Citation:

Vallanat, B., R. Judson, D. Haggard, L. Taylor, B. Chambers, J. Bundy, I. Shah, J. Rogers, K. Paul-Friedman, J. Harrill, AND L. Everett. Utilization of transcriptomics data in use class based grouping and classification of chemicals in toxicity testing. Society of Toxicology 61st Annual Meeting and ToxExpo 2022, San Diego, CA, March 27 - April 01, 2022.

Impact/Purpose:

Abstract submitted to the Society of Toxicology Meeting March 2022. EPA has adopted high-throughput transcriptomics (HTTr) as a new approach methodology to assess toxicity. This technology has been used to measure in vitro gene expression responses for hundreds of chemicals at multiple concentrations. We are grouping chemicals based on their use class. Initial analysis shows clusters of signatures with similar potency within use classes. Potency of targeted chemicals vs non-targeted chemicals can be evaluated based on these signature clusters. These data can provide supporting evidence for assessing the safety and potency of chemicals that are in daily use/contact, as well as pesticides and industrial chemicals.  

Description:

USEPA is developing non-animal studies known as New Approach Methodologies to assess toxicity. One such approach is high-throughput transcriptomics (HTTr) using the TempO-seq targeted RNA-seq platform. This technology has been used to measure gene expression responses for ~1500 chemicals at 8 different concentrations in MCF-7 cells. Using the R-based DESeq2 tool as part of a larger data processing pipeline, we estimated gene-level fold changes across chemicals and concentrations. We then used a modified implementation of ssGSEA and a customized catalog of pathways chosen primarily from MSigDB, Bioplanet and CMAP to generate signature-level scores. Signatures are directed or undirected gene sets that capture the coordinated response of pathway perturbations. Following this step, concentration-response models for signature scores were performed using a custom developed R-package (tcplFit2). Aggregation of signatures into meaningful collections focused primarily on molecular targets was performed by manually curating across multiple pathway databases. AUC value for these aggregated “super-targets” were calculated at different BMD ranges across all chemicals and examined in the context of use class based chemical grouping.  In our analysis of data from MCF-7 cells we identified clusters of signature aggregates with similar potency within use classes. These analyses can provide supporting evidence for assessing the safety and potency of chemicals that are in daily use/contact, as well as pesticides and industrial chemicals, and contrast them against more targeted chemicals belonging to the pharmaceutical use class. This approach can be utilized in providing supporting data for identifying toxicity of environmental chemicals. This abstract does not necessarily reflect U.S. EPA policy.    

Record Details: