scSSL-Bench: Benchmarking Self-Supervised Learning for Single-Cell Data

Olga Ovcharenko, Florian Barkmann, Philip Toma, Imant Daunhawer, Julia E Vogt, Sebastian Schelter, Valentina Boeva
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:47416-47442, 2025.

Abstract

Self-supervised learning (SSL) has proven to be a powerful approach for extracting biologically meaningful representations from single-cell data. To advance our understanding of SSL methods applied to single-cell data, we present scSSL-Bench, a comprehensive benchmark that evaluates nineteen SSL methods. Our evaluation spans nine datasets and focuses on three common downstream tasks: batch correction, cell type annotation, and missing modality prediction. Furthermore, we systematically assess various data augmentation strategies. Our analysis reveals task-specific trade-offs: the specialized single-cell frameworks, scVI, CLAIRE, and the finetuned scGPT excel at uni-modal batch correction, while generic SSL methods, such as VICReg and SimCLR, demonstrate superior performance in cell typing and multi-modal data integration. Random masking emerges as the most effective augmentation technique across all tasks, surpassing domain-specific augmentations. Notably, our results indicate the need for a specialized single-cell multi-modal data integration framework. scSSL-Bench provides a standardized evaluation platform and concrete recommendations for applying SSL to single-cell analysis, advancing the convergence of deep learning and single-cell genomics.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-ovcharenko25a, title = {sc{SSL}-Bench: Benchmarking Self-Supervised Learning for Single-Cell Data}, author = {Ovcharenko, Olga and Barkmann, Florian and Toma, Philip and Daunhawer, Imant and Vogt, Julia E and Schelter, Sebastian and Boeva, Valentina}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {47416--47442}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/ovcharenko25a/ovcharenko25a.pdf}, url = {https://proceedings.mlr.press/v267/ovcharenko25a.html}, abstract = {Self-supervised learning (SSL) has proven to be a powerful approach for extracting biologically meaningful representations from single-cell data. To advance our understanding of SSL methods applied to single-cell data, we present scSSL-Bench, a comprehensive benchmark that evaluates nineteen SSL methods. Our evaluation spans nine datasets and focuses on three common downstream tasks: batch correction, cell type annotation, and missing modality prediction. Furthermore, we systematically assess various data augmentation strategies. Our analysis reveals task-specific trade-offs: the specialized single-cell frameworks, scVI, CLAIRE, and the finetuned scGPT excel at uni-modal batch correction, while generic SSL methods, such as VICReg and SimCLR, demonstrate superior performance in cell typing and multi-modal data integration. Random masking emerges as the most effective augmentation technique across all tasks, surpassing domain-specific augmentations. Notably, our results indicate the need for a specialized single-cell multi-modal data integration framework. scSSL-Bench provides a standardized evaluation platform and concrete recommendations for applying SSL to single-cell analysis, advancing the convergence of deep learning and single-cell genomics.} }
Endnote
%0 Conference Paper %T scSSL-Bench: Benchmarking Self-Supervised Learning for Single-Cell Data %A Olga Ovcharenko %A Florian Barkmann %A Philip Toma %A Imant Daunhawer %A Julia E Vogt %A Sebastian Schelter %A Valentina Boeva %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-ovcharenko25a %I PMLR %P 47416--47442 %U https://proceedings.mlr.press/v267/ovcharenko25a.html %V 267 %X Self-supervised learning (SSL) has proven to be a powerful approach for extracting biologically meaningful representations from single-cell data. To advance our understanding of SSL methods applied to single-cell data, we present scSSL-Bench, a comprehensive benchmark that evaluates nineteen SSL methods. Our evaluation spans nine datasets and focuses on three common downstream tasks: batch correction, cell type annotation, and missing modality prediction. Furthermore, we systematically assess various data augmentation strategies. Our analysis reveals task-specific trade-offs: the specialized single-cell frameworks, scVI, CLAIRE, and the finetuned scGPT excel at uni-modal batch correction, while generic SSL methods, such as VICReg and SimCLR, demonstrate superior performance in cell typing and multi-modal data integration. Random masking emerges as the most effective augmentation technique across all tasks, surpassing domain-specific augmentations. Notably, our results indicate the need for a specialized single-cell multi-modal data integration framework. scSSL-Bench provides a standardized evaluation platform and concrete recommendations for applying SSL to single-cell analysis, advancing the convergence of deep learning and single-cell genomics.
APA
Ovcharenko, O., Barkmann, F., Toma, P., Daunhawer, I., Vogt, J.E., Schelter, S. & Boeva, V.. (2025). scSSL-Bench: Benchmarking Self-Supervised Learning for Single-Cell Data. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:47416-47442 Available from https://proceedings.mlr.press/v267/ovcharenko25a.html.

Related Material