You are here:
A cheminformatics workflow to select representative TSCA chemicals for New Approach Methodology (NAM) screening
Citation:
Patlewicz, G., A. Williams, M. Adams, I. Shah, AND K. Friedman. A cheminformatics workflow to select representative TSCA chemicals for New Approach Methodology (NAM) screening. SOT, Salt Lake City, UT, March 10 - 14, 2024. https://doi.org/10.23645/epacomptox.25395526
Impact/Purpose:
Presentation to the Society of Toxicology (SOT) 63rd Annual Meeting and ToxExpo March 2024
Description:
Background: The Toxic Substances Control Act (TSCA) requires the US Environmental Protection Agency (EPA) to evaluate the hazard and exposure to new and existing chemicals. New chemical notifications are typically data poor and EPA?s Office of Pollution Prevention and Toxics (OPPT) has historically relied upon the use of tools including chemical categories and read-across approaches to fill data gaps. As part of a multi-year, New Chemicals Collaborative Research Program between OPPT and EPA?s Office of Research & Development, opportunities are being explored to leverage New Approach Methods (NAMs). In vitro NAMs will be used to generate mechanistic, hazard, and toxicokinetic data that will be evaluated for their applicability to new chemical assessments. Herein, we describe a cheminformatics categorization workflow developed to identify a set of 200-300 representative candidate case study chemicals for screening from the TSCA non-confidential (non-CBI) active inventory. Methods: Chemicals were first categorized using EPA?s New Chemicals Categories (NCC) to evaluate their scope, coverage, and potential gaps. Chemical classification information generated from the freely available web application ClassyFire was used to categorize all discrete organic structures from the TSCA active inventory into one of 68 primary categories. Large primary categories were subcategorized into smaller categories using hierarchical clustering. The inventory was then profiled in a number of different ways to create filters that would aid in the selection of candidate substances from each terminal category. Informative physicochemical property predictions from OPERA QSAR models, such as melting point, boiling point, and vapour pressure were used to indicate likely physical form and volatility. PubChem was queried to identify the feasibility of finding a vendor from which a given substance could be procured. Analytical method detection amenability predictions for liquid-chromatography mass spectrometry were generated to provide an indication of which chemicals lent themselves to aqueous based-screening. Structural alerts were also used to identify potential explosive or highly reactive substances. The medoid of each primary category was taken as the nominal representative substance. Potential candidate substances were selected on the basis of being structurally similar to the medoid and meeting the other screenability conditions. The set of screenable, representative chemicals were evaluated further to characterize how representative they were of the predicted hazard of the broader TSCA active inventory through the lens of different in silico models to understand predictions for acute oral toxicity and endocrine activity (OPen structure-activity/property Relationship App, OPERA v2.8); structure alerts for carcinogenicity, chromosome damage, mutagenicity in vitro, skin sensitization, cardiotoxicity (Derek Nexus 2.5), and developmental toxicity and genotoxicity (Toxicity Estimation Software Tool, TEST v5.1.2). Results: Approximately half of the non-confidential active inventory with discrete structure representations could be assigned at least one NCC, indicating a need for additional structure-based categories. Using ClassyFire and other descriptors, 180 terminal categories were derived for the active inventory, and based on these categories and a target number of chemicals, a final set of 318 chemicals amenable to aqueous screening will undergo analytical quality control and screening in a range of broad and targeted biological technologies for human health relevant endpoints. The 318 chemicals were representative of the physicochemical and fate properties of the active inventory, but with slightly lower median vapor pressure and slightly higher median melting point, as would be expected for chemicals amenable to aqueous based screening. The final set of chemicals also demonstrated a representative balance of putative hazard profiles...
URLs/Downloads:
DOI: A cheminformatics workflow to select representative TSCA chemicals for New Approach Methodology (NAM) screening
POSTER.PDF (PDF, NA pp, 2294.983 KB, about PDF)