ToxiPred: A Server for Prediction of Aqueous Toxicity of Small Chemical Molecules in T. Pyriformis
Background: Toxicity Prediction is one of the crucial issues as various industrial chemicals are linked with acute and chronic human diseases like carcinogenicity, mutagenecity. Thus, there is a growing need to risk assessment of these chemicals. Tetrahymena pyriformis is used as a model organism to accessed the environmental fate of a chemical to address the toxicity potential of organic chemicals. Our study is based on large diverse dataset of 1208 compounds taken from an international open competition ICANN09 was organized for aqueous toxicity prediction of chemical molecules against Tetrahymena pyriformis. Results: This study described the development of Quantitative Structure Toxicity Relationship (QSTR) model for the prediction of aqueous toxicity against T. pyriformis. Firstly, model developed on 1002 V-life calculated molecular descriptors shows a R/R 20.874/0.76 with RMSE 0.523. Further, selection of relevant descriptors leads to only 9 descriptors, which shows a performance R/R 2 0.846/0.71 with RMSE 0.574 while on blind dataset 0.756/0.570 with RMSE 0.570 respectively. Second, model developed on CDK based 178 descriptors shows correlation (R) 0.876/0.85, R 2 0.77/0.72 with RMSE 0.518/0.556 on training and blind dataset respectively. Next, model developed on selected 6 descriptors from CDK shows nearly equal performance with R 0.866/0.823, R 2 0.75/0.66 with RMSE 0.541/0.609 on training and blind dataset respectively. Finally, a hybrid model based on selected 17 descriptors from both V-life and CDK shows significant improvement in performance on both training and blind dataset with R 0.89/0.85, R 2 0.79/0.72 with RMSE 0.491/0.557 respectively. It was also observed that Molecular mass (M.W.), and XLogP have very high correlation with toxicity of chemical molecules, it suggests that size and solubility of chemical molecules play major role in toxicity. Our results suggest that it is possible to develop web service for computing toxicity of chemicals using non-commercial software. Conclusions: Our present study demonstrates that performance of a QSTR model depends on the quality/quantity of descriptors as well as on used techniques. Based on these observations, we developed a web server ToxiPred (http://crdd.osdd.net/raghava/toxipred), for environmental risk assessment of small chemical compounds.
Keywords: MACHINE LEARNING TECHNIQUES; PREDICTION; QSAR/QSATR; SVM; TOXICITY; TOXIPRED
Document Type: Research Article
Publication date: March 1, 2014
More about this publication?
- Journal of Translational Toxicology is an international peer-reviewed journal, publishes original research articles, short communications, case reports and reviews dealing with all aspects of toxicology including in vitro, in vivo and mechanistic toxicology, human biomonitoring and environmental risk assessment, biochemical and molecular effects of drugs/toxicants, development of biomarkers for monitoring the drug/toxicant induced injury, in silico approaches for predictive toxicity testing, innovative methods and approaches in risk assessment, sensors, devices and chips for detection of biomolecules, toxicants and contaminants, nanomedicine, nanotoxicology, and much more.
- Editorial Board
- Information for Authors
- Subscribe to this Title
- Aims and Scope
- Ingenta Connect is not responsible for the content or availability of external websites
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content