What is scvi-tools used for?

scvi-tools is a comprehensive Python framework for probabilistic models in single-cell genomics, used for tasks like dimensionality reduction, batch correction, and multimodal integration.

Why does scvi-tools require raw counts?

The underlying probabilistic models (like VAEs) are designed to model the generative process of the count data directly, and pre-normalization can interfere with their statistical assumptions.

Can this skill help with spatial transcriptomics?

Yes, it supports specialized models like DestVI, Stereoscope, and Tangram for spatial deconvolution, mapping, and cell-environment analysis.

Does this skill work with Scanpy?

Yes, scvi-tools is designed to be AnnData-centric and integrates seamlessly with the Scanpy ecosystem for downstream analysis and visualization.

scvi-tools Assistant

Name: scvi-tools Assistant
Author: Microck

byMicrock

•

データサイエンスとML

Analyzes single-cell omics data using the scvi-tools framework for probabilistic modeling and batch correction.

概要

This skill empowers Claude to guide users through complex single-cell genomics workflows using the scvi-tools Python library. It provides deep domain expertise for implementing generative models for scRNA-seq, scATAC-seq, and spatial transcriptomics, ensuring best practices like using raw counts and proper covariate registration are followed. Whether you are performing batch correction, multimodal integration, or Bayesian differential expression analysis, this skill streamlines the process of building, training, and interpreting PyTorch-based models within the AnnData ecosystem.

主な機能

Probabilistic modeling for scRNA-seq and scATAC-seq
81 GitHub stars
Joint analysis for multimodal integration (CITE-seq, Multiome)
Bayesian differential expression and uncertainty quantification
Automated batch correction and technical variation adjustment
Spatial transcriptomics deconvolution and mapping

ユースケース

Integrating heterogeneous single-cell datasets across different labs or technologies
Predicting cell type labels using semi-supervised learning models like scANVI
Performing spatial deconvolution to map cell types to tissue locations

概要

主な機能

Probabilistic modeling for scRNA-seq and scATAC-seq
81 GitHub stars
Joint analysis for multimodal integration (CITE-seq, Multiome)
Bayesian differential expression and uncertainty quantification
Automated batch correction and technical variation adjustment
Spatial transcriptomics deconvolution and mapping

ユースケース

Integrating heterogeneous single-cell datasets across different labs or technologies
Predicting cell type labels using semi-supervised learning models like scANVI
Performing spatial deconvolution to map cell types to tissue locations