SCimilarity is a unifying representation of single cell expression profiles that quantifies similarity between expression states and generalizes to represent new studies without additional training.
This enables a novel cell search capability, which sifts through millions of profiles to find cells similar to a query cell state and allows researchers to quickly and systematically leverage massive public scRNA-seq atlases to learn about a cell state of interest.
Tutorials and API documentation can be found at: https://genentech.github.io/scimilarity/index.html
Pretrained model weights, embeddings, kNN graphs, a single-cell metadata can be downloaded from: https://zenodo.org/records/10685499
A docker container with SCimilarity preinstalled can be pulled from: https://ghcr.io/genentech/scimilarity