The Hidden Mutational Landscape of Extrachromosomal Circular DNA

Abstract

The human genome is not confined to chromosomes. Extrachromosomal circular DNA—eccDNA and its megabase-scale counterpart ecDNA—is a persistent, transcriptionally active, and mutationally accelerated compartment present in normal and neoplastic cells. Yet every stage of genomic analysis, from DNA extraction to variant annotation, is engineered for linear, diploid templates. This review argues that circular DNA mutations are not rare events; they are systematically erased by methodologies that mistake the map for the territory. We synthesize evidence that circular DNA exhibits a mutational spectrum fundamentally distinct from chromosomal DNA: microhomology-driven indels, complex structural rearrangements within single molecules, and interchromosomal chimeric fusions that generate novel oncogenic reading frames. Its mutation rate is accelerated by rolling-circle replication, suppressed homologous recombination, and torsional stress—quantified as μ_eff = μ_bp × CN × f_rep, exceeding chromosomal rates by two orders of magnitude. Experimental isolation and detection methods each carry specific biases that distort mutational discovery. Computational reconstruction requires graph-based assembly and copy-number-normalized burden metrics. Validation by inverse PCR, mutation-specific FISH, and single-cell sequencing reveals that ecDNA mutations exhibit non-Mendelian inheritance yet drive transient transcriptional dominance. Major barriers include the absence of a circular reference genome, indistinguishable replication errors, and zero clinical standardization. Nevertheless, ecDNA amplifications occur in 14% of cancers, confer prognosis independent of chromosomal status, and serve as privileged sites for resistance mutations. eccDNA burden tracks cellular age and genotoxic stress. Emerging therapies targeting ecDNA segregation or transcription are plausible but lack specificity. The central conclusion is not that circular DNA matters—it is that the genome is not what we thought it was.