Chemical hazard assessment relies heavily on accurate PBMT data—Persistence, Bioaccumulation, Mobility, and Toxicity. Yet, these critical parameters often exist across fragmented databases, making comprehensive evaluations challenging. Our semester project addressed this gap by developing an automated, integrated PBMT data pipeline within the Laravel framework.
Project Implementation Using Laravel
We developed automated Laravel jobs to interface directly with the EPA CompTox Dashboard via API, automatically retrieving and storing raw JSON data. This robust approach enables data retrieval both manually and on a scheduled basis, significantly improving efficiency.
Once data retrieval completes, another automated process activates to extract relevant PBMT parameters and populate structured tables. Upon completion, our system sends email notifications summarizing newly added, updated, or unchanged records, ensuring transparency.
Figure 1: Entity Relationship Diagram (ERD) showing the PBMT data pipeline
Integrating Janus Predictions
To complement the EPA data, we integrated predictions from the Janus modeling tool, initially provided in Excel format. A preprocessing step harmonized this data with existing EPA data, creating a unified, consistent dataset.
With both sources now integrated, we can clearly see which PBMT parameters are available from each and identify any gaps that future data sources could help fill.
Figure 2: Comparison of current and potential future data sources
Interactive Web Interface
This integrated data now powers an intuitive, searchable web interface. Users can explore PBMT parameters by substance ID, CAS number, or chemical name, clearly distinguishing between experimental and predicted values. Additionally, detailed views are available for users seeking deeper insights into specific parameters and values. On the page for each substance, symbols are displayed to indicate which data sources—such as the EPA Dashboard or Janus—have available information for that particular entry.
Figure 3: Index page showing PBMT data per substance
Expert Validation through Voting
Our system includes expert validation capabilities, enabling specialists to review and vote on chemical classifications, ensuring reliable and accurate hazard assessments. Each substance is first automatically scored based on the type and origin of the data. Expert votes carry the most influence and are used to finalize the classification.
Figure 4: Voting interface for expert classification reviews
Conclusion
By combining data from the EPA Dashboard and Janus, we’ve created a reliable, automated system that brings PBMT values into one place. The use of background tasks, structured storage, and an expert review layer ensures the data is not only accessible but trustworthy.
This project forms a solid foundation for future development. Next steps include integrating additional data sources to improve completeness and expanding our focus beyond PBMT to cover other hazard categories like CMR and ED. With Laravel at its core, the system is built to grow—supporting more features, more data, and ultimately, better chemical safety decisions.