Embed and annotate single-cell expression data from AnnData using scGPT.
npx skills add https://github.com/guanxiaol/open-science --skill scgpt
Please help me install this Agent Skill. Skill: scgpt Source: https://github.com/guanxiaol/open-science/tree/main/runtime/skills/scgpt Command: npx skills add https://github.com/guanxiaol/open-science --skill scgpt
This Skill addresses the challenge of embedding and annotating single-cell expression data with scGPT, a foundation model for single-cell biology.
Run the scgpt skill to produce cell embeddings from your AnnData file.
High-intent search queries and answers about installing and using this skill.
To generate cell embeddings from AnnData for clustering, you use the scGPT foundation model to process single-cell expression data and produce vector representations. This requires scGPT checkpoints and a GPU with CUDA 12.1 or higher.
Zero-shot cell-type annotation is the process of identifying cell types in single-cell expression data without prior task-specific training. The scGPT foundation model applies its pre-trained representations to classify cells directly from an AnnData input.
Yes, you need a GPU with CUDA 12.1 or higher to run scGPT for gene representation tasks. Accessing the scGPT checkpoints also requires this specific hardware setup to process single-cell expression data and obtain gene-level representations.
Yes, you can obtain gene-level representations for perturbation and GRN tasks using scGPT. The model embeds single-cell expression data to provide gene-specific representations suitable for downstream perturbation analysis.
Yes, scGPT supports fine-tuned cell-type annotation for large single-cell datasets. Researchers can fine-tune the foundation model on their specific expression data to improve annotation accuracy for their biological context.
PKU-YuanGroupEmbed and annotate single-cell expression data with the scGPT foundation model.
chenyhvvvvIdentify spatial cellular niches in single-slice spatial transcriptomics data using Harmonics hierarchical modeling.
StarlitnightlyCompare conditions per cell type in AnnData with Wilcoxon, t-test, or memento backends.
pur3v4d3rAnalyze single-cell data with probabilistic models to correct batch effects.
