Reproduce “CellPlex: Benchmarking Single-Cell Spatial Proteomics Prediction from H&E” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Zijun Gao, Yijiang Chen, Xiangde Luo, Pheng-Ann Heng

CellPlex provides a benchmark for predicting single-cell protein levels from H&E slides. It aligns multiplexed protein measurements with nuclei in the same tissue sections and tests models across seven cohorts. With CELLO, H-optimus-0 achieves mean PCC 0.532, versus 0.460 for fine-tuned Eva. The advantage is within-cohort: transfer to other cohorts remains poor, and scores measure agreement with assays, not clinical value.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.