TattaBio

TattaBio / FlashPPI

Public

Linear-time prediction of proteome-scale microbial protein interactions

19
1
100% credibility
Found Mar 06, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

FlashPPI is a tool and pre-trained model for rapidly predicting which proteins interact within a microbial proteome or between host and pathogen proteomes.

How It Works

1
📖 Discover FlashPPI

You hear about a smart tool that quickly figures out which proteins in microbes work together, from a science paper or online.

2
🧬 Gather your protein list

Collect the sequence of building blocks for all proteins in your organism or virus into one easy file.

3
Pick your prediction path
🌐
Try the web version

Upload your file to the friendly site and instantly explore the web of protein partnerships.

💻
Use your computer

Give your file to the tool and let it crunch the full details in minutes.

4
⚡ Watch it predict

The tool scans everything super fast and lists pairs of proteins likely to interact, with strength scores.

🎉 Uncover protein teams

You get a clear map of interactions, ready to spark new discoveries about how microbes function!

Sign up to see the full architecture

3 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is FlashPPI?

FlashPPI predicts proteome-scale microbial protein interactions in linear time using Python and Hugging Face Transformers. Upload a FASTA proteome via CLI for all-vs-all screening or cross-proteome runs like host-viral, getting a CSV of top interacting pairs with contact scores in minutes on GPU. It outputs interpretable residue-level 2D contact maps, dodging the O(N^2) slog of old-school all-vs-all structural methods.

Why is it gaining traction?

Scales full-proteome predictions from days to minutes with FAISS retrieval and optional Flash Attention, while baked-in genomic priors boost accuracy on microbial data. Users get a web server at seqhub.org for interactive networks, plus easy visualization of contact maps—far snappier than transformer quality in linear time alternatives on GitHub. The contrastive setup flags true cross-species hits reliably.

Who should use this?

Bioinformaticians predicting microbial interactions or host-pathogen PPIs, like viral proteins against bacterial hosts. Suits researchers in flashpoint-style campaigns needing quick proteome scans, or labs exploring protein interactions without massive compute.

Verdict

Promising for GPU users despite 19 stars and 1.0% credibility score—back it with the bioRxiv paper and strong docs, but verify on your data as it's early-stage with no tests. Grab it if proteome PPIs block your workflow.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.