Exploring Activity Cliffs in Medicinal Chemistry

FOCUS on NOD2: Advancing IBD Drug Discovery with a User-Informed Machine Learning Framework

In this study, we introduce the Framework for Optimized Customizable User-Informed Synthesis (FOCUS), a generative machine learning model tailored for drug discovery. FOCUS integrates domain expertise and uses Proximal Policy Optimization (PPO) to guide Monte Carlo Tree Search (MCTS) to efficiently explore chemical space. It generates SMILES representations of potential drug candidates, optimizing for druggability and binding efficacy to NOD2, PEP, and MCT1 receptors. The model is highly interpretive, allowing for user-feedback and expert-driven adjustments based on detailed cycle reports. Employing tools like SHAP and LIME, FOCUS provides a transparent analysis of decision-making processes, emphasizing features such as docking scores and interaction fingerprints. Comparative studies with Muramyl Dipeptide (MDP) demonstrate improved interaction profiles. FOCUS merges advanced machine learning with expert insight, accelerating the drug discovery pipeline.

iSIM: instant similarity

The quantification of molecular similarity has been present since the beginning of cheminformatics. Although several similarity indices and molecular representations have been reported, all of them ultimately reduce to the calculation of molecular similarities of only two objects at a time. Hence, to obtain the average similarity of a set of molecules, all the pairwise comparisons need to be computed, which demands a quadratic scaling in the number of computational resources. Here we propose an exact alternative to this problem: iSIM (instant similarity). iSIM performs comparisons of multiple molecules at the same time and yields the same value as the average pairwise comparisons of molecules represented by binary fingerprints and real-value descriptors. In this work, we introduce the mathematical framework and several applications of iSIM in chemical sampling, visualization, diversity selection, and clustering.

Benchmarking compound activity prediction for real-world drug discovery applications

Identifying active compounds for target proteins is fundamental in early drug discovery. Recently, data-driven computational methods have demonstrated promising potential in predicting compound activities. However, there lacks a well-designed benchmark to comprehensively evaluate these methods from a practical perspective. To fill this gap, we propose a Compound Activity benchmark for Real-world Applications (CARA). Through carefully distinguishing assay types, designing train-test splitting schemes and selecting evaluation metrics, CARA can consider the biased distribution of current real-world compound activity data and avoid overestimation of model performances. We observed that although current models can make successful predictions for certain proportions of assays, their performances varied across different assays. In addition, evaluation of several few-shot training strategies demonstrated different performances related to task types. Overall, we provide a high-quality dataset for developing and evaluating compound activity prediction models, and the analyses in this work may inspire better applications of data-driven models in drug discovery.

Discovery and optimization of narrow spectrum inhibitors of Tousled like kinase 2 (TLK2) using quantitative structure activity relationships

The oxindole scaffold has been the center of several kinase drug discovery programs, some of which have led to approved medicines. A series of two oxindole matched pairs from the literature were identified where TLK2 was potently inhibited as an off-target kinase. The oxindole has long been considered a promiscuous kinase inhibitor template, but across these four specific literature oxindoles TLK2 activity was consistent, while the kinome profile was radically different ranging from narrow to broad spectrum kinome coverage. We synthesized a large series of analogues, utilizing quantitative structure-activity relationship (QSAR) analysis, water mapping of the kinase ATP binding sites, kinome profiling, and small-molecule x-ray structural analysis to optimize TLK2 inhibition and kinome selectivity. This resulted in the identification of several narrow spectrum, sub-family selective, chemical tool compounds including 128 (UNC-CA2-103) that could enable elucidation of TLK2 biology.

Molecular property prediction based on graph structure learning

MOTIVATION

Molecular property prediction (MPP) is a fundamental but challenging task in the computer-aided drug discovery process. More and more recent works employ different graph-based models for MPP, which have achieved considerable progress in improving prediction performance. However, current models often ignore relationships between molecules, which could be also helpful for MPP.

RESULTS

For this sake, in this article we propose a graph structure learning (GSL) based MPP approach, called GSL-MPP. Specifically, we first apply graph neural network (GNN) over molecular graphs to extract molecular representations. Then, with molecular fingerprints, we construct a molecule similarity graph (MSG). Following that, we conduct GSL on the MSG, i.e. molecule-level GSL, to get the final molecular embeddings, which are the results of fuzing both GNN encoded molecular representations and the relationships among molecules. That is, combining both intra-molecule and inter-molecule information. Finally, we use these molecular embeddings to perform MPP. Extensive experiments on 10 various benchmark datasets show that our method could achieve state-of-the-art performance in most cases, especially on classification tasks. Further visualization studies also demonstrate the good molecular representations of our method.

AVAILABILITY AND IMPLEMENTATION

Source code is available at https://github.com/zby961104/GSL-MPP.

Protein characteristics substantially influence the propensity of activity cliffs among kinase inhibitors

Activity cliffs (ACs) are pairs of structurally similar molecules with significantly different affinities for a biotarget, posing a challenge in computer-assisted drug discovery. This study focuses on protein kinases, significant therapeutic targets, with some exhibiting ACs while others do not despite numerous inhibitors. The hypothesis that the presence of ACs is dependent on the target protein and its complete structural context is explored. Machine learning models were developed to link protein properties to ACs, revealing specific tripeptide sequences and overall protein properties as critical factors in ACs occurrence. The study highlights the importance of considering the entire protein matrix rather than just the binding site in understanding ACs. This research provides valuable insights for drug discovery and design, paving the way for addressing ACs-related challenges in modern computational approaches.

A High-Quality Data Set of Protein-Ligand Binding Interactions Via Comparative Complex Structure Modeling

High-quality protein-ligand complex structures provide the basis for understanding the nature of noncovalent binding interactions at the atomic level and enable structure-based drug design. However, experimentally determined complex structures are scarce compared with the vast chemical space. In this study, we addressed this issue by constructing the BindingNet data set via comparative complex structure modeling, which contains 69,816 modeled high-quality protein-ligand complex structures with experimental binding affinity data. BindingNet provides valuable insights into investigating protein-ligand interactions, allowing visual inspection and interpretation of structural analogues' structure-activity relationships. It can also be used for evaluating machine-learning-based scoring functions. Our results indicate that machine learning models trained on BindingNet could reduce the bias caused by buried solvent-accessible surface area, as we previously found for models trained on the PDBbind data set. We also discussed strategies to improve BindingNet and its potential utilization for benchmarking the molecular docking methods and ligand binding free energy calculation approaches. The BindingNet complements PDBbind in constructing a sufficient and unbiased protein-ligand binding data set and is freely available at http://bindingnet.huanglab.org.cn.

Chemical similarity of molecules with physiological response

Measuring the similarity among molecules is an important task in various chemically oriented problems. This elusive concept is hard to define and quantify. Moreover, the complexity of the problem is elevated by bifurcating the notion of molecular similarity to structural and chemical similarity. While the structural similarity of molecules is being extensively researched, the so-called chemical similarity is being mentioned scarcely. Here, we propose a way of converting the physico-chemical properties into molecular fingerprints. Then, using the apparatus of measuring the structural similarity, the chemical similarity can be assessed. The proof of a concept is demonstrated on a set of molecules that induce diverse physiological responses.

Multitask learning-driven identification of novel antitrypanosomal compounds

Background: Chagas disease and human African trypanosomiasis cause substantial death and morbidity, particularly in low- and middle-income countries, making the need for novel drugs urgent. Methodology & results: Therefore, an explainable multitask pipeline to profile the activity of compounds against three trypanosomes (Trypanosoma brucei brucei, Trypanosoma brucei rhodesiense and Trypanosoma cruzi) were created. These models successfully discovered four new experimental hits (LC-3, LC-4, LC-6 and LC-15). Among them, LC-6 showed promising results, with IC50 values ranging 0.01-0.072 μM and selectivity indices >10,000. Conclusion: These results demonstrate that the multitask protocol offers predictivity and interpretability in the virtual screening of new antitrypanosomal compounds and has the potential to improve hit rates in Chagas and human African trypanosomiasis projects.

Machine Learning Model for Screening Thyroid Stimulating Hormone Receptor Agonists Based on Updated Datasets and Improved Applicability Domain Metrics

Machine learning (ML) models for screening endocrine-disrupting chemicals (EDCs), such as thyroid stimulating hormone receptor (TSHR) agonists, are essential for sound management of chemicals. Previous models for screening TSHR agonists were built on imbalanced datasets and lacked applicability domain (AD) characterization essential for regulatory application. Herein, an updated TSHR agonist dataset was built, for which the ratio of active to inactive compounds greatly increased to 1:2.6, and chemical spaces of structure-activity landscapes (SALs) were enhanced. Resulting models based on 7 molecular representations and 4 ML algorithms were proven to outperform previous ones. Weighted similarity density (ρ_s) and weighted inconsistency of activities (I_A) were proposed to characterize the SALs, and a state-of-the-art AD characterization methodology AD_SAL{ρ_s, I_A} was established. An optimal classifier developed with PubChem fingerprints and the random forest algorithm, coupled with AD_SAL{ρ_s ≥ 0.15, I_A ≤ 0.65}, exhibited good performance on the validation set with the area under the receiver operating characteristic curve being 0.984 and balanced accuracy being 0.941 and identified 90 TSHR agonist classes that could not be found previously. The classifier together with the AD_SAL{ρ_s, I_A} may serve as efficient tools for screening EDCs, and the AD characterization methodology may be applied to other ML models.

The Good, The Bad, and The Perplexing: Structural Alerts and Read-Across for Predicting Skin Sensitization Using Human Data

In our earlier work (Golden et al., 2021), we showed 70-80% accuracies for several skin sensitization computational tools using human data. Here, we expanded the data set using the NICEATM human skin sensitization database to create a final data set of 1355 discrete chemicals (largely negative, ∼70%). Using this expanded data set, we analyzed model performance and evaluated mispredictions using Toxtree (v 3.1.0), OECD QSAR Toolbox (v 4.5), VEGA's (1.2.0 BETA) CAESAR (v 2.1.7), and a k-nearest-neighbor (kNN) classification approach. We show that the accuracy on this data set was lower than previous estimates, with balanced accuracies being 63% and 65% for Toxtree and OECD QSAR Toolbox, respectively, 46% for VEGA, and 59% for a kNN approach, with the lower accuracy likely due to the higher percentage of nonsensitizing chemicals. Two hundred eighty seven chemicals were mispredicted by both Toxtree and OECD QSAR Toolbox, which was approximately 20% of the entire data set, and 84% of these were false positives. The absence or presence of metabolic simulation in OECD QSAR Toolbox made no overall difference. While Toxtree is known for overpredicting, 60% of the chemicals in the data set had no alert for skin sensitization, and a substantial number of these chemicals were in fact sensitizers, pointing to sensitization mechanisms not recognized by Toxtree. Interestingly, we observed that chemicals with more than one Toxtree alert were more likely to be nonsensitizers. Finally, a kNN approach tended to mispredict different chemicals than either OECD QSAR Toolbox or Toxtree, suggesting that there was additional information to be garnered from a kNN approach. Overall, the results demonstrate that while there is merit in structural alerts as well as QSAR or read-across approaches (perhaps even more so in their combination), additional improvement will require a more nuanced understanding of mechanisms of skin sensitization.

A consensual machine-learning-assisted QSAR model for effective bioactivity prediction of xanthine oxidase inhibitors using molecular fingerprints

Xanthine oxidase inhibitors (XOIs) have been widely studied due to the promising potential as safe and effective therapeutics in hyperuricemia and gout. Currently, available XOI molecules have been developed from different experiments but they are with the wide structure diversity and significant varying bioactivities. So it is of great practical significance to present a consensual QSAR model for effective bioactivity prediction of XOIs based on a systematic compiling of these XOIs across different experiments. In this work, 249 XOIs belonging to 16 scaffolds were collected and were integrated into a consensual dataset by introducing the concept of IC₅₀ values relative to allopurinol (RIC₅₀). Here, extended connectivity fingerprints (ECFPs) were employed to represent XOI molecules. By performing effective feature selection by machine-learning method, 54 crucial fingerprints were indicated to be valuable for predicting the inhibitory potency (IP) of XOIs. The optimal predictor yields the promising performance by different cross-validation tests. Besides, an external validation of 43 XOIs and a case study on febuxostat also provide satisfactory results, indicating the powerful generalization of our predictor. Here, the predictor was interpreted by shapely additive explanation (SHAP) method which revealed several important substructures by mapping the featured fingerprints to molecular structures. Then, 15 new molecules were designed and predicted by our predictor to show superior IP than febuxostat. Finally, molecular docking simulation was performed to gain a deep insight into molecular binding mode with xanthine oxidase (XO) enzyme, showing that molecules with selenazole moiety, cyano group and isopropyl group tended to yield higher IP. The absorption, distribution, metabolism, excretion and toxicity (ADMET) prediction results further enhanced the potential of these novel XOIs as drug candidates. Overall, this work presents a QSAR model for accurate prediction of IP of XOIs, and is expected to provide new insights for further structure-guided design of novel XOIs.

A Perspective on Explanations of Molecular Prediction Models

Chemists can be skeptical in using deep learning (DL) in decision making, due to the lack of interpretability in "black-box" models. Explainable artificial intelligence (XAI) is a branch of artificial intelligence (AI) which addresses this drawback by providing tools to interpret DL models and their predictions. We review the principles of XAI in the domain of chemistry and emerging methods for creating and evaluating explanations. Then, we focus on methods developed by our group and their applications in predicting solubility, blood-brain barrier permeability, and the scent of molecules. We show that XAI methods like chemical counterfactuals and descriptor explanations can explain DL predictions while giving insight into structure-property relationships. Finally, we discuss how a two-step process of developing a black-box model and explaining predictions can uncover structure-property relationships.

In Silico Tools to Extract the Drug Design Information Content of Degradation Data: The Case of PROTACs Targeting the Androgen Receptor

Proteolysis-Targeting Chimeras (PROTACs) have recently emerged as a promising technology in the drug discovery landscape. Large interest in the degradation of the androgen receptor (AR) as a new anti-prostatic cancer strategy has resulted in several papers focusing on PROTACs against AR. This study explores the potential of a few in silico tools to extract drug design information from AR degradation data in the format often reported in the literature. After setting up a dataset of 92 PROTACs with consistent AR degradation values, we employed the Bemis-Murcko method for their classification. The resulting clusters were not informative in terms of structure-degradation relationship. Subsequently, we performed Degradation Cliff analysis and identified some key aspects conferring a positive contribution to activity, as well as some methodological limits when applying this approach to PROTACs. Linker structure degradation relationships were also investigated. Then, we built and characterized ternary complexes to validate previous results. Finally, we implemented machine learning classification models and showed that AR degradation for VHL-based but not CRBN-based PROTACs can be predicted from simple permeability-related 2D molecular descriptors.

Large-scale prediction of activity cliffs using machine and deep learning methods of increasing complexity

Activity cliffs (AC) are formed by pairs of structural analogues that are active against the same target but have a large difference in potency. While much of our knowledge about ACs has originated from the analysis and comparison of compounds and activity data, several studies have reported AC predictions over the past decade. Different from typical compound classification tasks, AC predictions must be carried out at the level of compound pairs representing ACs or nonACs. Most AC predictions reported so far have focused on individual methods or comparisons of two or three approaches and only investigated a few compound activity classes (from 2 to 10). Although promising prediction accuracy has been reported in most cases, different system set-ups, AC definitions, methods, and calculation conditions were used, precluding direct comparisons of these studies. Therefore, we have carried out a large-scale AC prediction campaign across 100 activity classes comparing machine learning methods of greatly varying complexity, ranging from pair-based nearest neighbor classifiers and decision tree or kernel methods to deep neural networks. The results of our systematic predictions revealed the level of accuracy that can be expected for AC predictions across many different compound classes. In addition, prediction accuracy did not scale with methodological complexity but was significantly influenced by memorization of compounds shared by different ACs or nonACs. In many instances, limited training data were sufficient for building accurate models using different methods and there was no detectable advantage of deep learning over simpler approaches for AC prediction. On a global scale, support vector machine models performed best, by only small margins compared to others including simple nearest neighbor classifiers.

Computational Prediction of Drug Phenotypic Effects Based on Substructure-Phenotype Associations

Identifying drug phenotypic effects, including therapeutic effects and adverse drug reactions (ADRs), is an inseparable part for evaluating the potentiality of new drug candidates (NDCs). However, current computational methods for predicting phenotypic effects of NDCs are mainly based on the overall structure of an NDC or a related target. These approaches often lead to inconsistencies between the structures and functions and limit the prediction space of NDCs. In this study, first, we constructed quantitative associations of substructure-domain, domain-ADR, and domain-ATC (Anatomical Therapeutic Chemical Classification System code) through L1LOG and L1SVM machine learning models. These associations represent relationships between phenotypes (ADRs and ATCs) and local structures of drugs and proteins. Then, based on these established associations, substructure-phenotype relationships were constructed which were utilized to quantify drug-phenotype relationships. Thus, this approach could achieve high-throughput and effective evaluations of the druggability of NDCs by referring to the established substructure-phenotype relationships and structural information of NDCs without additional prior knowledge. Using this computational pipeline, 83,205 drug-ATC relationships (including 1,479 drugs and 178 ATCs) and 306,421 drug-ADR relationships (including 1,752 drugs and 454 ADRs) were predicted in total. The prediction results were validated at four levels: five-fold cross validation, public databases, literature, and molecular docking. Furthermore, three case studies demonstrated the feasibility of our method. 79 ATCs and 269 ADRs were predicted to be related to Maraviroc, an approved drug, including the existing antiviral effect in clinical use. Additionally, we also found risk substructures of severe ADRs, for example, SUB215 (>= 1, saturated or only aromatic carbon ring size 7) can result in shock. And we analyzed the mechanism of action (MOA) of interested drugs based on the established drug-substructure-domain-protein associations. In a word, this approach through establishing drug-substructure-phenotype relationships can achieve quantitative prediction of phenotypes for a given NDC or drug without any prior knowledge except its structure information. Using that way, we can directly obtain the relationships between substructure and phenotype of a compound, which is more convenient to analyze the phenotypic mechanism of drugs and accelerate the process of rational drug design.

Exposing the Limitations of Molecular Machine Learning with Activity Cliffs

Machine learning has become a crucial tool in drug discovery and chemistry at large, e.g., to predict molecular properties, such as bioactivity, with high accuracy. However, activity cliffs─pairs of molecules that are highly similar in their structure but exhibit large differences in potency─have received limited attention for their effect on model performance. Not only are these edge cases informative for molecule discovery and optimization but also models that are well equipped to accurately predict the potency of activity cliffs have increased potential for prospective applications. Our work aims to fill the current knowledge gap on best-practice machine learning methods in the presence of activity cliffs. We benchmarked a total of 24 machine and deep learning approaches on curated bioactivity data from 30 macromolecular targets for their performance on activity cliff compounds. While all methods struggled in the presence of activity cliffs, machine learning approaches based on molecular descriptors outperformed more complex deep learning methods. Our findings highlight large case-by-case differences in performance, advocating for (a) the inclusion of dedicated "activity-cliff-centered" metrics during model development and evaluation and (b) the development of novel algorithms to better predict the properties of activity cliffs. To this end, the methods, metrics, and results of this study have been encapsulated into an open-access benchmarking platform named MoleculeACE (Activity Cliff Estimation, available on GitHub at: https://github.com/molML/MoleculeACE). MoleculeACE is designed to steer the community toward addressing the pressing but overlooked limitation of molecular machine learning models posed by activity cliffs.

Design of MMP-1 inhibitors via SAR transfer and experimental validation

New matrix metalloproteinase 1 (MMP-1) inhibitors were predicted using the structure-activity relationship (SAR) transfer method based on a series of analogues of kinesin-like protein 11 (KIF11) inhibitors. Compounds 5-7 predicted to be highly potent against MMP-1 were synthesized and tested for MMP-1 inhibitory activity. Among these, compound 6 having a Cl substituent at the R¹ site was found to possess ca. 3.5 times higher inhibitory activity against MMP-1 than the previously reported compound 4. The observed potency was consistent with the presence of an SAR transfer event between analogous MMP-1 and KIF11 inhibitors. Pharmacophore fitting revealed that the higher inhibitory activity of compound 6 compared to compound 4 against MMP-1 might be due to a halogen bond interaction between the Cl substituent of compound 6 and residue ARG214 of MMP-1.

Navigating Transcriptomic Connectivity Mapping Workflows to Link Chemicals with Bioactivities

Screening new compounds for potential bioactivities against cellular targets is vital for drug discovery and chemical safety. Transcriptomics offers an efficient approach for assessing global gene expression changes, but interpreting chemical mechanisms from these data is often challenging. Connectivity mapping is a potential data-driven avenue for linking chemicals to mechanisms based on the observation that many biological processes are associated with unique gene expression signatures (gene signatures). However, mining the effects of a chemical on gene signatures for biological mechanisms is challenging because transcriptomic data contain thousands of noisy genes. New connectivity mapping approaches seeking to distinguish signal from noise continue to be developed, spurred by the promise of discovering chemical mechanisms, new drugs, and disease targets from burgeoning transcriptomic data. Here, we analyze these approaches in terms of diverse transcriptomic technologies, public databases, gene signatures, pattern-matching algorithms, and statistical evaluation criteria. To navigate the complexity of connectivity mapping, we propose a harmonized scheme to coherently organize and compare published workflows. We first standardize concepts underlying transcriptomic profiles and gene signatures based on various transcriptomic technologies such as microarrays, RNA-Seq, and L1000 and discuss the widely used data sources such as Gene Expression Omnibus, ArrayExpress, and MSigDB. Next, we generalize connectivity mapping as a pattern-matching task for finding similarity between a query (e.g., transcriptomic profile for new chemical) and a reference (e.g., gene signature of known target). Published pattern-matching approaches fall into two main categories: vector-based use metrics like correlation, Jaccard index, etc., and aggregation-based use parametric and nonparametric statistics (e.g., gene set enrichment analysis). The statistical methods for evaluating the performance of different approaches are described, along with comparisons reported in the literature on benchmark transcriptomic data sets. Lastly, we review connectivity mapping applications in toxicology and offer guidance on evaluating chemical-induced toxicity with concentration-response transcriptomic data. In addition to serving as a high-level guide and tutorial for understanding and implementing connectivity mapping workflows, we hope this review will stimulate new algorithms for evaluating chemical safety and drug discovery using transcriptomic data.

Pharmacophore Modeling of Targets Infested with Activity Cliffs via Molecular Dynamics Simulation Coupled with QSAR and Comparison with other Pharmacophore Generation Methods: KDR as Case Study

Activity cliffs (ACs) are defined as pairs of structurally similar compounds with large difference in their potencies against certain biotarget. We recently proposed that potent AC members induce significant entropically-driven conformational modifications of the target that unveil additional binding interactions, while their weakly-potent counterparts are enthalpically-driven binders with little influence on the protein target. We herein propose to extract pharmacophores for ACs-infested target(s) from molecular dynamics (MD) frames of purely "enthalpic" potent binder(s) complexed within the particular target. Genetic function algorithm/machine learning (GFA/ML) can then be employed to search for the best possible combination of MD pharmacophore(s) capable of explaining bioactivity variations within a list of inhibitors. We compared the performance of this approach with established ligand-based and structure-based methods. Kinase inserts domain receptor (KDR) was used as a case study. KDR plays a crucial role in angiogenic signaling and its inhibitors have been approved in cancer treatment. Interestingly, GFA/ML selected, MD-based, pharmacophores were of comparable performances to ligand-based and structure-based pharmacophores. The resulting pharmacophores and QSAR models were used to capture hits from the national cancer institute list of compounds. The most active hit showed anti-KDR IC50 of 2.76 µM.

Drug Chemical Space as a Guide for New Herbicide Development: A Cheminformatic Analysis

Herbicides are critical resources for meeting agricultural demand. While similar in structure and function to pharmaceuticals, the development of new herbicidal mechanisms of action and new scaffolds against known mechanisms of action has been much slower than in pharmaceutical sciences. We hypothesized that this may be due in part to a relative undersampling of possible herbicidal chemistries and set out to test whether this difference in sampling existed and whether increasing the diversity of possible herbicidal chemistries would be likely to result in more efficacious herbicides. To conduct this work, we first identified databases of commercially available herbicides and clinically approved pharmaceuticals. Using these databases, we created a two-dimensional embedding of the chemical, which provides a qualitative visualization of the degree to which each chemotype is distributed within the combined chemical space and shows a moderate degree of overlap between the two sets. Next, we trained several machine learning models to classify herbicides versus drugs based on physicochemical characteristics. The most accurate of these models has an accuracy of 93% with the key differentiating characteristics being the number of polar hydrogens, number of amide bonds, LogP, and polar surface area. We then used several types of scaffold decomposition to quantitatively evaluate the chemical diversity of each molecular family and showed herbicides to have considerably fewer unique structural fragments. Finally, we used molecular docking as an in silico evaluation of further structural diversification in herbicide development. To this end, we identified herbicides with well-characterized binding sites and modified those scaffolds based on similar structural subunits from the drug dataset not present in any commercial herbicide while using the machine-learned model to ensure that required herbicide properties were maintained. Redocking the original and modified scaffolds of several herbicides showed that even this simple design strategy is capable of yielding new molecules with higher predicted affinity for the target enzymes. Overall, we show that herbicides are distinct from drugs based on physicochemical properties but less diverse in their chemistry in a way not governed by these properties. We also demonstrate in silico that increasing the diversity of herbicide scaffolds has the potential to increase potency, potentially reducing the amount needed in agricultural practice.

A new quinolinone-chalcone hybrid with potential antibacterial and herbicidal properties using in silico approaches

Quinolinone-chalcones are hybrid compounds consisting of chalcone and quinolone moieties with biological activity related to their hybrid structure. This work seeks to describe the structural and theoretical parameters related to the physicochemical properties and biological activity of a new quinolinone-chalcone. The synthesis, structural characterization by X-ray diffraction, molecular topology by Hirshfeld surfaces and QTAIM, molecular electronic calculations, and pharmacophore analysis were described. The weak interactions C-H^…O, C-H^…π, and C-H^…Br were responsible for crystal growth and stabilized the crystalline state. The DFT analysis shows that the sulfonamide group region is susceptible to observed interactions, and the frontier molecular orbitals indicate high kinetic stability. Also, pharmacophore analysis revealed potential antibacterial and herbicidal activity; by docking within the active site of TtgR, a transcription regulator for the efflux pump TtgABC from the highly resistant Pseudomonas putida (P. putida) strain DOT-TIE, we showed that the activation of TtgR relies upon the binding of aromatic-harboring compounds, which plays a crucial role in bacterial evasion. In this context, a new quinolinone-chalcone has a higher binding affinity than tetracycline, which suggests it might be a better effector for TtgR.

Quantitative Structure-Activity Relationship (QSAR) Study Predicts Small-Molecule Binding to RNA Structure

The diversity of RNA structural elements and their documented role in human diseases make RNA an attractive therapeutic target. However, progress in drug discovery and development has been hindered by challenges in the determination of high-resolution RNA structures and a limited understanding of the parameters that drive RNA recognition by small molecules, including a lack of validated quantitative structure-activity relationships (QSARs). Herein, we develop QSAR models that quantitatively predict both thermodynamic- and kinetic-based binding parameters of small molecules and the HIV-1 transactivation response (TAR) RNA model system. Small molecules bearing diverse scaffolds were screened against TAR using surface plasmon resonance. Multiple linear regression (MLR) combined with feature selection afforded robust models that allowed direct interpretation of the properties critical for both binding strength and kinetic rate constants. These models were validated with new molecules, and their accurate performance was confirmed via comparison to ensemble tree methods, supporting the general applicability of this platform.

Exploration of chemical space with partial labeled noisy student self-training and self-supervised graph embedding

BACKGROUND

Drug discovery is time-consuming and costly. Machine learning, especially deep learning, shows great potential in quantitative structure-activity relationship (QSAR) modeling to accelerate drug discovery process and reduce its cost. A big challenge in developing robust and generalizable deep learning models for QSAR is the lack of a large amount of data with high-quality and balanced labels. To address this challenge, we developed a self-training method, Partially LAbeled Noisy Student (PLANS), and a novel self-supervised graph embedding, Graph-Isomorphism-Network Fingerprint (GINFP), for chemical compounds representations with substructure information using unlabeled data. The representations can be used for predicting chemical properties such as binding affinity, toxicity, and others. PLANS-GINFP allows us to exploit millions of unlabeled chemical compounds as well as labeled and partially labeled pharmacological data to improve the generalizability of neural network models.

RESULTS

We evaluated the performance of PLANS-GINFP for predicting Cytochrome P450 (CYP450) binding activity in a CYP450 dataset and chemical toxicity in the Tox21 dataset. The extensive benchmark studies demonstrated that PLANS-GINFP could significantly improve the performance in both cases by a large margin. Both PLANS-based self-training and GINFP-based self-supervised learning contribute to the performance improvement.

CONCLUSION

To better exploit chemical structures as an input for machine learning algorithms, we proposed a self-supervised graph neural network-based embedding method that can encode substructure information. Furthermore, we developed a model agnostic self-training method, PLANS, that can be applied to any deep learning architectures to improve prediction accuracies. PLANS provided a way to better utilize partially labeled and unlabeled data. Comprehensive benchmark studies demonstrated their potentials in predicting drug metabolism and toxicity profiles using sparse, noisy, and imbalanced data. PLANS-GINFP could serve as a general solution to improve the predictive modeling for QSAR modeling.

AFSE: towards improving model generalization of deep graph learning of ligand bioactivities targeting GPCR proteins

Ligand molecules naturally constitute a graph structure. Recently, many excellent deep graph learning (DGL) methods have been proposed and used to model ligand bioactivities, which is critical for the virtual screening of drug hits from compound databases in interest. However, pharmacists can find that these well-trained DGL models usually are hard to achieve satisfying performance in real scenarios for virtual screening of drug candidates. The main challenges involve that the datasets for training models were small-sized and biased, and the inner active cliff cases would worsen model performance. These challenges would cause predictors to overfit the training data and have poor generalization in real virtual screening scenarios. Thus, we proposed a novel algorithm named adversarial feature subspace enhancement (AFSE). AFSE dynamically generates abundant representations in new feature subspace via bi-directional adversarial learning, and then minimizes the maximum loss of molecular divergence and bioactivity to ensure local smoothness of model outputs and significantly enhance the generalization of DGL models in predicting ligand bioactivities. Benchmark tests were implemented on seven state-of-the-art open-source DGL models with the potential of modeling ligand bioactivities, and precisely evaluated by multiple criteria. The results indicate that, on almost all 33 GPCRs datasets and seven DGL models, AFSE greatly improved their enhancement factor (top-10%, 20% and 30%), which is the most important evaluation in virtual screening of hits from compound databases, while ensuring the superior performance on RMSE and $r^2$. The web server of AFSE is freely available at http://noveldelta.com/AFSE for academic purposes.

Evolution of Support Vector Machine and Regression Modeling in Chemoinformatics and Drug Discovery

The support vector machine (SVM) algorithm is one of the most widely used machine learning (ML) methods for predicting active compounds and molecular properties. In chemoinformatics and drug discovery, SVM has been a state-of-the-art ML approach for more than a decade. A unique attribute of SVM is that it operates in feature spaces of increasing dimensionality. Hence, SVM conceptually departs from the paradigm of low dimensionality that applies to many other methods for chemical space navigation. The SVM approach is applicable to compound classification, and ranking, multi-class predictions, and –in algorithmically modified form– regression modeling. In the emerging era of deep learning (DL), SVM retains its relevance as one of the premier ML methods in chemoinformatics, for reasons discussed herein. We describe the SVM methodology including strengths and weaknesses and discuss selected applications that have contributed to the evolution of SVM as a premier approach for compound classification, property predictions, and virtual compound screening.

Application of deep metric learning to molecular graph similarity

Graph based methods are increasingly important in chemistry and drug discovery, with applications ranging from QSAR to molecular generation. Combining graph neural networks and deep metric learning concepts, we expose a framework for quantifying molecular graph similarity based on distance between learned embeddings separate from any endpoint. Using a minimal definition of similarity, and data from the ZINC database of public compounds, this work demonstrate the properties of the embedding and its suitability for a range of applications, among them a novel reconstruction loss method for training deep molecular auto-encoders. Finally, we compare the applications of the embedding to standard practices, with a focus on known failure points and edge cases; concluding that our approach can be used in conjunction to existing methods.

Assisting Multitargeted Ligand Affinity Prediction of Receptor Tyrosine Kinases Associated Nonsmall Cell Lung Cancer Treatment with Multitasking Principal Neighborhood Aggregation

A multitargeted therapeutic approach with hybrid drugs is a promising strategy to enhance anticancer efficiency and overcome drug resistance in nonsmall cell lung cancer (NSCLC) treatment. Estimating affinities of small molecules against targets of interest typically proceeds as a preliminary action for recent drug discovery in the pharmaceutical industry. In this investigation, we employed machine learning models to provide a computationally affordable means for computer-aided screening to accelerate the discovery of potential drug compounds. In particular, we introduced a quantitative structure–activity-relationship (QSAR)-based multitask learning model to facilitate an in silico screening system of multitargeted drug development. Our method combines a recently developed graph-based neural network architecture, principal neighborhood aggregation (PNA), with a descriptor-based deep neural network supporting synergistic utilization of molecular graph and fingerprint features. The model was generated by more than ten-thousands affinity-reported ligands of seven crucial receptor tyrosine kinases in NSCLC from two public data sources. As a result, our multitask model demonstrated better performance than all other benchmark models, as well as achieving satisfying predictive ability regarding applicable QSAR criteria for most tasks within the model’s applicability. Since our model could potentially be a screening tool for practical use, we have provided a model implementation platform with a tutorial that is freely accessible hence, advising the first move in a long journey of cancer drug development.

Systematic review on the application of machine learning to quantitative structure-activity relationship modeling against Plasmodium falciparum

Malaria accounts for over two million deaths globally. To flatten this curve, there is a need to develop new and high potent drugs against Plasmodium falciparum. Some major challenges include the dearth of suitable animal models for anti-P. falciparum assays, resistance to first-line drugs, lack of vaccines and the complex life cycle of Plasmodium. Gladly, newer approaches to antimalarial drug discovery have emerged due to the release of large datasets by pharmaceutical companies. This review provides insights into these new approaches to drug discovery covering different machine learning tools, which enhance the development of new compounds. It provides a systematic review on the use and prospects of machine learning in predicting, classifying and clustering IC₅₀ values of bioactive compounds against P. falciparum. The authors identified many machine learning tools yet to be applied for this purpose. However, Random Forest and Support Vector Machines have been extensively applied though on a limited dataset of compounds.

Systematic identification of activity cliffs with dual-atom replacements and their rationalization on the basis of single-atom replacement analogs and X-ray structures

Very small chemical changes in active compounds causing large potency effects are of particular interest in medicinal chemistry and drug design. We have systematically searched active compounds with available high-confidence activity data for pairs of structural analogs with dual-atom replacements and additional analogs with corresponding single-atom replacements. From ~287,000 unique qualifying compounds with activity against nearly 1900 unique targets, ~3500 target-based analog pairs with dual-atom replacements were identified. These included 852 pairs with significant differences in compound potency, representing a set of previously unobserved activity cliffs. Comparing these pairs with corresponding single-atom replacement analogs, which were frequently identified, made it possible to systematically analyze how potency changes propagated from single- to dual-atom replacements. The analysis uncovered different potency effects and revealed that individual atom replacements were often decisive for activity cliff formation. For a limited number of activity cliffs, X-ray structures of targets in complex with cliff compounds were available, which aided in rationalizing potency alterations among analogs with single- or dual-atom replacements. The analog pairs identified herein provide a rich resource of structure-activity relationship information and attractive test cases for calibrating computational methods.

Deep generative design with 3D pharmacophoric constraints

Generative models have increasingly been proposed as a solution to the molecular design problem. However, it has proved challenging to control the design process or incorporate prior knowledge, limiting their practical use in drug discovery. In particular, generative methods have made limited use of three-dimensional (3D) structural information even though this is critical to binding. This work describes a method to incorporate such information and demonstrates the benefit of doing so. We combine an existing graph-based deep generative model, DeLinker, with a convolutional neural network to utilise physically-meaningful 3D representations of molecules and target pharmacophores. We apply our model, DEVELOP, to both linker and R-group design, demonstrating its suitability for both hit-to-lead and lead optimisation. The 3D pharmacophoric information results in improved generation and allows greater control of the design process. In multiple large-scale evaluations, we show that including 3D pharmacophoric constraints results in substantial improvements in the quality of generated molecules. On a challenging test set derived from PDBbind, our model improves the proportion of generated molecules with high 3D similarity to the original molecule by over 300%. In addition, DEVELOP recovers 10× more of the original molecules compared to the baseline DeLinker method. Our approach is general-purpose, readily modifiable to alternate 3D representations, and can be incorporated into other generative frameworks. Code is available at https://github.com/oxpig/DEVELOP.

Planning new Trypanosoma cruzi CYP51 inhibitors using QSAR studies

Chagas disease kills over 10,000 people per year, and approximately 8 million people are infected by Trypanosoma cruzi. The reference drug for treatment of the disease, benznidazole, is the same since the 70s. In recent years, many CYP51 inhibitors were tested against this parasite's target. One of them, posaconazole, was even tested in clinical trials that unfortunately were not successful. Nevertheless, there are still many evidences that CYP51 is a great potential target to treat T. cruzi infection. The research for new effective molecules that can cure the chronic phase of the disease is essential. 2D and 3D-quantitative structure activity relationship (QSAR) studies were conducted in this work to create three QSAR models using the chemical structures of 197 published compounds that already went through either in vivo or in vitro tests. After the analysis of the models, new analogues not yet synthesized were suggested here and had their biological activity and synthetic availability assessed.

Therapeutic target database update 2022: facilitating drug discovery with enriched comparative data of targeted agents

Drug discovery relies on the knowledge of not only drugs and targets, but also the comparative agents and targets. These include poor binders and non-binders for developing discovery tools, prodrugs for improved therapeutics, co-targets of therapeutic targets for multi-target strategies and off-target investigations, and the collective structure-activity and drug-likeness landscapes of enhanced drug feature. However, such valuable data are inadequately covered by the available databases. In this study, a major update of the Therapeutic Target Database, previously featured in NAR, was therefore introduced. This update includes (a) 34 861 poor binders and 12 683 non-binders of 1308 targets; (b) 534 prodrug-drug pairs for 121 targets; (c) 1127 co-targets of 672 targets regulated by 642 approved and 624 clinical trial drugs; (d) the collective structure-activity landscapes of 427 262 active agents of 1565 targets; (e) the profiles of drug-like properties of 33 598 agents of 1102 targets. Moreover, a variety of additional data and function are provided, which include the cross-links to the target structure in PDB and AlphaFold, 159 and 1658 newly emerged targets and drugs, and the advanced search function for multi-entry target sequences or drug structures. The database is accessible without login requirement at: https://idrblab.org/ttd/.

Evaluating High-Variance Leaves as Uncertainty Measure for Random Forest Regression

Uncertainty measures estimate the reliability of a predictive model. Especially in the field of molecular property prediction as part of drug design, model reliability is crucial. Besides other techniques, Random Forests have a long tradition in machine learning related to chemoinformatics and are widely used. Random Forests consist of an ensemble of individual regression models, namely, decision trees and, therefore, provide an uncertainty measure already by construction. Regarding the disagreement of single-model predictions, a narrower distribution of predictions is interpreted as a higher reliability. The standard deviation of the decision tree ensemble predictions is the default uncertainty measure for Random Forests. Due to the increasing application of machine learning in drug design, there is a constant search for novel uncertainty measures that, ideally, outperform classical uncertainty criteria. When analyzing Random Forests, it appears obvious to consider the variance of the dependent variables within each terminal decision tree leaf to obtain predictive uncertainties. Hereby, predictions that arise from more leaves of high variance are considered less reliable. Expectedly, the number of such high-variance leaves yields a reasonable uncertainty measure. Depending on the dataset, it can also outperform ensemble uncertainties. However, small-scale comparisons, i.e., considering only a few datasets, are insufficient, since they are more prone to chance correlations. Therefore, large-scale estimations are required to make general claims about the performance of uncertainty measures. On several chemoinformatic regression datasets, high-variance leaves are compared to the standard deviation of ensemble predictions. It turns out that high-variance leaf uncertainty is meaningful, not superior to the default ensemble standard deviation. A brief possible explanation is offered.

Predicting the Relative Binding Affinity for Reversible Covalent Inhibitors by Free Energy Perturbation Calculations

Covalent inhibitors are assuming central importance in drug discovery projects, especially in this pandemic scenario. Many research groups have focused their attention on inhibiting viral proteases or human proteases such as cathepsin L (hCatL). The inhibition of these critical enzymes may impair viral replication. However, molecular modeling of covalent ligands is challenging since covalent and noncovalent ligand-bound states must be considered in the binding process. In this work, we evaluated the suitability of free energy perturbation (FEP) calculations as a tool for predicting the binding affinity of reversible covalent inhibitors of hCatL. Our strategy relies on the relative free energy calculated for both covalent and noncovalent complexes and the free energy changes have been compared with experimental data for eight nitrile-based inhibitors, including three new inhibitors of hCatL. Our results demonstrate that the covalent complex can be employed to properly rank the inhibitors. Nevertheless, a comparison of the free energy changes in both noncovalent and covalent states is valuable to interpret the effect triggered by the formation of the covalent bond on the interactions played by functional groups distant from the warhead. Overall, FEP can be employed as a powerful predictor tool in developing and understanding the activity of reversible covalent inhibitors.

SimilarityLab: Molecular Similarity for SAR Exploration and Target Prediction on the Web

Exploration of chemical space around hit, experimental, and known active compounds is an important step in the early stages of drug discovery. In academia, where access to chemical synthesis efforts is restricted in comparison to the pharma-industry, hits from primary screens are typically followed up through purchase and testing of similar compounds, before further funding is sought to begin medicinal chemistry efforts. Rapid exploration of druglike similars and structure–activity relationship profiles can be achieved through our new webservice SimilarityLab. In addition to searching for commercially available molecules similar to a query compound, SimilarityLab also enables the search of compounds with recorded activities, generating consensus counts of activities, which enables target and off-target prediction. In contrast to other online offerings utilizing the USRCAT similarity measure, SimilarityLab’s set of commercially available small molecules is consistently updated, currently containing over 12.7 million unique small molecules, and not relying on published databases which may be many years out of date. This ensures researchers have access to up-to-date chemistries and synthetic processes enabling greater diversity and access to a wider area of commercial chemical space. All source code is available in the SimilarityLab source repository.

Interpretation of Ligand-Based Activity Cliff Prediction Models Using the Matched Molecular Pair Kernel

Activity cliffs (ACs) are formed by two structurally similar compounds with a large difference in potency. Accurate AC prediction is expected to help researchers' decisions in the early stages of drug discovery. Previously, predictive models based on matched molecular pair (MMP) cliffs have been proposed. However, the proposed methods face a challenge of interpretability due to the black-box character of the predictive models. In this study, we developed interpretable MMP fingerprints and modified a model-specific interpretation approach for models based on a support vector machine (SVM) and MMP kernel. We compared important features highlighted by this SVM-based interpretation approach and the SHapley Additive exPlanations (SHAP) as a major model-independent approach. The model-specific approach could capture the difference between AC and non-AC, while SHAP assigned high weights to the features not present in the test instances. For specific MMPs, the feature weights mapped by the SVM-based interpretation method were in agreement with the previously confirmed binding knowledge from X-ray co-crystal structures, indicating that this method is able to interpret the AC prediction model in a chemically intuitive manner.

Antiplasmodial 2-thiophenoxy-3-trichloromethyl quinoxalines target the apicoplast of Plasmodium falciparum

The identification of a plant-like Achille's Heel relict, i.e. the apicoplast, that is essential for Plasmodium spp., the causative agent of malaria lead to an attractive drug target for new antimalarials with original mechanism of action. Although it is not photosynthetic, the apicoplast retains several anabolic pathways that are indispensable for the parasite. Based on previously identified antiplasmodial hit-molecules belonging to the 2-trichloromethylquinazoline and 3-trichloromethylquinoxaline series, we report herein an antiplasmodial Structure-Activity Relationships (SAR) study at position two of the quinoxaline ring of 16 newly synthesized compounds. Evaluation of their activity toward the multi-resistant K1 Plasmodium falciparum strain and cytotoxicity on the human hepatocyte HepG2 cell line revealed a hit compound (3k) with a PfK1 EC₅₀ value of 0.3 μM and a HepG2 CC₅₀ value of 56.0 μM (selectivity index = 175). Moreover, hit-compound 3k was not cytotoxic on VERO or CHO cell lines and was not genotoxic in the in vitro comet assay. Activity cliffs were observed when the trichloromethyl group was replaced by CH₃, CF₃ or H, showing that this group played a key role in the antiplasmodial activity. Biological investigations performed to determine the target and mechanism of action of the compound 3k strongly suggest that the apicoplast is the putative target as showed by severe alteration of apicoplaste biogenesis and delayed death response. Considering that there are very few molecules that affect the Plasmodium apicoplast, our work provides, for the first time, evidence of the biological target of trichloromethylated derivatives.

2-Phenoxy-3-Trichloromethylquinoxalines Are Antiplasmodial Derivatives with Activity against the Apicoplast of Plasmodium falciparum

The malaria parasite harbors a relict plastid called the apicoplast. Although not photosynthetic, the apicoplast retains unusual, non-mammalian metabolic pathways that are essential to the parasite, opening up a new perspective for the development of novel antimalarials which display a new mechanism of action. Based on the previous antiplasmodial hit-molecules identified in the 2-trichloromethylquinoxaline series, we report herein a structure–activity relationship (SAR) study at position two of the quinoxaline ring by synthesizing 20 new compounds. The biological evaluation highlighted a hit compound (3i) with a potent PfK1 EC₅₀ value of 0.2 µM and a HepG2 CC₅₀ value of 32 µM (Selectivity index = 160). Nitro-containing (3i) was not genotoxic, both in the Ames test and in vitro comet assay. Activity cliffs were observed when the 2-CCl₃ group was replaced, showing that it played a key role in the antiplasmodial activity. Investigation of the mechanism of action showed that 3i presents a drug response by targeting the apicoplast and a quick-killing mechanism acting on another target site.

Mechanism-Driven Read-Across of Chemical Hepatotoxicants Based on Chemical Structures and Biological Data

Hepatotoxicity is a leading cause of attrition in the drug development process. Traditional preclinical and clinical studies to evaluate hepatotoxicity liabilities are expensive and time consuming. With the advent of critical advancements in high-throughput screening, there has been a rapid accumulation of in vitro toxicity data available to inform the risk assessment of new pharmaceuticals and chemicals. To this end, we curated and merged all available in vivo hepatotoxicity data obtained from the literature and public resources, which yielded a comprehensive database of 4089 compounds that includes hepatotoxicity classifications. After dividing the original database of chemicals into modeling and test sets, PubChem assay data were automatically extracted using an in-house data mining tool and clustered based on relationships between structural fragments and cellular responses in in vitro assays. The resultant PubChem assay clusters were further investigated. During the cross-validation procedure, the biological data obtained from several assay clusters exhibited high predictivity of hepatotoxicity and these assays were selected to evaluate the test set compounds. The read-across results indicated that if a new compound contained specific identified chemical fragments (ie, Molecular Initiating Event) and showed active responses in the relevant selected PubChem assays, there was potential for the chemical to be hepatotoxic in vivo. Furthermore, several mechanisms that might contribute to toxicity were derived from the modeling results including alterations in nuclear receptor signaling and inhibition of DNA repair. This modeling strategy can be further applied to the investigation of other complex chemical toxicity phenomena (eg, developmental and reproductive toxicities) as well as drug efficacy.

Extended similarity indices: the benefits of comparing more than two objects simultaneously. Part 1: Theory and characteristics†

Quantification of the similarity of objects is a key concept in many areas of computational science. This includes cheminformatics, where molecular similarity is usually quantified based on binary fingerprints. While there is a wide selection of available molecular representations and similarity metrics, there were no previous efforts to extend the computational framework of similarity calculations to the simultaneous comparison of more than two objects (molecules) at the same time. The present study bridges this gap, by introducing a straightforward computational framework for comparing multiple objects at the same time and providing extended formulas for as many similarity metrics as possible. In the binary case (i.e. when comparing two molecules pairwise) these are naturally reduced to their well-known formulas. We provide a detailed analysis on the effects of various parameters on the similarity values calculated by the extended formulas. The extended similarity indices are entirely general and do not depend on the fingerprints used. Two types of variance analysis (ANOVA) help to understand the main features of the indices: (i) ANOVA of mean similarity indices; (ii) ANOVA of sum of ranking differences (SRD). Practical aspects and applications of the extended similarity indices are detailed in the accompanying paper: Miranda-Quintana et al. J Cheminform. 2021. https://doi.org/10.1186/s13321-021-00504-4 . Python code for calculating the extended similarity metrics is freely available at: https://github.com/ramirandaq/MultipleComparisons .

Prediction of activity cliffs on the basis of images using convolutional neural networks

An activity cliff (AC) is formed by a pair of structurally similar compounds with a large difference in potency. Accordingly, ACs reveal structure–activity relationship (SAR) discontinuity and provide SAR information for compound optimization. Herein, we have investigated the question if ACs could be predicted from image data. Therefore, pairs of structural analogs were extracted from different compound activity classes that formed or did not form ACs. From these compound pairs, consistently formatted images were generated. Image sets were used to train and test convolutional neural network (CNN) models to systematically distinguish between ACs and non-ACs. The CNN models were found to predict ACs with overall high accuracy, as assessed using alternative performance measures, hence establishing proof-of-principle. Moreover, gradient weights from convolutional layers were mapped to test compounds and identified characteristic structural features that contributed to successful predictions. Weight-based feature visualization revealed the ability of CNN models to learn chemistry from images at a high level of resolution and aided in the interpretation of model decisions with intrinsic black box character.

Protocol for hit-to-lead optimization of compounds by auto in silico ligand directing evolution (AILDE) approach

Hit-to-lead (H2L) optimization is crucial for drug design, which has become an increasing concern in medicinal chemistry. A virtual screening strategy of auto in silico ligand directing evolution (AILDE) has been developed to yield promising lead compounds rapidly and efficiently. The protocol includes instructions for fragment compound library construction, conformational sampling by molecular dynamics simulation, ligand modification by fragment growing, as well as the binding free energy prediction. For complete details on the use and execution of this protocol, please refer to Wu et al. (2020).

Comparative study of Danshen and Siwu decoction based on the molecular structures of the components and predicted targets

BACKGROUND

The sentence of "Danshen (Salvia Miltiorrhizae Radix et Rhizoma) and Siwu decoction are similar in function" was first recorded in an ancient Chinese medical book "Fu Ren Ming Li Lun". This theory has far-reaching influence on the clinical practice of Chinese medicine and is highly respected by Chinese medical doctors. However, the theory has limitations and controversial part for there is no in-depth and system comparative study.

METHODS

We collected the molecular structures of 129 compounds of Danshen and 81 compounds of Siwu decoction from the literatures. MACCS fingerprints and Tanimoto similarity were calculated based on the molecular structures for comparing the structural feature. Molecular descriptors which represent physical and chemical properties were calculated by Discovery Studio. Principal component analysis (PCA) of was performed based on the descriptors. The ADMET properties were predicted by FAF-Drugs4. The effect targets for the compounds with good ADMET properties were confirmed from experimental data and predicted using the algorithm comprising Bernoulli Naive Bayes profiling.

RESULTS

Based on the molecular structures, the presented study compared the structural feature, physical and chemical properties, ADMET properties, and effect targets of compounds of Danshen and Siwu decoction. It is found that Danshen and Siwu decoction do not have the same main active components. Moreover, the 2D structure of compounds from Danshen and Siwu decoction is not similar. Some of the compounds of Danshen and Siwu decoction are similar in 3D structure. The compounds with good ADMET properties of Danshen and Siwu decoction have same predicted targets, but some have different targets.

CONCLUSIONS

It can be inferred from the result that Danshen and Siwu decoction have some similarities, but also present differences from each other in the structure of the compounds and predicted targets. This may be the material basis of the similar and different traditional efficacy of Danshen and Siwu decoction. The setence of " Danshen and Siwu decoction are similar in function. " which is used in clinical has its material basis and target connotation to some extent. However, the traditional effects of Danshen and Siwu decoction are not exactly the same.

Evaluation of the Structure-Activity Relationship of Microtubule-Targeting 1,2,4-Triazolo[1,5-a]pyrimidines Identifies New Candidates for Neurodegenerative Tauopathies

Studies in tau and Aβ plaque transgenic mouse models demonstrated that brain-penetrant microtubule (MT)-stabilizing compounds, including the 1,2,4-triazolo[1,5-a]pyrimidines, hold promise as candidate treatments for Alzheimer's disease and related neurodegenerative tauopathies. Triazolopyrimidines have already been investigated as anticancer agents; however, the antimitotic activity of these compounds does not always correlate with stabilization of MTs in cells. Indeed, previous studies from our laboratories identified a critical role for the fragment linked at C6 in determining whether triazolopyrimidines promote MT stabilization or, conversely, disrupt MT integrity in cells. To further elucidate the structure-activity relationship (SAR) and to identify potentially improved MT-stabilizing candidates for neurodegenerative disease, a comprehensive set of 68 triazolopyrimidine congeners bearing structural modifications at C6 and/or C7 was designed, synthesized, and evaluated. These studies expand upon prior understanding of triazolopyrimidine SAR and enabled the identification of novel analogues that, relative to the existing lead, exhibit improved physicochemical properties, MT-stabilizing activity, and pharmacokinetics.

Cheminformatics Analysis of Fluoroquinolones and their Inhibition Potency Against Four Pathogens

Drug-resistant bacteria are a worldwide public health concern. As the prevalence of multi-drug resistant pathogens outpaces the discovery of new antibacterials, it is of importance to explore the structure-activity relationships for series of known bactericides with proven scaffolds. Herein, we assembled a set of 507 fluoroquinolone analogues all experimentally tested for their inhibition potency against four pathogens: Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus, and Streptococcus pneumoniae. We relied on cheminformatics techniques to characterize and cluster them based on their structural similarity and analyzed the structure-activity relationships identified for each cluster of fluoroquinolones. Then, we utilized machine learning techniques to develop and validate predictive QSAR models for computing the inhibition potencies (pMIC) of analogues for each pathogen. These QSAR models afforded reasonable external prediction performances (R2≥0.6, MAE∼0.4). This study confirmed that (i) there are both global and local inter-pathogen concordance regarding the antibacterial potency of fluoroquinolones, (ii) small clusters of fluoroquinolone analogues are characterized by unique patterns of strain selectivity and potency, the latter being potentially useful to design new analogues with enhanced potency and/or selectivity towards a given pathogen, and (iii) robust QSAR models were obtained allowing for future design of new bioactive fluoroquinolones.