The abstract is read at the publisher; the summary is JClub's.
Multicellular Transformer (MCT) learns context-aware gene embeddings from sets of cells, recovering ligand-receptor interactions and spatial co-localization from scRNA-seq across >4200 tissue samples.
- Why it matters: Most gene-regulation models use single-cell inputs, limiting how well they capture tissue-level coordination, cross-cell signaling, and disease-linked rewiring.
- What they did: MCT is a foundation model trained on >4200 tissue samples spanning more than 30 tissue types, using multicell inputs to learn gene-gene interactions across the tissue environment.
- The result: MCT identifies disease-specific rewiring in systemic lupus erythematosus and COVID-19, predicts population-level perturbation effects in cancer without extra training, and enables in silico discovery of therapeutic targets and tissue-specific dependencies.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.