CellPatch: a flexible and efficient framework for single-cell foundation models empowered by heuristic gene patching
Abstract Foundation models have advanced single-cell omics analysis but often remain too complex for practical use. We present CellPatch, a lightweight foundation model combining cross-attention with heuristic gene patching to reduce computational complexity during pretraining and fine-tuning while enabling inspection of gene-attention patterns. Across single-cell RNA-sequencing datasets spanning multiple organs and tissue states, CellPatch achieves state-of-the-art performance on diverse prediction tasks. Its modular design supports linear, adapter-based and task-specific architectures, enabling CellPatch to be adapted to diverse biological tasks.
Authors
- Senxin Zhang (ORCID: https://orcid.org/0000-0002-3006-7340)
- Nana Wei (ORCID: https://orcid.org/0000-0002-8039-9626)
- Futing Lai
- Kangwen Cai
- Yujuan Wang (ORCID: https://orcid.org/0000-0002-0404-3599)
- Sun Yuanchen
- Wen-Jie Jiang
- An Liu
- Jiyuan Yang
- Hanwen Zhu
- Yushun Yuan
- Hua-Jun Wu
- Lu Wang
- Xiaoqi Zheng
- Zeyu Ma
Institutions
- Shanghai Jiao Tong University (CN)
- Peking University (CN)
- Shanghai Normal University (CN)
- Peking University Cancer Hospital (CN)
- Peking University Third Hospital (CN)
- Shanghai Key Laboratory of Computer Software Testing and Evaluating (CN)
- Shanghai Development Center of Computer Software Technology (CN)
Publication Details
- Journal
- Genome biology
- Published
- 2026-10-30
- DOI
- https://doi.org/10.1186/s13059-026-04329-y
- Primary Topic
- Single-cell and spatial transcriptomics
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China
- Fundamental Research Funds for the Central Universities