Inferring ecDNA Evolution from Allele Frequency Spectra

Extrachromosomal DNA (ecDNA) may form early in aggressive tumors' development or later during progression, but it is unclear how tumor chronology influences patient outcome. No method exists to discern when ecDNA emerged during tumor evolution using current genome sequencing data. Researchers have now created ecEvol, an analytic method capable of inferring ecDNA emergence time, selection strength, mutation rate, and clonal dominance from bulk whole-genome sequencing data's allele frequency spectra.

EcEvol proves superior to other methods, accurately predicting these parameters across a broad range of values. EcEvol, when applied to untreated primary tumors, reveals that ecDNA usually arises early and experiences strong positive selection, although with considerable variation among cancer types. EcDNAs containing known oncogenes tend to emerge earlier and achieve greater clonal dominance compared to ecDNAs lacking oncogenes.

In advanced, often treated tumors, ecDNA emergence significantly lags behind that in matched primary tumors, aligning with the generation and selection of new ecDNA throughout tumor progression and therapy. Follow-up studies involving longitudinal samples corroborate the ongoing expansion of ecDNA-bearing lineages. Ultimately, earlier ecDNA emergence is linked to worse overall survival.

These findings offer insights into the timeline of ecDNA formation from a single tumor biopsy taken at any point during disease progression, demonstrating that allele-frequency spectra contain ecDNA's evolutionary history and providing a quantitative framework for reconstructing its emergence and expansion.