‹ back to 2026-09-11

Proximity-guided graph learning reveals tumour-associated proximity antigens

Nature Medical/Healthcare AI Cody Scandore, Clare F. Malone, Christopher K. May, Anna K. de Regt, Jeff Guernsey, Hayley Ma, Noah Dephoure, Ben Setter, Rebecca A. Howell, Kendall R. Johnson, Carol L. Farr, Sophia Romero, Lydia Vignale, Tali Vittum, Emma Dawson, Tsadik Habtetsion, Francesca Nardi, Brian Woodruff, Martin Mathay, Julia Swanson, Mikaela Rusnak, Quynh Ton, Payam E. Farahani, Robert W. Gene, Jason Misurelli, Zach Caldwell, Hengyu Xu, Michael Hornsby, Marc A. Gavin, Heath E. Klock, Ertan Eryilmaz, Pamela M. Holland, Scott A. Lesley, Rob C. Oslund, Olugbeminiyi O. Fadeyi 2026-09-09

TL;DR - A proximity-guided graph-learning atlas maps spatially associated membrane antigens in tumours and identifies EGFR–CDCP1 as a promising co-target pair. This could improve tumour killing by enabling multispecific therapeutics to exploit disease-associated antigen communities.

Defines tumour-associated proximity antigens through proximity mapping and graph learning.
Reveals disease-linked spatial communities of membrane proteins.
Identifies EGFR and CDCP1 as a co-target pair for multispecific therapeutic strategies.
Reports enhanced tumour killing from joint targeting, though the provided summary does not include quantitative results.

view merged work →