Inverse Game Theory Characterizes Frequency-Dependent Selection Driven by Karyotypic Diversity in Triple Negative Breast Cancer.

Chromosomal instability (CIN), characterized by pervasive copy number alterations (CNAs), significantly contributes to cancer progression and therapeutic resistance. CNAs drive intratumoral genetic heterogeneity, creating distinct subpopulations whose interactions shape tumor evolution through frequency-dependent selection. Here, we introduce, ECO-K (Ecological-Karyotypes), an inverse game theory framework that infers subpopulation interactions from longitudinal single-cell whole genome sequencing data. Applying this approach to serially-passaged, triple-negative breast cancer (TNBC) cell lines and patient-derived xenografts (PDXs), we systematically identified frequency-dependent selection dynamics governed by karyotypic diversity. Notably, in one PDX lineage, we found consistent evidence of karyotypically defined subpopulations acting as interaction hubs, associated specifically with chromosome 1 loss and chromosome 14p gain. Our framework provides testable predictions of intratumoral ecological dynamics, highlighting opportunities to strategically target key subpopulations to disrupt tumor evolution.