Comparative analysis of spatial transcriptomics datasets can be performed using traditional non-spatial differential gene expression analysis but these approaches can miss cases where a gene has similar gene expression magnitude despite having distinct spatial expression patterns.

STcompare provides orthogonal insights to such differential expression analysis by incorporating spatial information by enabling two differential spatial comparison tests: one based on spatial correlation, and one based on spatial fold change.

Using simulated data, we demonstrate how STcompare provides distinct insights from bulk differential gene expression analysis and robustly controls for false positives even in the presence of spatial autocorrelation common in spatial transcriptomics data.

We also apply STcompare to real spatial transcriptomics data of biological replicates to confirm high spatial correspondence for spatially variable genes as well as identify some genes that maintain spatial patterning but change in magnitude potentially indicative of batch effects.

Finally, we apply STcompare to spatial transcriptomics data of healthy control and diseased acute kidney injury tissue to discover potentially disease-relevant changes in spatial gene expression patterns.