Chemometric modeling of acute toxicity of diverse aromatic compounds against Rana japonica

Does the accounting of the local symmetry fragments in quasi-SMILES improve the predictive potential of the QSAR models of toxicity toward tadpoles?

Models of toxicity to tadpoles have been developed as single parameters based on special descriptors which are sums of correlation weights, molecular features, and experimental conditions. This information is presented by quasi-SMILES. Fragments of local symmetry (FLS) are involved in the development of the model and the use of FLS correlation weights improves their predictive potential. In addition, the index of ideality correlation (IIC) and correlation intensity index (CII) are compared. These two potential predictive criteria were tested in models built through Monte Carlo optimization. The CII was more effective than IIC for the models considered here.

QSAR assessment of aquatic toxicity potential of diverse agrochemicals

The fast-increasing number of commercially produced chemicals challenges the experimental ecotoxicity assessment methods, which are costly, time-consuming, and dependent on the sacrifice of animals. In this regard, Quantitative Structure-Property/Activity Relationships (QSPR/QSAR) have led the way in developing ecotoxicity assessment models. In this study, QSAR models have been developed using the pEC50 values of 82 diverse agrochemicals or agro-molecules against a planktonic crustacean Daphnia magna with easily interpretable 2D descriptors. Moreover, a link among octanol-water partition coefficient (KOW), bio-concentration factor (BCF), and critical body residue (CBR) has been addressed, and their imputation for the prediction of the toxicity endpoint (EC50) has been done with an objective of the advanced exploration of several ecotoxicological parameters for toxic chemicals. The developed partial least squares (PLS) models were validated rigorously and proved to be robust, sound, and immensely well-predictive. The final Daphnia toxicity model derived from experimental derived properties along with computed descriptors emerged better in statistical quality and predictivity than those obtained solely from computed descriptors. Additionally, the pEC50 and other important properties (log KOW, log BCF, and log CBR) for a set of external agro-molecules, not employed in model development, were predicted to show the predictive ability of the models.

Computational modeling of aquatic toxicity of polychlorinated naphthalenes (PCNs) employing 2D-QSAR and chemical read-across

Polychlorinated naphthalenes (PCNs) are produced from a variety of industrial sources, and they reach the aquatic ecosystems by the dry-wet deposition from the atmosphere and also by the drainage from the land surfaces. Then the PCNs can be transmitted through the food chain to humans and show toxic effects on different aquatic animals as well as humans. Considering this scenario, it is an obligatory task to explore the toxicity data of PCNs more deeply for the species of an aquatic ecosystem (green algae-Daphnia magna-fish), and to extrapolate those data for humans. But the toxicity data for different aquatic species are quite limited. The laboratory experimentations are complicated and ethically troublesome to fill toxicity data gaps; therefore, different in silico methods (e.g., QSAR, quantitative read-across predictions) are emerging as crucial ways to fill the data gaps and hazard assessments. In the present study, we developed individual toxicity models as well as interspecies models from the 75 PCN toxicity data against three aquatic species (green algae-Daphnia magna-fish) by employing easily interpretable 2D descriptors; these models were validated rigorously employing different globally accepted internal and external validation metrics. Then we interpreted the modelled descriptors mechanistically with the endpoint values for better understanding. And finally, we endeavored to improve the prediction quality in terms of external validation metrics by employing a novel quantitative read-across approach by pooling the descriptors from the developed individual QSAR models.

Unsupervised machine learning, QSAR modelling and web tool development for streamlining the lead identification process of antimalarial flavonoids

Identification of lead compounds with the traditional laboratory approach is expensive and time-consuming. Nowadays, in silico techniques have emerged as a promising approach for lead identification. In this study, we aim to develop robust and predictive 2D-QSAR models to identify lead flavonoids by predicting the IC₅₀ against Plasmodium falciparum. We applied machine learning algorithms (Principal component analysis followed by K-means clustering) and Pearson correlation analysis to select 9 molecular descriptors (MDs) for model building. We selected and validated the three best QSAR models after execution of multiple linear regression (MLR) 100 times with different combinations of MDs. The developed models have fulfilled the five principles for QSAR models as specified by the Organization for Economic Co-operation and Development. The outcome of the study is a reliable and sustainable in silico method of IC₅₀ (Mean ± SD) prediction that will positively impact the antimalarial drug development process by reducing the money and time required to identify potential antimalarial lead compounds from the class of flavonoids. We also developed a web tool (JazQSAR, https://etflin.com/news/4) to offer an easily accessible platform for the developed QSAR models.