The abstract is read at the publisher; the summary is JClub's.
SpaAxFormer improves spatial-domain identification by using a spatially constrained, gene-resolved axial transformer that preserves explicit input-gene correspondence and yields higher mean ARI after refinement on Visium.
- Why it matters: Spatial transcriptomics needs models that use neighborhood context without losing which input gene each representation corresponds to, enabling both gene-level prediction and cell-level spatial understanding.
- What they did: The method contextualizes genes within each cell, exchanges corresponding gene states across spatial neighbors, and uses CLS for cell summaries; it is trained with gene-expression reconstruction plus CLS-conditioned reconstruction and contrastive learning, then evaluated on imputation and spatial-domain identification across multiple technologies.
- The result: Imputation is most favorable on DLPFC datasets but mixed on other technologies, while post-hoc spatial refinement consistently improves mean ARI on Visium datasets; adding CLS-conditioned reconstruction and contrastive supervision increases mean clustering ARI in a component-addition analysis.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.