Objective
This project will address critical collection and dissemination of per- and polyfluoroalkyl substance (PFAS) transformation pathways through enviPath via collaboration with enviPath’s development team (University of Zurich and Swiss Federal Institute of Aquatic Science and Technology). The intent of the project is to more rapidly and effectively disseminate the results of previous and ongoing SERDP research through enviPath, which itself is used as an input source for additional chemical transformation programs such as the U.S. Environmental Protection Agency’s Chemical Transformation Simulator. Specific objectives include:
Consolidate and characterize existing PFAS biotransformation data in enviPath and add new data
Develop PFAS-specific biotransformation algorithms in enviPath
Engage in technology transfer efforts to ensure broader use and adoption of the final work products
Technology Description
The project tasks build from work conducted under SERDP project ER20-1375 by expanding enviPath curation efforts and development specific to polyfluorinated substance transformation. Moreover, though different in intent, this work dovetails with work from SERDP Project ER23-3697. enviPath can store information on biotransformation rates and pathways, along with extensive metadata on study conditions, allowing for a thorough mechanistic understanding and prediction of how biotransformation pathways and rates depend on experimental and environmental conditions. enviPath is currently the only fully openly accessible, internationally available resource to collect and store such information and to make it available to interested parties, including environmental scientists and engineers, public authorities, and the chemical industry.
Different research groups have started compiling biotransformation data for PFAS on enviPath in separate data packages. As a first task for this project, the different packages will be merged into one PFAS package to avoid duplication and maximize the volume of data available. The project team will also add new data from recently published biotransformation studies. Working with the data collected under Task 1, the project team will retrain the pathway prediction algorithms in enviPath to optimally apply to polyfluorinated substances. PFAS-specific transformation rules will be trained from the data for those reactions that are not covered by enviPath’s existing transformation rules. Then the machine learning models used for prioritizing predicted reactions to avoid combinatorial explosion will also be re-trained on the PFAS data, using the complete set of old and new PFAS-specific transformation rules.
Benefits
This demonstration is expected to provide valuable and novel information regarding PFAS transformation pathways through enviPath. Having all data collected in one place, annotated with necessary metadata on study conditions and reliability scores for the observed transformation products, will allow for a thorough evaluation of known and predicted enzymatic transformations of PFAS (particularly those relevant to aqueous film-forming foam-impacted sites) and their dependence on environmental conditions such as redox status. This project is also expected to enable the optimization of existing biotransformation pathway prediction algorithms for specific polyfluoroalkyl structures. Such information is critical for predicting PFAS transformation pathways, and for developing appropriate conceptual site models. (Anticipated Project Completion - 2024)