We are in the process of building a centralized cheminformatics repository which will include popular toxicological databases (e.g. ACToR, ChEMBL, Chemistry Dashboard Comparative Toxicogenomics Data Base, CTD, EPA Integrated Risk Information System (IRIS), T3DB, TOXNET and TOXCAST) and tools.
To make our resource more powerful, we will develop a customized data analysis and visualization platform so the users can perform some preliminary analysis as well as retrieve ensembled data. For this purpose, we are testing the fracking web hub that lists the chemicals used in the hydraulic fracturing industry with predictions of toxicity. We collected hydraulic fracturing fluid data from Fracfocus and toxicity data from the ChEMBL, TOXNET, ACToR and etc. databases and developed a simple platform for the integration, analysis, and visualization of the data. We are searching for and including comprehensive geographic data, CAS data and toxic endpoints in this system.
In building our cheminformatics platform to predict the toxicity of unknowns we will use machine learning of toxicological big data sets. A critical component will be to conduct validation of the predictive tools on a large library of compounds whose toxic end points are known. This approach fulfills one of the tenets of Tox21st Century.