Scvi-hub: Pre-trained models to simplify single-cell analysis
- DNA writer
- Sep 12
- 1 min read
Single-cell data analysis typically requires significant processing power and long processing times, especially when working with atlases containing millions of cells. Scvi-hub is an open repository of pre-trained models for single-cell analysis, that integrates with the scvi-tools suite and the scverse community. This approach allows models trained on large atlases, such as the CZI CELLxGENE Discover Census, to be reused and applied directly to new datasets without having to retrain them from scratch.
Scvi-hub not only reduces computing and storage costs but also facilitates common tasks such as cell annotation, data imputation, visualization, and spatial analysis. Any laboratory can benefit from the knowledge encapsulated in large-scale models and adapt them to their own experiments.
Read more:
Comments