AI and the World's Hard Problems · Life sciences · arXiv · 2026-10-21
Spatially Informed and Direction Sensitive Representation Learning for Spatial Transcriptomics
Cellular states are shaped by both gene expression programs and the spatial microenvironment, but existing methods can overlook spatially structured signals or reduce a cell's surroundings to proximity alone, discarding direction. The study introduces Tessera, a spatial representation learning framework that combines spatially informed gene selection with a geometric inductive bias modeling relative cell-cell relationships, remaining translation-invariant while preserving direction. Across held-out tissues and technologies, Tessera outperforms existing spatial foundation models in representation quality and generalizes zero-shot to contexts up to 16 times larger than those seen in training.