Export to Word
(http://www.niehs.nih.gov//portfolio/index.cfm?do=portfolio.grantdetail&&grant_number=R43ES035359&format=word)
| Principal Investigator: Luechtefeld, Thomas | |
|---|---|
| Institute Receiving Award | Insilica, Llc |
| Location | Bethesda, MD |
| Grant Number | R43ES035359 |
| Funding Organization | National Institute of Environmental Health Sciences |
| Award Funding Period | 01 Jan 2024 to 31 Dec 2024 |
| DESCRIPTION (provided by applicant): | Environmental & Health-Related Risk Assessments (EHRRA) are an integral part of new product formulation, as well as product safety stewardship. Unfortunately, the current process for conducting EHRRAs is expensive and slow. According to the EPA report: "FY 2021 Contributions to EPA's Portfolio of Evidence", the cost of conducting a Toxic Substance Control Act (TSCA) Chemical Evaluation is approximately $8.4MM over 3.5 years. Currently, most EHRRAs rely on the manual curation of data via Systematic Literature Review. This process, while integral to the overall accuracy and completeness of EHRRAs, requires hundreds of research hours. Therefore, academia, industry, and government regulators would all likewise benefit from a platform which leverages machine learning to optimize EHRRA related literature and systematic reviews. More specifically, there is currently a strong value proposition in the $534BB global cosmetic industry, $221BB global agrochemical industry, $235BB global home cleaning supplies industry, and $1.5TT global consumer packaged goods industry for tools that optimize SLRs for EHRRAs. To minimize the amount of time required by SLRs for EHRRAs, advances in machine learning (ML) are increasingly helping researchers more quickly identify relevant pieces of information. One major obstacle to the development and use of additional ML tools for EHRRAs is integrating such tools into existing SLR platforms. While an increasing number of these platforms utilize, or permit the utility of, ML models, their complex codebase makes it too cumbersome to deploy new features or integrate ML tools for specific use cases. As a result, risk assessments continue to rely on long, manual literature reviews which, in turn, delays accessibility to new data to make informed chemical and product safety decisions. The 'srvc-EHSR' platform will solve this growing market need though a git-based, remote-deployment capable, application which enables EHRRA researchers to build customizable workflows based on standardized, interchangeable SLR and ML components. The platform will prioritize internal and external interoperability of components to ensure fit-for purpose adaptability. As a result, srvc-EHSR will enable more efficient SLRs and more thorough EHRAAs, thereby decreasing the time-to-market for new products and increasing the overall safety of consumer and industrial chemicals. While the fully commercialized srvc-EHSR will integrate the features above, Phase I will target feasibility for component modularity and interchangeability, ML development, and prototype interface. Development will leverage existing assets, BioBricks.ai and Sysrev, to develop core Phase I srvc-EHSR components as cost-efficiently as possible. A prototype cloud based application, terminal interface, software “packages”, and API will be developed and deployed in a usability study by the Phase I Commercial Partner wherein researchers will use srvc-EHSR to review documents for environmental, chemical, and health data. |
| Science Code(s)/Area of Science(s) |
Primary: 75 - Computational Biology/Computational Methods for Exposure Assessment Secondary: 03 - Carcinogenesis/Cell Transformation |
| Publications | No publications associated with this grant |
| Program Officer | Lingamanaidu Ravichandran |