Xaira-Therapeutics

Xaira-Therapeutics / X-Cell

Public

X-Cell: a diffusion language model for genome-scale perturbation prediction

46
3
100% credibility
Found Mar 25, 2026 at 46 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

X-Cell is a tool for predicting how genetic changes affect gene activity in various cell types using control cell data.

How It Works

1
📰 Discover X-Cell

You stumble upon X-Cell, a smart tool from a biotech company that simulates how tweaking genes changes cell behavior.

2
💻 Set up X-Cell

You easily add the tool to your computer, getting everything ready in moments.

3
🤖 Load the model

You bring in the pre-trained brain that understands millions of real cell experiments.

4
🔬 Prepare your cells

You share measurements from normal cells, setting the stage for your gene simulation.

5
✨ Run a prediction

You name the gene to change, like BRCA1, and watch the tool predict the exciting cell transformations.

6
📊 Review results

You explore the detailed new cell profiles, seeing exactly how everything shifts.

🎉 Gain research insights

You uncover gene effects across cell types without costly lab work, speeding up your discoveries.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 46 to 46 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 X-Cell?

X-Cell is a Python package for predicting genome-scale transcriptional responses to gene perturbations, like CRISPRi knockdowns of BRCA1, starting from control single-cell RNA data in AnnData format. It uses a diffusion language model trained on 25.6 million perturbed cells to simulate cell x gene effects across diverse contexts, outputting predicted expression profiles via simple API calls like model.predict(adata, perturbation="BRCA1"). This tackles the pain of running costly wet-lab screens by enabling fast, virtual perturbation prediction on cell github datasets.

Why is it gaining traction?

It beats baselines by 5x on held-out perturbation accuracy, generalizes zero-shot to new cell types like T cells, and runs its 55M-parameter Mini variant on a single GPU. Developers dig the HuggingFace integration for models and the massive X-Atlas Pisces dataset, plus multi-modal priors from tools like ESM-2 and scGPT for sharper genome-scale predictions. Early Python API previews hook bio devs eyeing diffusion models for perturbation forecasting.

Who should use this?

Computational biologists analyzing Perturb-seq screens or single-cell atlases who need quick what-if simulations for drug targets. Single-cell RNA experts tired of sparse perturbation data in tools like scanpy. Genome researchers at biotechs modeling cell responses without new experiments.

Verdict

Promising alpha from Xaira Therapeutics with solid docs and pyproject setup, but inference code and weights are not ready—star count at 46 and 1.0% credibility score reflect pre-release status. Watch the x-cell github repo for the full drop if genome-scale diffusion perturbation prediction fits your pipeline.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.