Reproduce “SEPAL: Spatial Gene Expression Prediction from Local Graphs” 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
Gabriel Mejia, Paula Cárdenas, Daniela Ruiz, Angela Castillo, Pablo Arbeláez
SEPAL estimates gene-expression patterns from H&E images by predicting each patch’s deviation from the training-set mean, then refining it with a graph of nearby patches. On a Visium held-out slide, its MSE was 0.708, versus 0.741 for EGGN. On STNet, its gene-wise Pearson correlation was −0.004, below EGN’s 0.064, limiting evidence for reliable gene-map prediction across datasets.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.