Deconstructing evolutionary histories of complex genomic rearrangements in lung malignancies

Complex genomic rearrangements (CGRs) are hallmarks of tumor genomes, yet their architectures and origins remain poorly understood. We utilized optical mapping to profile the genomic landscapes of lung tissues across various stages - from normal and benign lesions to advanced carcinomas - to characterize genome disorganization during carcinogenesis. We developed omcplR to resolve the CGR architectures and predict their genesis using the genome-graph framework. Our findings reveal that many CGRs arise from hierarchical combinations of localized structural variations that are unlikely due to break-fusion-bridge, chromothripsis, or chromoplexy. We termed this process chromodiadochi: a cascade of nested rearrangements accumulating over multiple generations during carcinogenesis. Early intra-chromosomal junctions in CGRs typically localized within specific chromatin domains, while late junctions in advanced tumors occurred in more heterogeneous contexts and showed signs of positive selection. A composite signature of genomic alterations, including the CGRs, captures the trajectory of progressive genomic structural disorganization and instability from benign tissue to malignancy.