Lineage-mixed neurodevelopmental programs in pediatric sarcomas nominate tumor-selective surface antigen pairs.

An AI-guided analysis of a pan-pediatric single-cell atlas (266 patient- and xenograft-derived samples across 22 tumor types) links fusion-driven sarcomas to ectopic neuronal marker expression and tumor-specific mixed surface marker co-expression.

• Why it matters: Pediatric solid tumors often share developmental origins and oncogenic fusions, but translating pan-tumor single-cell gene programs into actionable biology is slow and requires broad expertise.
• What they did: The study built an AI agent framework using off-the-shelf large language models (e.g., Gemini 2.5) to interpret scRNA-seq programs across the atlas, then nominated a surface antigen pair.
• The result: It validated STEAP1/NPTXR as a Ewing sarcoma candidate target pair, supporting AND-gated bispecific ADC and CAR-T strategies for these fusion-driven tumor types.

AI summary of the abstract · may contain errors · not medical advice · read the paper.