ToxiPred: A Server for Prediction of Aqueous Toxicity of Small Chemical Molecules in T. pyriformis

ToxiPred: A Server for Prediction of Aqueous Toxicity of Small Chemical Molecules in T. pyriformis ToxiPred is a computational server developed for predicting the aqueous toxicity of small chemical molecules against Tetrahymena pyriformis. The tool uses Quantitative Structure Toxicity Relationship models to predict the toxicity endpoint pIGC50, which represents the logarithm of 50% growth inhibitory concentration. ToxiPred supports environmental risk assessment and early-stage screening of small chemical compounds. ToxiPred is a computational server developed for predicting the aqueous toxicity of small chemical molecules against Tetrahymena pyriformis. The tool uses Quantitative Structure Toxicity Relationship models to predict the toxicity endpoint pIGC50, which represents the logarithm of 50% growth inhibitory concentration. ToxiPred supports environmental risk assessment and early-stage screening of small chemical compounds. Web Server: https://webs.iiitd.edu.in/raghava/toxipred/ Citation Mishra, N. K., Singla, D., Agarwal, S., Open Source Drug Discovery Consortium, and Raghava, G. P. S. ToxiPred: A Server for Prediction of Aqueous Toxicity of Small Chemical Molecules in T. pyriformis. Journal of Translational Toxicology, 1, 21-27, 2014. https://doi.org/10.1166/jtt.2014.1005 About the Research Toxicity prediction is an important step in chemical safety assessment because many industrial and environmental chemicals are associated with harmful biological effects. Experimental toxicity testing is costly, time-consuming, and not always feasible for large chemical libraries. Tetrahymena pyriformis is commonly used as a model organism for studying the environmental toxicity of organic chemicals. ToxiPred was developed to predict aqueous toxicity of small chemical molecules against T. pyriformis using computational descriptors and machine learning methods. Data Compilation: The study used a large and diverse dataset of 1208 chemical compounds obtained from the ICANN09 aqueous toxicity prediction competition. The dataset included optimized molecular structures and experimentally reported pIGC50 values. Methodology: ToxiPred uses Quantitative Structure Toxicity Relationship modeling based on molecular descriptors calculated using CDK and V-life software. The models were developed using linear and non-linear statistical methods, including Multiple Linear Regression, Partial Least Squares, and SMO-based machine learning approaches.