Search

Abstract Collecting labeled data for many important tasks in chemoinformatics is time consuming and requires expensive experiments. In recent years, machine learning has been used to learn rich representations of molecules using large scale unlabeled molecular datasets and transfer the knowledge to solve the more challenging tasks with limited datasets. Variational autoencoders are one of the tools that have been proposed to perform the transfer for both chemical property prediction and molecular generation tasks. In this work we propose a simple method to improve chemical property prediction performance of machine learning models by incorporating additional information on correlated molecular descriptors in the representations learned by variational autoencoders. We verify the method on three property prediction tasks. We explore the impact of the number of incorporated descriptors, correlation between the descriptors and the target properties, sizes of the datasets etc. Finally, we show the relation between the performance of property prediction models and the distance between property prediction dataset and the larger unlabeled dataset in the representation space.

Introduction: Plants containing β-sitosterol and oleamide are important for various diseases. So, Dillenia indica, D. obovata, and D. pentagyna were investigated for phytochemicals, cytotoxicity and genotoxicity levels on peripheral blood mononuclear cells (PBMCs) and Hela cells. The protective effect of D. pentagyna extract on a HepG2 cell line was also investigated. Methods: Gas chromatography-mass spectrometry (GC-MS) and high-performance liquid chromatography (HPLC) were used for phytochemical analysis. 3-(4,5-dimethylthiazol-2- yl)-2,5-diphenyltetrazolium bromide) tetrazolium reduction (MTT) and comet assays were performed for toxicity testing and protective effects against DNA oxidative damage. Results: The major components were oleamide and β-sitosterol at 38.464-58.247% and 5.585- 6.887% with concentration and quantity of β-sitosterol at 0.2-0.37 mg/mL and 0.42-0.964 mg/g leaf. The D. indica, D. obovata, and D. pentagyna toxicities on PBMCs showed IC50 values at >430, >430, and 350 µg/mL respectively, with no significant DNA damage (P > 0.05) compared to the negative control group. All plant extracts showed toxic activity on Hela cell with IC50 values at 430 μg/mL), no significant (P > 0.05) DNA damage induction, significantly (P < 0.05) decreased DNA damage level, and tremendous antioxidant effect. Additionally, a combined mixture of all plants in an equal proportion revealed no IC50 value and insignificant DNA damage. Conclusion: All the studied species contained oleamide and β-sitosterol, with toxicity on Hela cells without toxicity on PBMC. The D. pentagyna species showed high antioxidant effects and no toxicity on HepG2.

In silico (quantitative) structure–activity relationship modeling is an approach that provides a fast and cost-effective alternative to assess the genotoxic potential of chemicals. However, one of the limiting factors for model development is the availability of consolidated experimental datasets. In the present study, we collected experimental data on micronuclei in vitro and in vivo, utilizing databases and conducting a PubMed search, aided by text mining using the BioBERT large language model. Chemotype enrichment analysis on the updated datasets was performed to identify enriched substructures. Additionally, chemotypes common for both endpoints were found. Five machine learning models in combination with molecular descriptors, twelve fingerprints and two data balancing techniques were applied to construct individual models. The best-performing individual models were selected for the ensemble construction. The curated final dataset consists of 981 chemicals for micronuclei in vitro and 1309 for mouse micronuclei in vivo, respectively. Out of 18 chemotypes enriched in micronuclei in vitro, only 7 were found to be relevant for in vivo prediction. The ensemble model exhibited high accuracy and sensitivity when applied to an external test set of in vitro data. A good balanced predictive performance was also achieved for the micronucleus in vivo endpoint.

Introduction: Many plant-derived products despite wide usage are not scientifically evaluated for their safety and efficacy. Therefore, in the present study, we aimed to evaluate the cytotoxic and genotoxic activities of Polygonum aviculare L., Equisetum arvense L., Plantago lanceolata L. and Artemisia absinthium L. ethanolic extracts in human white blood cells. Methods: Cell viability was assayed by trypan blue exclusion method, while the genotoxicity was tested by cytokinesis-block micronucleus (CBMN) assay upon cells stimulation with noncytotoxic concentrations of the plant extracts. Results: None of the plant extracts showed high cytotoxic activity. At the same time, only extract of P. lanceolata did not present any mutagenic activity, while E. arvense, P. aviculare and A. absinthium were clearly genotoxic. Conclusion: Caution is advice in the case of long-term use of E. arvense, P. aviculare and A. absinthium herbal medicines by population.

Schiff bases and their metal-complexes are versatile compounds exhibiting a broad range of biological activities and thus actively used in the drug development process. The aim of the present study was the synthesis and characterization of new Schiff bases and their copper (II) complexes, derived from L-tryptophan and isomeric (2-; 3-; 4-) pyridinecarboxaldehydes, as well as the assessment of their toxicity in vitro. The optimal conditions of the Schiff base synthesis resulting in up to 75-85% yield of target products were identified. The structure-activity relationship analysis indicated that the location of the carboxaldehyde group at 2-, 3- or 4-position with regard to nitrogen of the pyridine ring in aldehyde component of the L-tryptophan derivative Schiff bases and corresponding copper complexes essentially change the biological activity of the compounds. The carboxaldehyde group at 2- and 4-positions leads to the higher cytotoxic activity, than that of at 3-position, and the presence of the copper in the complexes increases the cytotoxicity. Based on toxicity classification data, the compounds with non-toxic profile were identified, which can be used as new entities in the drug development process using Schiff base scaffold.