You are here:
Developing an Integrated Model Management Solution to Assure Quality of Predicted Data at the US EPA’s National Center of Computational Toxicology
Citation:
Grulke, C., A. Williams, A. Singh, AND J. Edwards. Developing an Integrated Model Management Solution to Assure Quality of Predicted Data at the US EPA’s National Center of Computational Toxicology. Presented at American Chemical Society Spring Meeting, Orlando, FL, March 31 - April 04, 2019. https://doi.org/10.23645/epacomptox.8089379
Impact/Purpose:
Presentation at the 2019 Spring meeting of the American Chemical Society. The Computational Toxicology Program within the U.S. Environmental Protection Agency (EPA) integrates advances in biology, chemistry, and computer science to help prioritize chemicals for further research based on potential human and environmental health risks. A key component of this prioritization effort is the use of New Approach Methods (NAMs) which include both in vitro and in silico methods of estimating more traditional risk-assessment inputs (e.g. in vivo toxicity).
Description:
The Computational Toxicology Program within the U.S. Environmental Protection Agency (EPA) integrates advances in biology, chemistry, and computer science to help prioritize chemicals for further research based on potential human and environmental health risks. A key component of this prioritization effort is the use of New Approach Methods (NAMs) which include both in vitro and in silico methods of estimating more traditional risk-assessment inputs (e.g. in vivo toxicity). Without NAMs, the number of chemicals is too large and the available data too sparse to effectively prioritize. However, the application of NAM generated data within EPA to support a regulatory context requires a high degree of quality assurance. While this need is apparent for in silico NAMs (e.g. QSAR models), it is also vital for most in vitro NAMs which use computational tools to process screening results into informative scores. To meet minimal requirements, well-defined workflows must be applied to versioned and well-documented computational methods to yield reproducible results. In addition, the management solution must be flexible enough to efficiently integrate newly developed models from our cheminformatics researchers that could inform a prioritization task without limiting the options afforded to our researchers. To meet these requirements, we have designed a series of loosely coupled microservices that are genericized to enable many of our use-cases but specialized enough to efficiently support our most pressing need of supplying predictions for the Comptox Chemicals Dashboard (https://comptox.epa.gov/dashboard). This abstract does not necessarily represent the views or policies of the U.S. Environmental Protection Agency.
URLs/Downloads:
DOI: Developing an Integrated Model Management Solution to Assure Quality of Predicted Data at the US EPA’s National Center of Computational Toxicology
GRULKE_ACS_MODELING_20190329.PDF (PDF, NA pp, 2849.661 KB, about PDF)