PuSH - Publication Server of Helmholtz Zentrum München

Engelmann, J.P. ; Hetzel, L. ; Palla, G. ; Sikkema, L. ; Luecken, M. ; Theis, F.J.

Uncertainty Quantification for Atlas-Level Cell Type Transfer.

In: (ICML Workshop on Computational Biology). 2022.
Publ. Version/Full Text
Single-cell reference atlases are large-scale, cell- level maps that capture cellular heterogeneity within an organ using single cell genomics. Given their size and cellular diversity, these atlases serve as high-quality training data for the transfer of cell type labels to new datasets. Such label transfer, however, must be robust to domain shifts in gene expression due to measurement technique, lab specifics and more general batch effects. This re- quires methods that provide uncertainty estimates on the cell type predictions to ensure correct in- terpretation. Here, for the first time, we introduce uncertainty quantification methods for cell type classification on single-cell reference atlases. We benchmark four model classes and show that cur- rently used models lack calibration, robustness, and actionable uncertainty scores. Furthermore, we demonstrate how models that quantify uncer- tainty are better suited to detect unseen cell types in the setting of atlas-level cell type transfer.
Additional Metrics?
[➜Log in]
Edit extra informations Login
Publication type Article: Conference contribution
Conference Title ICML Workshop on Computational Biology
Institute(s) Institute of Computational Biology (ICB)