Reproduce “scCBGM: Interpretable Single-Cell Counterfactual Editing” 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
Alma Andersson, Aya Abdelsalam Ismail, Edward De Brouwer, Doron Haviv, Tommaso Biancalani, Kyunghyun Cho, Gabriele Scalia, Aïcha BenTaieb and 1 more
scCBGM predicts counterfactual gene-expression profiles by separating named concepts from residual cell variation, then changing selected concepts while retaining the rest. In a held-out TCDD stellate-cell test, edited low-dose controls shifted toward high-dose responders. On the Kang PBMC IFN-β task, CD4-group rMMD was 0.156 ± 0.066, versus 0.487 ± 0.098 for Vanilla-FM; these are population-level scores, not individual-cell accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.