Spatially Informed and Direction Sensitive Representation Learning for Spatial Transcriptomics
Teaching AI to read gene activity by understanding where cells sit in tissue
Cells behave differently depending on their neighbors and position in tissue, but existing methods for analyzing this spatial context miss important details or rely on arbitrary coordinates. A new framework called Tessera captures both which genes are active and how cells are organized relative to each other—encoding direction and local relationships rather than just distance—and outperforms current methods even when applied to tissues it has never seen before.
Spatial transcriptomics is becoming routine in medical research, but current tools lose information about tissue organization that shapes whether cells become healthy or diseased. Better representation learning means researchers can more accurately predict cell behavior, identify disease mechanisms tied to tissue structure, and potentially design better treatments. Tessera also generalizes to much larger tissue samples than it was trained on, making it practical for real research workflows without constant retraining.