https://github.com/xnought/hyperprot

Many-to-many protein search and querying with Foldseek
https://github.com/xnought/hyperprot

foldseek graph protein similarity-search

Last synced: 12 days ago
JSON representation

Many-to-many protein search and querying with Foldseek

• Host: GitHub
• URL: https://github.com/xnought/hyperprot
• Owner: xnought
• License: mit
• Created: 2024-09-16T16:55:03.000Z (4 months ago)
• Default Branch: main
• Last Pushed: 2024-09-18T19:51:21.000Z (4 months ago)
• Last Synced: 2024-11-10T22:45:25.734Z (2 months ago)
• Topics: foldseek, graph, protein, similarity-search
• Language: Jupyter Notebook
• Homepage:
• Size: 263 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        


---

Extending protein search for many proteins vs. many other proteins. Visualized with a hypergraph.

## TODO

- [ ] Experiment

     - [x] Use foldseek in python

     - [x] Visualize a graph of the proteins

     - [ ] Select group on click

     - [ ] Query PDB for the clicked group

     - [ ] Parse the output graph with paper, bio data, structure from PDB

     - [ ] Use Graph RAG on the structures and paper to come up with function hypotheses (by Monday)

- [ ] Create user interface for anyone to use this process (By next Friday)

- [ ] Write the paper (By next Monday)

## Development

Install the foldseek executable based on your system from https://github.com/steineggerlab/foldseek directly into the `exec/` folder. Or if you have it installed globally, you can change the executable path config option (TODO). This is just so we can call Foldseek from Python.

## References

- https://github.com/steineggerlab/foldseek

- https://venome.cqls.oregonstate.edu/proteins

- https://github.com/upphiminn/jLouvain