The exploration of chemical space is a fundamental aspect of chemoinformatics, particularly when one explores a large compound data set to relate chemical structures with molecular properties. In this study, we extend our previous work on chemical space visualization at the pharmacophoric level. Instead of using conventional binary classification of affinity (active vs inactive), we introduce a refined approach that categorizes compounds into four distinct classes based on their activity levels: super active, very active, active, and inactive. This classification enriches the color scheme applied to pharmacophore space, where the color representation of a pharmacophore hypothesis is driven by the associated compounds. Using the BCR-ABL tyrosine kinase as a case study, we identified intriguing regions corresponding to pharmacophore activity discontinuities, providing valuable insights for structure-activity relationships analysis.

The therapeutic effects of Chinese herbal compounds are often achieved through the synergistic interactions of multiple ingredients. However, current research predominantly focuses on individual ingredients, neglecting the holistic nature of Chinese herbal compounds. This study proposes a novel strategy to elucidate the pharmacodynamic material basis of Chinese herbal compounds based on their multi-components (components named \'ZuFen\' in China, it refers to multiple ingredients with similar chemical structures) composition, using the Xian-Ling-Gu-Bao (XLGB) capsule as a case study. Cheminformatics-based components partitioning was conducted after sourcing ingredients from various databases, resulting in a total of 856 ingredients which were categorized into nine major components. Furthermore, the pharmacodynamic ingredients of XLGB capsule were determined by analyzing the ingredients that were absorbed into the bloodstream. Through a combination of these ingredients and screening for absorption, the Dipsacus asper saponin components, Psoralea corylifolia coumarin components, and Epimedium flavonoid polyglycosides components were isolated. The anti-osteoporosis efficacy of these components were evaluated in zebrafish, demonstrating their capability to reverse mineralization reduction caused by prednisolone. These findings further support the idea that these components serve as the material basis for the pharmacological efficacy of XLGB capsule. This study provides a novel systematic strategy for discovering the pharmacodynamic material basis of the efficacy of Chinese herbal compounds based on a \'multi-components\' perspective.

Scientific workflows facilitate the automation of data analysis tasks by integrating various software and tools executed in a particular order. To enable transparency and reusability in workflows, it is essential to implement the FAIR principles. Here, we describe our experiences implementing the FAIR principles for metabolomics workflows using the Metabolome Annotation Workflow (MAW) as a case study. MAW is specified using the Common Workflow Language (CWL), allowing for the subsequent execution of the workflow on different workflow engines. MAW is registered using a CWL description on WorkflowHub. During the submission process on WorkflowHub, a CWL description is used for packaging MAW using the Workflow RO-Crate profile, which includes metadata in Bioschemas. Researchers can use this narrative discussion as a guideline to commence using FAIR practices for their bioinformatics or cheminformatics workflows while incorporating necessary amendments specific to their research area.

To accurately simulate the inner workings of an enzyme active site with quantum mechanics (QM), not only must the reactive species be included in the model but also important surrounding residues, solvent, or coenzymes involved in crafting the microenvironment. Our lab has been developing the Residue Interaction Network Residue Selector (RINRUS) toolkit to utilize interatomic contact network information for automated, rational residue selection and QM-cluster model generation. Starting from an x-ray crystal structure of catechol-O-methyltransferase, RINRUS was used to construct a series of QM-cluster models. The reactant, product, and transition state of the methyl transfer reaction were computed for a total of 550 models, and the resulting free energies of activation and reaction were used to evaluate model convergence. RINRUS-designed models with only 200-300 atoms are shown to converge. RINRUS will serve as a cornerstone for improved and automated cheminformatics-based enzyme model design.

Antimicrobial resistance (AMR) is one of the most serious global public health threats as it compromises the successful treatment of deadly infectious diseases like tuberculosis. New therapeutics are constantly needed but it takes a long time and is expensive to explore new biochemical space. One way to address this issue is to repurpose the validated targets and identify novel chemotypes that can simultaneously bind to multiple binding pockets of these targets as a new lead generation strategy. This study reports such a strategy, dynamic hybrid pharmacophore model (DHPM), which represents the combined interaction features of different binding pockets contrary to the conventional approaches, where pharmacophore models are generated from single binding sites. We have considered Mtb-DapB, a validated mycobacterial drug target, as our model system to explore the effectiveness of DHPMs to screen novel unexplored compounds. Mtb-DapB has a cofactor binding site (CBS) and an adjacent substrate binding site (SBS). Four different model systems of Mtb-DapB were designed where, either NADPH/NADH occupies CBS in presence/absence of an inhibitor 2, 6-PDC in the adjacent SBS. Two more model systems were designed, where 2, 6-PDC was linked to NADPH and NADH to form hybrid molecules. The six model systems were subjected to 200 ns molecular dynamics simulations and trajectories were analyzed to identify stable ligand-receptor interaction features. Based on these interactions, conventional pharmacophore models (CPM) were generated from the individual binding sites while DHPMs were created from hybrid-molecules occupying both binding sites. A huge library of 1,563,764 publicly available molecules were screened by CPMs and DHPMs. The screened hits obtained from both types of models were compared based on their Hashed binary molecular fingerprints and 4-point pharmacophore fingerprints using Tanimoto, Cosine, Dice and Tversky similarity matrices. Molecules screened by DHPM exhibited significant structural diversity, better binding strength and drug like properties as compared to the compounds screened by CPMs indicating the efficiency of DHPM to explore new chemical space for anti-TB drug discovery. The idea of DHPM can be applied for a wide range of mycobacterial or other pathogen targets to venture into unexplored chemical space.

Recently, biogenic toxins have received increasing attention owing to their high contamination levels in feed and food as well as in the environment. However, there is a lack of an integrative platform for seamless linking of data-driven computational methods with \'wet\' experimental validations. To this end, we constructed a novel platform that integrates the technical aspects of toxin biotransformation methods. First, a biogenic toxin database termed ToxinDB (http://www.rxnfinder.org/toxindb/), containing multifaceted data on more than 4836 toxins, was built. Next, more than 8000 biotransformation reaction rules were extracted from over 300,000 biochemical reactions extracted from ~580,000 literature reports curated by more than 100 people over the past decade. Based on these reaction rules, a toxin biotransformation prediction model was constructed. Finally, the global chemical space of biogenic toxins was constructed, comprising ~550,000 toxins and putative toxin metabolites, of which 94.7% of the metabolites have not been previously reported. Additionally, we performed a case study to investigate citrinin metabolism in Trichoderma, and a novel metabolite was identified with the assistance of the biotransformation prediction tool of ToxinDB. This unique integrative platform will assist exploration of the \'dark matter\' of a toxin\'s metabolome and promote the discovery of detoxification enzymes.

Over the last 5 years deep learning has progressed tremendously in both image recognition and natural language processing. Now it is increasingly applied to other data rich fields. In drug discovery, recurrent neural networks (RNNs) have been shown to be an effective method to generate novel chemical structures in the form of SMILES. However, ligands generated by current methods have so far provided relatively low diversity and do not fully cover the whole chemical space occupied by known ligands. Here, we propose a new method (DrugEx) to discover de novo drug-like molecules. DrugEx is an RNN model (generator) trained through reinforcement learning which was integrated with a special exploration strategy. As a case study we applied our method to design ligands against the adenosine A2A receptor. From ChEMBL data, a machine learning model (predictor) was created to predict whether generated molecules are active or not. Based on this predictor as the reward function, the generator was trained by reinforcement learning without any further data. We then compared the performance of our method with two previously published methods, REINVENT and ORGANIC. We found that candidate molecules our model designed, and predicted to be active, had a larger chemical diversity and better covered the chemical space of known ligands compared to the state-of-the-art.

Chemical liabilities, such as adverse effects and toxicity, play a significant role in modern drug discovery process. In silico assessment of chemical liabilities is an important step aimed to reduce costs and animal testing by complementing or replacing in vitro and in vivo experiments. Herein, we propose an approach combining several classification and chemography methods to be able to predict chemical liabilities and to interpret obtained results in the context of impact of structural changes of compounds on their pharmacological profile. To our knowledge for the first time, the supervised extension of Generative Topographic Mapping is proposed as an effective new chemography method. New approach for mapping new data using supervised Isomap without re-building models from the scratch has been proposed. Two approaches for estimation of model\'s applicability domain are used in our study to our knowledge for the first time in chemoinformatics. The structural alerts responsible for the negative characteristics of pharmacological profile of chemical compounds has been found as a result of model interpretation.