scvi-tools

Batch correction for single-cell RNA-seq data is handled using deep generative models like scVI to correct technical variation and extract reliable biological signals. You can run the scVI model directly on your h5ad data file to resolve batch effects.

Multi-modal integration for single-cell omics data supports combining diverse data types like RNA-seq, ATAC-seq, and spatial transcriptomics. Deep learning models map these different modalities together to correct technical variation and extract unified biological signals.

Deep learning-based differential expression analysis requires Python, PyTorch, and PyTorch Lightning to run. This Skill is designed for advanced users familiar with these frameworks to perform deep generative modeling on high-throughput omics data.

Semi-supervised annotation of single-cell types is best performed using the scANVI model, an advanced deep learning approach. It extends unsupervised scVI models by leveraging labeled data to accurately classify cell types and correct technical variation.

Deep learning batch correction is ideal for high-throughput single-cell omics studies where technical variation obscures biological signals. Models like scVI handle complex batch effects across RNA-seq, ATAC-seq, and spatial transcriptomics better than traditional methods.

RNA velocity analysis for single-cell transcriptomics uses the VeloVI deep learning model to infer the splicing dynamics of cells. This unsupervised approach helps predict the future state of cells based on high-throughput RNA-seq data.