FlashPPI: Linear-time prediction of proteome-scale microbial protein interactions
FlashPPI is a contrastively trained model for protein-protein interaction (PPI) prediction, grounded in residue-level interactions, that enables full-proteome interaction prediction in minutes.
By reframing PPI prediction as a dense retrieval task, FlashPPI circumvents the quadratic computational bottleneck of traditional all-vs-all structural screening.
Scalable: Reduces proteome-wide screening from days/months to minutes.
Genomic Priors: Leverages gLM2 to capture cross-protein, multi-gene co-evolutionary signals.
Web Server
FlashPPI is integrated into seqhub.org. You can upload a FASTA and interactively explore whole-proteome networks and contact maps. Explore an example network here.
The model code and inference scripts in the GitHub repo are licensed under the Apache License 2.0.
The FlashPPI model weights are hosted on Hugging Face and released under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license. The weights are freely available for academic and research purposes.
Citing
If you use FlashPPI or our datasets in your research, please cite:
@article{
doi:10.1073/pnas.2610619123,
author = {Andre Cornman and Matt Tranzillo and Nicolo G. Zulaybar and Imane Bouzit and Yunha Hwang },
title = {Linear-time prediction of proteome-scale microbial protein interactions},
journal = {Proceedings of the National Academy of Sciences},
volume = {123},
number = {25},
pages = {e2610619123},
year = {2026},
doi = {10.1073/pnas.2610619123},
URL = {https://www.pnas.org/doi/abs/10.1073/pnas.2610619123},
eprint = {https://www.pnas.org/doi/pdf/10.1073/pnas.2610619123},
}
