You are here:
Inverting the SAR paradigm: Applications of a Chemotype-Enrichment Approach within EPA’s Computational Toxicology Programs (QSAR 2021)
Citation:
Richard, A., R. Lougee, Chris Grulke, J. Wang, Mahmoud A. Shobair, G. Patlewicz, AND C. Yang. Inverting the SAR paradigm: Applications of a Chemotype-Enrichment Approach within EPA’s Computational Toxicology Programs (QSAR 2021). QSAR 2021 International Workshop on QSAR in Environmental and Health Sciences June 2021, Virtual, NC, June 07 - 10, 2021. https://doi.org/10.23645/epacomptox.15106026
Impact/Purpose:
Presentation to the QSAR 2021 International Workshop on QSAR in Environmental and Health Sciences June 2021. Chemotyping, the annotation of observed activities at a fragment, substructure, or scaffold level, provides a way to abstract chemical detail from the broader activity signature that is correlated with a particular fragment. This approach is an effective profiling mechanism and allows navigation of complex data landscapes, such as ToxCast, to understand relationships between an activity and fragment to develop a weight¿of¿evidence regarding a particular chemical and its potential liabilities. Applying and refining the approach towards an array of datasets and research applications will inform the development of standardized workflows and tools. The results of these efforts will be of direct benefit to program and regional offices as well as the greater scientific community in expanding the utility of NAMs through structure¿based concepts.
Description:
Traditional structure-activity relationship (SAR) approaches to modeling in toxicology are determined and bounded by the chemical-activity test set, the chemical descriptors used, and the modeling method employed, resulting in models with limited applicability or comparability outside of the test set and its bounding conditions. Thus, SAR models in toxicology are effectively siloed into a multitude of separate, non-overlapping applications and fail to effectively build on each other’s limited successes. Overcoming the profound data limitations, structural diversity, and mechanistic complexity challenges in this field requires new ways of thinking. The “Comparative QSAR” approach, promoted by the late Corwin Hansch, extracted mechanistic insights by examining patterns of regression coefficients across thousands of fixed-format, linear-regression QSAR models of biological potencies. In effect, diverse QSAR models were compared by projecting them onto an aggregation layer comprised of a small “basis set” of Hansch/Free-Wilson type QSAR descriptors (s.a., logP). A somewhat analogous chemotype-enrichment approach is being applied to a wide variety of binary activity datasets and problems in EPA’s ToxCast, Tox21, and computational toxicology research programs. The approach uses a standardized enrichment algorithm and a fixed “basis set” of ToxPrint chemical features to create an aggregation layer on which chemical enrichments across diverse activity datasets can be projected and compared. A series of examples will be presented illustrating the power of the approach to provide chemically intuitive results whereby weak chemical-activity signals are amplified within local chemical domains, and global patterns are extracted from diverse activity datasets. Abstract does not represent EPA policy.
URLs/Downloads:
DOI: Inverting the SAR paradigm: Applications of a Chemotype-Enrichment Approach within EPA’s Computational Toxicology Programs (QSAR 2021)
QSAR2021_CTEW-PLENARY_ARICHARD_V6_DC.PDF (PDF, NA pp, 6915.05 KB, about PDF)