ComptoxAI: A toolkit for AI research in computational toxicology
Note: We have recently migrated ComptoxAI from Neo4j to Memgraph! If you prefer to use Neo4j, you can still do so by downloading the Neo4j database dump and installing it locally in Neo4j Desktop: Penn+Box Download Link
We are also prototyping a new data model for representing individual people's environmental exposures and the way their bodies respond to those exposures. This will be a larger-scale project developed over several years, but we have an initial prototype using synthetic patient data (from the US CMS SynPUF dataset) available at the following link: oekg.comptox.ai. Note that only a few patients are included in this prototype, and we have only implemented a few types of exposures/observations. To access the database, leave the username and password fields blank and click "Connect".
ComptoxAI provides:
- A graph database for storing and retrieving data used in computational toxicology research (implemented in Neo4j)
- Algorithms for analyzing the structure of data stored in the graph database
- A full-featured OWL ontology for computational toxicology (which is used to structure and query the graph database)
- This website, which includes a blog and showcase of research resulting from ComptoxAI
- (Coming soon:) A gallery of machine learning models for making scientific discoveries from ComptoxAI data
Data in ComptoxAI are integrated from a large number of third-party open access databases. We especially want to thank our collaborators at the US Environmental Protection Agency for their help integrating data from the Adverse Outcome Pathway Database (AOP-DB) and DSSTox resources, which together comprise much of the core functionality of ComptoxAI. To see all of the data sources included in ComptoxAI, check out our About page.