AnnoGCD: a generalized category discovery framework for automatic cell type annotation
2024

AnnoGCD: A Framework for Automatic Cell Type Annotation

publication Evidence: high

Author Information

Author(s): Francesco Ceccarelli, Pietro Liò, Sean B Holden

Primary Institution: University of Cambridge

Hypothesis

Can a semi-supervised and unsupervised approach improve cell type annotation in single-cell RNA sequencing data?

Conclusion

The AnnoGCD framework effectively identifies both known and novel cell types in scRNA-seq data, outperforming existing methods.

Supporting Evidence

  • AnnoGCD demonstrated superior performance in identifying known and novel cell types compared to existing methods.
  • The framework is robust in handling datasets with significant class imbalance.
  • Publicly available datasets were used to validate the effectiveness of the proposed approach.

Takeaway

This study created a new tool that helps scientists automatically label different types of cells from complex data, even when they don't have all the information they need.

Methodology

The study used a semi-supervised block to classify known cell types and an unsupervised block to identify novel cell types from single-cell RNA sequencing data.

Limitations

The framework's performance may depend on the quality and quantity of labeled data available.

Digital Object Identifier (DOI)

10.1093/nargab/lqae166

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication