scvi-tools

To remove batch effects from single-cell RNA-seq datasets, this skill uses deep generative models to perform probabilistic batch correction. It trains on annotated data matrices to identify shared cell states across different donors while eliminating technical variations.

Yes, you can perform multi-modal integration by jointly analyzing RNA, protein, and chromatin accessibility data within a single generative model. This unified probabilistic framework supports CITE-seq and scATAC-seq modalities to provide comprehensive multi-omics analysis.

Probabilistic modeling for single-cell omics uses deep generative models to perform dimensionality reduction and differential expression with built-in uncertainty quantification. It handles high-dimensional data complexity by automatically addressing noise during the integration process.

Yes, you need PyTorch and scanpy installed along with scvi-tools to execute variational inference on annotated data matrices. These dependencies provide the necessary deep learning and single-cell data infrastructure to run the generative models.

The best way to perform differential expression with uncertainty quantification is using a probabilistic framework that models high-dimensional data complexity. This skill provides built-in uncertainty quantification during differential expression analysis of single-cell omics data.

You should use deep generative models when your single-cell analysis requires simultaneous batch correction, multi-modal integration, and explicit uncertainty quantification. This unified probabilistic framework is designed to handle high-dimensional noise and technical batch effects automatically.