Researchers have proposed GATE-ST, a gene-aware text-image encoder designed for spatial transcriptomics. The method integrates gene expression profiles with slide-level images, allowing the model to learn from both modalities while retaining morphological context. According to the paper, this joint representation is intended to support analysis of disease mechanisms and the development of treatments.
The approach is presented as a way to make use of spatially resolved gene expression data at scale. Because the abstract notes the method preserves morphological features, the work appears focused on keeping tissue structure visible to the model rather than relying on expression values alone. No evaluation results or comparisons are included in the available abstract, so the claims are limited to the proposed architecture and its intended use.