Joint localization and testing reveals spatial context of differential expression

SpaceDX jointly tests differential expression and localizes spatial expression patterns in multi-subject mouse brain data, detecting 45.8% more DE genes than regional DESeq2 pseudobulk in chronic stress.

• Why it matters: Spatial transcriptomics often requires predefining regions for subject-level DE, separating localization from testing and obscuring signals when DE does not match predefined spatial boundaries.
• What they did: SpaceDX uses gene-specific attention-based spatial weighting to jointly test DE and infer spatial maps; in simulations it outperforms cluster-based DE when DE misaligns with spatial boundaries.
• The result: In an Alzheimer’s disease model, combining SpaceDX with DESeq2 increased discovery by 19.1% versus DESeq2 alone, and learned maps provided spatial context for condition-associated signals.