# bowang-lab/scGPT

Repository: https://github.com/bowang-lab/scGPT
Canonical: https://ross.abutalabs.com/products/scgpt
Homepage: https://scgpt.readthedocs.io/en/latest/
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: foundation-model, gpt, single-cell
Last push: 2026-04-29T17:41:42+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 79, release rhythm 40, longevity 87
- inputs: {"age_days": 1228, "days_push": 126, "days_rel": 521, "gap_med": 0, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1620, forks 341 (observed 2026-08-28T04:05:12.329002+00:00)

## What it is
scGPT is a foundation model for single-cell multi-omics built with a generative transformer, providing pretrained checkpoints and a Python package for analysis tasks like cell embedding, integration, and reference mapping. It is the official codebase from Bowang Lab with tutorials, documentation, and PyPI distribution.

## Use cases
- analyze single-cell RNA-seq data with a pretrained foundation model
- generate cell embeddings for scRNA-seq datasets
- map query cells to a large reference atlas like CellXGene
- integrate single-cell datasets across batches
- pretrain a transformer model on single-cell multi-omics data
- run zero-shot cell type annotation

## When to choose
- you need state-of-the-art pretrained models for single-cell analysis
- you want reference mapping against millions of cells efficiently
- you work in Python with PyTorch and want notebook tutorials

## When to avoid
- you need bulk RNA-seq or non-single-cell genomics tools
- you lack GPU resources for large model inference or training
- you need a production web service rather than a research library

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-training
- domain: bioinformatics, machine-learning, artificial-intelligence
- platform: python
- tags: single-cell, foundation-model, transformer, genomics, multi-omics, generative-ai, gpu

## Member repositories
- bowang-lab/scGPT (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:12.329002+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:49:16.814397+00:00, confidence not recorded.
  - readme: https://github.com/bowang-lab/scGPT (fetched 2026-08-28T04:05:12.329002+00:00, sha 2d155aee0191)
  - registry_pypi: https://pypi.org/pypi/scgpt/json (fetched 2026-08-29T11:22:22.569676+00:00, sha 79b3631f17ce)
- Data as of 2026-08-30T08:39:29.467469+00:00.