Chloroplasts and mitochondria are, famously, the only organelles to have their own genomes, separate from that housed in the nucleus. According to a study published October 5 in Nature, it turns out that every so often, DNA escapes from mitochondria and integrates itself into our chromosomes. The findings shed light on how the nuclear genome is evolving.

Mitochondria are thought to trace their origins to a bacterium swallowed by a eukaryotic cell two billion years ago. The primitive prokaryote stuck around, whittling its genome down to just thirteen protein-coding genes in humans. Along the way, some genes simply vanished, while the rest migrated to the nucleus before we evolved from our primate ancestors. Since then, mitochondrial DNA (mtDNA) has been thought to remain separate from the nucleus, inherited wholesale from our mothers.

But in 2018, a study challenged this simple picture of maternal mitochondrial inheritance, describing paternal mtDNA in 17 individuals and suggesting that the organelle might be inherited from either parent. “This was a really wild idea,” Patrick Chinnery, a neurologist at the University of Cambridge in the UK, tells The Scientist. “We had a healthy skepticism about the discovery and went looking for other explanations.”

Chinnery and his colleagues’ initial analysis of people testing positive for paternal mtDNA revealed that they hadn’t inherited the organelle itself from their fathers. Instead, pieces of paternal mtDNA had plastered themselves onto the nuclear genome. In the new study, the team scoured whole genome sequences from around 66,000 participants in Genomics England’s 100,000 Genomes Project. The researchers specifically searched for sections of DNA where one half could be mapped to the nucleus and the other half to the mitochondria.

The search turned up a total of 1,637 mitochondrial fragments in the nuclear genome, with more than 99 percent of people harboring at least one. An average person possesses around five inserts that have not been previously described, suggesting they hopped onto the genome recently in human evolution, the team reports.

The results suggest that “these transfers of genomic information from the mitochondria to the nucleus are still happening today,” says cell biologist Cole Haynes of the University of Massachusetts Chan Medical School, who was not involved in the study.

The researchers’ database contained sequencing data from around 8,000 single-child families, allowing them to measure how frequently the process happens. A new mitochondrial fragment appears in the nuclear genome once in every 4,000 births, they found—a surprisingly common event considering that worldwide, around 385,0000 babies are born every day. For Haynes, pinpointing these events in which transfer of mtDNA to the nucleus occurred was the study’s “coolest observation.” Such transfer “is happening all the time,” he says.

The study has broad implications for eukaryotic evolution, says Iain Johnston, a computational biologist at the University of Bergen in Norway who was not involved in the work. Instead of mitochondria being frozen in time, their information transfer with the nucleus is “continuous and dynamic,” he adds.

Fragments of mtDNA were usually found next to binding sites for PRDM9, a protein involved in repairing double-stranded breaks in DNA. This suggests that mtDNA may integrate into the genome at sites where both DNA strands are severed, the authors write in their paper. “Bits of mitochondrial DNA preferentially stick in these holes and act as a kind of Band-Aid to repair the nuclear genome,” says Chinnery.

Of the 66,000 genomes in the initial analysis, 12,000 were from people with cancer. A comparison of blood cells and tumor cells from these individuals revealed more mitochondrial segments in cancer cells, possibly because they experience more DNA damage.

The mtDNA fragments tend to slot into noncoding DNA and are inactivated by the addition of methyl groups, suggesting that they pose no health risks, Chinnery says. But his team uncovered rare cases of segments sticking to regions known to be drivers of cancer, which he says may trigger tumor development.

A “more fine-grained investigation” into the role of mitochondrial DNA in cancer may prove informative, says Johnston.

The next step, according to Chinnery, is working out how these insertions happen. One possibility is that the DNA somehow maneuvers its way through the mitochondrial double membrane. Alternatively, mitochondrial messenger RNA may be to blame, in which case it would need to be reverse-transcribed before making its way into the nuclear genome. “Of course, that’s what viruses do,” says Chinnery. Our own microbial squatters—mitochondria—may perform a similar trick.