About ComptoxAI

ComptoxAI is a toolkit designed to enable AI and data science research in computational toxicology.

Like many areas in biomedicine, environmental toxicology is facing a deluge of valuable data being generated in a wide array of modalities and levels of biological organization. However, the tools for interacting with these data are - at best - inconsistent, outdated, and fail to use state-of-the-art computational methods.

See also:

Authors

ComptoxAI is developed and maintained by Joseph D. Romano, who is a postdoctoral researcher at the University of Pennsylvania. Yun Hao, a PhD student at the University of Pennsylvania, provides assistance with data preprocessing/curation and analysis.

Other important contributors include:

  • Jason Moore (Cedars-Sinai)

  • Trevor Penning (Penn)

  • Holly Mortensen (US EPA)

  • Jonathan Senn (formerly US EPA)

  • Mei Liang (US EPA)

Citing ComptoxAI

We are preparing two papers for publication which, together, will comprehensively describe ComptoxAI and its current features. Until those are released, please cite us using the following:

Romano et al. (2022) ComptoxAI: A toolkit for AI research in computational toxicology. https://comptox.ai.

BibTeX entry:

@misc{comptoxai,
  title = "ComptoxAI: A toolkit for AI research in computational toxicology",
  author = "Joseph~D.~Romano",
  howpublished = "\url{https://comptox.ai}",
  year = 2021
}

Other Publications

  • Romano JD, Hao Y, & Moore JH. (2022) Improving QSAR Modeling for Predictive Toxicology using Publicly Aggregated Semantic Graph Data and Graph Neural Networks. Pacific Symposium on Biocomputing 27: 187-198.

@inproceedings{romano2021improving,
  title={Improving QSAR Modeling for Predictive Toxicology using Publicly Aggregated Semantic Graph Data and Graph Neural Networks},
  author={Romano, Joseph D and Hao, Yun and Moore, Jason H},
  booktitle={PACIFIC SYMPOSIUM ON BIOCOMPUTING 2022},
  pages={187--198},
  year={2021},
  organization={World Scientific}
}

Source databases

ComptoxAI consists of data integrated from a wide range of third-party, open access databases (many of which are relational databases rather than graph databases). These include:

  • AOP-DB

  • AOP-Wiki

  • Drugbank

  • DSSTox

  • Hetionet

  • NCBI Gene

  • NCBI OMIM

  • PubChem

  • Reactome

  • Tox21

See Source Databases for complete details on each of these.

Contact Us

This wesite is maintained by Joseph D. Romano, PhD. He can be reached via email at:

joseph.romano [at] pennmedicine.upenn.edu

Similar projects can be found at his personal website, or at the Artificial Intelligence Innovation Lab’s home page.

Contributing

If you believe you’ve found a bug, would like to request a new feature, or are interested in contributing to the continued development of ComptoxAI, please see CONTRIBUTING.md on GitHub.

Funding and Acknowledgements

ComptoxAI is supported by grant funding from the US National Institutes of Health, including:

ComptoxAI would also not be possible without essential contributions from researchers at the US Environmental Protection Agency (EPA), including Dr. Holly Mortensen, Jonathan Senn, and Mei Liang, who have contributed essential data from the AOP-DB project. We also would like to acknowledge Daniel Himmelstein’s hetionet resource, which is used to derive many of the graph relationships between different classes of biological entities.