Proximity-guided graph learning reveals tumour-associated proximity antigens.

Proximity-guided graph learning identifies 248 tumour-associated proximity antigens, revealing spatial relationships that enhance targeted cancer therapy.

• Why it matters: Current methods focus mainly on protein expression, overlooking the importance of membrane protein spatial organization in tumor biology and therapeutic targeting. Understanding spatial proximity can uncover novel co-targets and improve precision treatments.
• What they did: Researchers developed an industrialized proximity-mapping workflow and created 248 maps across receptor tyrosine kinases and tumor systems, then built MetaMap to analyze spatial protein communities and identify tumour-associated proximity antigens.
• The result: They validated EGFR-CDCP1 as a TAA-TAPA pair, demonstrating improved tumor cell killing, and established membrane proximity as a key principle for designing more effective multispecific therapeutics.