scvi-scanvi

What problem does it solve?

scANVI enables semi-supervised label transfer for single-cell RNA-seq data by initializing from a trained scVI model and propagating seed labels to new datasets, enabling more accurate and conserved annotations.

Core Features & Use Cases

• Seed-label propagation: seed labels guide annotation transfer to new cells/datasets.
• Model-driven transfer: leverage a trained scVI model to preserve biological consistency across experiments.
• Downstream outputs: obtain predicted labels and latent representations suitable for clustering and visualization.

Quick Start

Provide a trained scVI model and seed labels, then run a scANVI training to transfer cell-type annotations.