In vitro cell-line cytotoxicity is widely used in the experimental studies of potential antineoplastic agents and evaluation of safety in drug discovery. In silico estimation of cytotoxicity against hundreds of tumor cell lines and dozens of normal cell lines considerably reduces the time and costs of drug development and the assessment of new pharmaceutical agent perspectives. In 2018, we developed the first freely available web application (CLC-Pred) for the qualitative prediction of cytotoxicity against 278 tumor and 27 normal cell lines based on structural formulas of 59,882 compounds. Here, we present a new version of this web application: CLC-Pred 2.0. It also employs the PASS (Prediction of Activity Spectra for Substance) approach based on substructural atom centric MNA descriptors and a Bayesian algorithm. CLC-Pred 2.0 provides three types of qualitative prediction: (1) cytotoxicity against 391 tumor and 47 normal human cell lines based on ChEMBL and PubChem data (128,545 structures) with a mean accuracy of prediction (AUC), calculated by the leave-one-out (LOO CV) and the 20-fold cross-validation (20F CV) procedures, of 0.925 and 0.923, respectively; (2) cytotoxicity against an NCI60 tumor cell-line panel based on the Developmental Therapeutics Program’s NCI60 data (22,726 structures) with different thresholds of IG₅₀ data (100, 10 and 1 nM) and a mean accuracy of prediction from 0.870 to 0.945 (LOO CV) and from 0.869 to 0.942 (20F CV), respectively; (3) 2170 molecular mechanisms of actions based on ChEMBL and PubChem data (656,011 structures) with a mean accuracy of prediction 0.979 (LOO CV) and 0.978 (20F CV). Therefore, CLC-Pred 2.0 is a significant extension of the capabilities of the initial web application.

Computational Approaches To Identify A Hidden Pharmacological Potential In Large Chemical Libraries

Most of pharmaceutical agents exhibit several or even many biological activities. It is clear that testing even one compound for thousands of biological activities is a practically not feasible task. Therefore, computer-aided prediction is the method-of-the-choice to select the most promising bioassays for particular compounds. Using PASS Online software, we determined the likely anti-inflammatory action of the 13 dithioloquinolinethione derivatives with antimicrobial activities. Chemical similarity search in the Cortellis Drug Discovery Intelligence database did not reveal close structural analogues with anti-inflammatory action. Experimental testing of anti-inflammatory activity of the synthesized compounds in carrageenan-induced inflammation mouse model confirmed the computational predictions. The anti-inflammatory activity of the studied compounds was comparable with or higher than the reference drug Indomethacin. Thus, based on the in silico predictions, novel class of the anti-inflammatory agents was discovered.

Every year, more than 50 new therapeutic agents are approved for medical use worldwide. Despite the increase in the number of biologics approvals, the majority of drugs currently available are chemical entities (de la Torre and Albericio, 2024). The development of small-molecule drugs continues to be favored due to the lower cost of research (Wouters et al., 2020), the wider range of applications, and availability to patients (Makurvet, 2021). However, drug development is a long-term process taking on average 10-15 years, and the cost of bringing a drug to market can be $2.5 billion (Catacutan et al., 2024).The wide application of computer-aided drug design (CADD) methods allowed for a reduction in time and financial costs in the drug discovery process (Athanasiou and Cournia, 2019). Knowledge of existing therapeutics, including their structure, biological activity, physicochemical parameters, etc., is a necessary part of the development and application of CADD methods. Currently, numerous freely accessible curated chemical databases (DBs) containing information about approved drugs are available via the World Wide Web (Masoudi-Sobhanzadeh et al., 2020;Elkashlan et al., 2023).Most of these resources contain data only on medicines approved in the US and Europe. However, up to 30% of medicines are approved for the first time by authorities in other countries (Wu et al., 2021), and this information is often missing from the existing DBs. In July 2024, for example, one could not find records in the widely used resource DrugBank (https://go.drugbank.com) for the anticoronavirus drugs simnotrelvir and leritrelvir, which were approved in China in 2023 (Zhu, 2023;Zheng et al., 2024), despite their relevance to the therapy of COVID-19.The lack of locally authorized drug information in freely available resources does not allow researchers to use this knowledge in the development and application of CADD methods. It significantly constricts the investigated pharmacotherapeutic chemical space and reduces the known structure-activity relationships. Thus, there is an urgent need to create an open-access database containing information on medicines, irrespective of the country of their origin and use.The National Medicine Registers (NMRs) are the most reliable sources of data on pharmaceuticals authorized for therapeutic use. Despite the free availability of NMRs, a search for them is challenging due to the large number of regulatory authorities in the world (World Health Organization, 2024) and the lack of a unified resource providing the appropriate access (Siramshetty et al., 2022). The variety of report presentations and document formats published by regulatory authorities also complicates the collection and aggregation of information on the approved medicines, since it is not possible to apply a uniform processing algorithm to retrieve such data.Taking into account the needs for a single source of aggregated information on worldwide approved medicines, we created the World Wide Approved Drugs (WWAD) database. WWAD contains information collected manually from the NMRs and official documentation from the medicines regulatory authorities (MRAs) of different countries, which are freely available on the Internet.The information collection and handling process for the WWAD database consisted of eight stages (Figure 1). The relational database management system MySQL and the program suite Instant JChem (IJC) from ChemAxon (ChemAxon, 2024a) were used for storing and processing the collected data.First of all, we composed a preliminary set of pharmaceutical substances presented in the existing DBs.Next, we searched the NMRs of the MRAs of different countries, which are freely available on the Internet. Currently, we have found the freely available pharmaceutical registers of 71 countries and regions: Argentina, Armenia, Australia, Austria, Bahrain, Belarus, Belgium, Bhutan, Bosnia and Herzegovina, Brazil, Bulgaria, Cameroon, Canada, China, Croatia, Czech Republic, Denmark, Egypt, Estonia, European Union, Finland, France, Germany, Greece, Hong Kong, Hungary, Iceland, India, Ireland, Israel, Italy, Japan, Kazakhstan, Republic of Korea, Latvia, Lebanon, Liberia, Lithuania, Malaysia, Malta, Mexico, Netherlands, New Zealand, Nigeria, Norway, Pakistan, Philippines, Poland, Portugal, Romania, Russia, Saudi Arabia, Serbia, Singapore, Slovakia, South Africa, Spain, Sweden, Switzerland, Tanzania, Thailand, Turkey, United Arab Emirates, Uganda, Ukraine, United Kingdom, United States, Uruguay, Uzbekistan, Zambia, and Zimbabwe. The preliminary set of low molecular weight pharmaceutical substances approved for medical use was significantly expanded using this data.The lists of approved drug names were extracted from the identified NMRs. Then, the names of vaccines, biologics, traditional medicines, and homeopathic products were deleted from the collected lists because the information about such products is not used in most CADD methods (Fourches et al., 2010;Mansouri et al., 2024).Further, using the MedNet information service provided by the World Health Organization (INN-Mednet, 2024), the drug names from the collected lists were matched with their international nonproprietary names (INN) in seven languages (English, Russian, French, Spanish, Latin, Arabic, and Chinese) and national non-proprietary names (for example, the United States adopted name and British approved name). Thus, we created a unified dictionary of the therapeutics, where drug names were presented by their INN in English. If INN was not available for a compound, we used the shortest widespread name: a well-established (e.g., morphine, codeine), national (e.g., carotegrast methyl), trivial (e.g., acetic acid), or trade name (e.g., dimolegin), etc. The Anatomical Therapeutic Chemical (ATC) classification system codes of the substances were also obtained using the MedNet service.At the next stage, a search for official documents published by the MRAs...

Scientific relevance. Currently, machine learning (ML) methods are widely used in the research and development of new pharmaceuticals. ML methods are particularly important for assessing the safety of pharmacologically active substances early in the research process because such safety assessments significantly reduce the risk of obtaining negative results in the future.Aim. This study aimed to review the main information and prediction resources that can be used for the assessment of the safety of pharmacologically active substances in silico.Discussion. Novel ML methods can identify the most likely molecular targets for a specific compound to interact with, based on structure–activity relationship analysis. In addition, ML methods can be used to search for potential therapeutic and adverse effects, as well as to study acute and specific toxicity, metabolism, and other pharmacodynamic, pharmacokinetic, and toxicological characteristics of investigational substances. Obtained at early stages of research, this information helps to prioritise areas for experimental testing of biological activity, as well as to identify compounds with a low probability of producing adverse and toxic effects. This review describes free online ML-based information and prediction resources for assessing the safety of pharmacologically active substances using their structural formulas. Special attention is paid to the Russian computational products presented on the Way2Drug platform (https://www.way2 d rug.com/dr/).Conclusions. Contemporary approaches to the assessment of pharmacologically active substances in silico based on structure–activity relationship analysis using ML methods provide information about various safety characteristics and allow developers to select the most promising candidates for further in-depth preclinical and clinical studies.