^CT-opt, CT^CT-opt, and ComboT^CT-opt were 0.54 ± 0.10, 0.07 ± 0.05, 0.62 ± 0.13, 0.41 ± 0.11, 0.18 ± 0.06, and 0.59 ± 0.14, respectively. Considering that this random distribution of biologically tested compounds was constructed using a single theoretical conformer per compound (the "default" conformer provided by PubChem), further study may be necessary using multiple diverse conformers per compound; however, given the breadth of the compound set, the single conformer per compound results may still apply to the case of multi-conformer per compound 3-D similarity value distributions. As such, this work is a critical step, covering a very wide corpus of chemical structures and biological assays, creating a statistical framework to build upon.

The second part of this study explored the question of whether it was possible to realize a statistically meaningful 3-D similarity value separation between reputed biological assay "inactives" and "actives". Using the terminology of noninactive-noninactive (NN) pairs and the noninactive-inactive (NI) pairs to represent comparison of the "active/active" and "active/inactive" spaces, respectively, each of the 1,389 biological assays was examined by their 3-D similarity score differences between the NN and NI pairs and analyzed across all assays and by assay category types. While a consistent trend of separation was observed, this result was not statistically unambiguous after considering the respective standard deviations. While not all "actives" in a biological assay are amenable to this type of analysis, e.g., due to different mechanisms of action or binding configurations, the ambiguous separation may also be due to employing a single conformer per compound in this study. With that said, there were a subset of biological assays where a clear separation between the NN and NI pairs found. In addition, use of combo Tanimoto (ComboT) alone, independent of superposition optimization type, appears to be the most efficient 3-D score type in identifying these cases.

Conclusion

This study provides a statistical guideline for analyzing biological assay data in terms of 3-D similarity and PubChem structure-activity analysis tools. When using a single conformer per compound, a relatively small number of assays appear to be able to separate "active/active" space from "active/inactive" space.

Toward mechanistic classification of enzyme functions

Classification of enzyme function should be quantitative, computationally accessible, and informed by sequences and structures to enable use of genomic information for functional inference and other applications. Large-scale studies have established that divergently evolved enzymes share conserved elements of structure and common mechanistic steps and that convergently evolved enzymes often converge to similar mechanisms too, suggesting that reaction mechanisms could be used to develop finer-grained functional descriptions than provided by the Enzyme Commission (EC) system currently in use. Here we describe how evolution informs these structure-function mappings and review the databases that store mechanisms of enzyme reactions along with recent developments to measure ligand and mechanistic similarities. Together, these provide a foundation for new classifications of enzyme function.

Similarity searching using 2D structural fingerprints

This chapter reviews the use of molecular fingerprints for chemical similarity searching. The fingerprints encode the presence of 2D substructural fragments in a molecule, and the similarity between a pair of molecules is a function of the number of fragments that they have in common. Although this provides a very simple way of estimating the degree of structural similarity between two molecules, it has been found to provide an effective and an efficient tool for searching large chemical databases. The review describes the historical development of similarity searching since it was first described in the mid-1980s, reviews the many different coefficients, representations, and weightings that can be combined to form a similarity measure, describes quantitative measures of the effectiveness of similarity searching, and concludes by looking at current developments based on the use of data fusion and machine learning techniques.