CNVSIM: an interpretable copy-number reference atlas for cancer subtyping

CNVSIM builds cancer root-specific CNV reference spaces and defines 112 CNV-defined subtypes across 19 cancer roots, enabling calibrated projection of 85.7% held-out profiles.

• Why it matters: Interpreting new genome-wide CNV profiles against a reusable reference is difficult, limiting consistent subtype comparison beyond existing diagnostic labels.
• What they did: The method organizes CNV profiles into root-specific reference spaces with chromosome-scale and gene-level signatures, then applies calibrated projection on the Progenetix dataset using 155,881 neoplastic-sample analyses.
• The result: Within each root, broad diagnoses split into distinct CNV subtypes, while individual subtypes span multiple diagnostic labels; projection of 4,605 cell-line profiles yields representatives for 90/112 subtypes and reveals representation gaps.

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