Analyze single-cell omics data with probabilistic models to extract latent structure and harmonize across batches.
npx skills add https://github.com/Ketomihine/my_skills --skill scvitools
Please help me install this Agent Skill. Skill: scvitools Source: https://github.com/Ketomihine/my_skills/tree/main/scvitools Command: npx skills add https://github.com/Ketomihine/my_skills --skill scvitools
scvi-tools provides a unified, probabilistic framework for analyzing single-cell omics data, enabling scalable modeling, integration, and interpretation across diverse modalities.
High-intent search queries and answers about installing and using this skill.
To integrate multimodal single-cell data across batches, load your data as an AnnData object and train a probabilistic model like totalVI or MultiVI. This extracts latent representations to harmonize RNA, protein, and chromatin modalities across samples.
Identifying cell-type-specific markers in scRNA-seq and CITE-seq datasets is best done by training a probabilistic model on your AnnData object, then applying built-in differential expression and differential abundance analyses across conditions and cell types.
Yes, you need PyTorch to run probabilistic single-cell analysis with scvi-tools because it provides the backend for scalable model training. Python and scvi-tools are required, while Ray Tune is optional for autotuning hyperparameters.
You can analyze scATAC-seq data using the same probabilistic framework as scRNA-seq by using the MultiVI model. This supports chromatin modalities within AnnData-compatible objects for integrated cross-dataset benchmarking.
Latent representation extraction for single-cell omics data works by training a chosen probabilistic model like SCVI on an AnnData object. This process learns a harmonized latent structure that can be used for visualization and downstream analyses.
When using probabilistic models for single-cell integration, ensure your data is organized as AnnData-compatible objects. While supporting RNA, protein, and chromatin modalities, the implementation requires Python and PyTorch, making it dependent on specific environment configurations.
Victory-HugoRun scvi-tools deep generative models on single-cell RNA sequencing data.
omicverseIntegrate paired and unpaired single-cell AnnData modalities into harmonized embeddings and joint latent factors.
MesferAliOrchestrate distributed machine learning training across clusters with Ray Train.
fuzzy-dynamicsStreamline probabilistic single-cell analysis workflows with scvi-tools.
