ProST: an image prompt-guided multimodal representation learning framework for spatial domain identification
Chenlan Sun, Zhengxia Wang, Qingchen Zhang, Xiaodong Bai, Chong Yin, Zhipei Sang, Hongsheng Xie, Yuxing Li · 2026
Read ProST: an image prompt-guided multimodal representation learning framework for spatial domain identification by Chenlan Sun, Zhengxia Wang, Qingchen Zhang, Xiaodong Bai, Chong Yin, Zhipei Sang, Hongsheng Xie, Yuxing Li (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13619311.txt.
Description
Sun and colleagues present in 2026 ProST, a computer framework that identifies spatial domains in spatial transcriptomics data. Spatial domains are regions of tissue with coherent biology, and the data map gene activity while keeping each measurement's location in the tissue slice.
Earlier methods chased clustering scores while neglecting whether the regions drawn actually match tissue architecture, and they largely ignored the microscope image sitting beside the gene data. ProST instead uses the stained tissue image as a prompt, guiding the learning toward morphology while filtering out irrelevant visual noise, and it sharpens how image features are extracted. Across several public datasets the method beat existing approaches at drawing domains, visualizing structure, and inferring spatial trajectories. It also filled in missing gene expression values better in the benchmark comparisons the authors report. The code is freely available.
Catalog record
| CREATOR | Chenlan Sun, Zhengxia Wang, Qingchen Zhang, Xiaodong Bai, Chong Yin, Zhipei Sang, Hongsheng Xie, Yuxing Li |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |