ILTox: A Curated Toxicity Database for Machine Learning and Design of Environmentally Friendly Ionic Liquids

Abstract

A comprehensive online database on the toxicity of ionic liquids (ILs) is urgently needed to facilitate machine learning to design environmentally friendly ILs. In this direction, we present ILTox, a manually curated database of 1183 ILs with over 6700 pieces of toxicity data across different living organisms, including mammalian cells, bacteria, and plants. All the toxicity values and structural information on ILs have been rigorously assessed to ensure data quality. Using this database, various machine learning models have been constructed to quantitatively analyze the relationship between the ILs’ structures and their toxicities. Furthermore, the optimized models were used for a virtual screening of desired properties from 8 million ILs. Our results demonstrated that the ILTox database could accelerate the transformation of toxicity data into critical structure–toxicity relationship knowledge. As far as we know, ILTox is the only available database on IL toxicity and is now openly accessible at http://www.iltox.com.