Jumping DNA Fragments Drive Genomic Chaos in Early Cancer, Not Just Late-Stage Tumors

Inside the human genome lie the remnants of ancient genetic passengers - fragments of DNA that have been copying and reinserting themselves into new chromosomal locations for hundreds of millions of years. Most of the roughly 500,000 copies of LINE-1 (L1) elements in the human genome are dormant fossils. But each person carries between 150 and 200 L1 elements still capable of jumping, and in certain cancers, that activity accelerates dramatically.

For years, L1 retrotransposition was viewed primarily as a source of local disruption - the occasional gene knocked out when a copy lands in the wrong place. A study published in Science, led by Professor Jose Tubio at the Universidade de Santiago de Compostela and conducted in collaboration with the Centre for Genomic Regulation in Barcelona, the Francis Crick Institute, and MD Anderson Cancer Center, reveals a substantially more disruptive picture. L1 elements do not just interrupt genes. They also drive large-scale architectural rearrangements of chromosomes - and they do so predominantly in the early stages of tumor formation, not the late stages as previously assumed.

What Short-Read Sequencing Was Missing

The story changed with technology. Standard DNA sequencing methods, which read DNA in short fragments, can detect L1 insertions reasonably well. What they cannot do is reconstruct the broader structural consequences of those insertions - the deletions, inversions, and translocations that can follow when an L1 element jumps and the surrounding DNA is rearranged in the process.

Tubio's team used long-read sequencing, which reads much longer continuous stretches of DNA, to examine ten tumors selected specifically for unusually high L1 activity: five head-and-neck squamous carcinomas, four lung squamous carcinomas, and one colorectal adenoma. Across these ten tumors, they identified 6,418 retrotransposition events in total.

The vast majority - thousands of cases - were straightforward insertions, where L1 elements deposited copies of themselves into new chromosomal locations. Most were truncated and unlikely to jump again. But within that large set, the team identified 152 instances where an L1 event triggered something far more consequential: large-scale structural rearrangements involving deletions, translocations, and other major changes to chromosomal architecture. That translates to an incidence rate of roughly 1 in 40 for high-L1-activity tumors.

Three-quarters of these structural events would have been invisible to short-read sequencing. As Dr. Bernardo Rodriguez-Martin, one of the study's main authors, noted, the falling cost of long-read sequencing - expected to drop by roughly half in 2026 alone - means this kind of deep structural analysis will not remain a niche capability for long.

A Previously Unknown Chromosomal Exchange

Among the rearrangements the team documented was a mechanism not previously described in the scientific literature: a balanced reciprocal translocation apparently caused by two L1 events occurring on different chromosomes nearly simultaneously. First author Sonia Zumalave described it as two pages of a book torn at the same time, with fragments exchanged between them and L1 elements acting as the glue holding the joined pages together. The existence of this mechanism suggests that L1's architectural effects on the genome may be even broader than current data capture.

Early Events, Not Late Consequences

The study's most significant finding concerns timing. A standard milestone in tumor evolution is whole-genome doubling - a catastrophic event in which a cancer cell duplicates its entire chromosome set, typically occurring a median of 4.77 years before diagnosis in the tumors examined. The researchers found that most L1 activity in their samples preceded this doubling event.

That means L1 retrotransposition is not primarily a symptom of an already chaotic late-stage cancer genome. It is an early contributor to the genomic instability that makes cancer possible in the first place. Sixty-five percent of L1 events in the study occurred during early tumor evolution. The implication is that these ancient DNA elements are actively destabilizing chromosomes during the years-long preclinical phase of cancer development, creating more opportunities for the further mutations that drive disease progression.

A side analysis also found that L1 promoter regions were less methylated in tumors than in nearby healthy tissue - consistent with the idea that epigenetic changes to the genome can awaken dormant L1 sequences. The mechanism by which cancer cells lose the chemical silencing that keeps most L1 elements inactive remains an open question.

Deliberate Selection for Extreme Activity

The study's design carries a key limitation that the authors address directly. The ten tumors were deliberately chosen for extremely high L1 activity, making rare structural events detectable. That selection means the findings cannot be extrapolated to all cancers, or even to all cancers of the same types studied. In tumors with lower L1 activity - the majority - structural rearrangements driven by L1 would be less frequent, and some mechanisms observed here might not occur at all.

What the study establishes, within those constraints, is a framework: long-read sequencing can reveal L1's structural effects, those effects include a novel translocation mechanism, and they begin early. Whether that early activity represents a viable target for detection or intervention is the question the field will take forward.