In the late 1970s, scientists noticed that soft-shell clams from Maine were dying from a strange kind of leukemia. Large, cannonball-shaped cancer cells would fill their blood, turning it milky white, and eventually fatally clogging the mollusks’ organs.

For almost 40 years, scientists struggled to work out what was causing the cancer. But once they noticed that the disease seemed to spread from infected clams to uninfected ones, they suspected that a virus might be involved. That’s when Stephen Goff from Columbia University, who studies viruses that cause leukemia in mice, got a call.

He couldn’t find any viruses in the affected mollusks. Instead, his team discovered that the leukemia was associated with a new gene, which they called Steamer. It’s a retrotransposon—a jumping gene that can make copies of itself and paste those facsimiles elsewhere in the clam’s genome. Healthy clams have between two and 10 copies of Steamer in their DNA. But the cancer-afflicted ones had between 150 and 300 copies.

At the time, Goff wondered if something was making Steamer run amok, pasting new copies of itself throughout the clams’ DNA, disrupting important genes, and ultimately leading to cancer. If that was the case, each diseased clam should have a unique pattern of Steamer copies, strewn across its DNA. But instead, Goff’s student Michael Metzger found that clams across the eastern seaboard all had Steamer in the same places. Their tumors were all genetically identical.

The team eventually realized that the leukemia is a contagious cancer. Unlike almost every other tumor, which begins in an individual and then dies with it, the clam cancer cells are immortal, independent parasites that can move through the water from one host to another. So far, only eight such cancers have been discovered—two in Tasmanian devils, one in dogs, one in the clams, and four more in other kinds of shellfish.

Steamer’s role in all of this is unclear. It’s still possible that the gene’s cut-and-paste antics were the original trigger that created the contagious leukemia in the first place. But cancer aside, Goff and Metzger have found that Steamer has an extraordinary story of its own.

They found Steamer-like genes in the DNA of many species of shellfish—clams, oysters, mussels, cockles, and more. Bizarrely, some of these genes were incredibly similar, even when their hosts were only distantly related. For example, Atlantic razor clams and soft-shell clams have been evolving separately for between 300 and 500 million years, and many of their genes are only 65 percent identical. By contrast, their copies of Steamer are around 97 percent identical.

Another weird pattern: When the team created a family tree of the various species in which Steamer genes appear, and put it next to Steamer’s own genealogy, the two look completely different. This discrepancy means that Steamer isn’t just inherited in the usual vertical way, from parent to offspring. It also moves horizontally between individuals, and even between species. “These events are occurring in evolutionary time,” says Goff. “It could be one event every 10,000 years, or 100,000 years. But once that seeding has occurred, that single gene can easily expand within the new species.”

There are other examples of similar jumping genes, but “the scale documented here is unprecedented,” says Nancy Craig from Johns Hopkins University. Aside from shellfish, Goff’s team also discovered Steamer-like genes in animals from six other major groups (phyla), including corals, sponges, worms, and sea urchins. They’re even present in fish like zebrafish, carp, and salmon, and seem to have jumped from one fish genome to another at least five times in the past. In fact, the Steamer gene in zebrafish is almost identical to the original one that was first identified in the soft-shell clams. “It’s astounding,” says Goff. “We don’t have any bright insights into how it could have happened.”

Here’s a clue: Steamer genes only seem to jump between aquatic animals. It’s not in birds or mammals, insects or spiders. This suggests that it’s moving through the water—but how?

One possibility is that it’s acting like a virus. Steamer belongs to the same large group of retrotransposons that gave rise to retroviruses like HIV. That transition happened when some of these genes gained the ability to enclose themselves in an envelope of proteins, allowing them to travel outside their host cells and enter new ones. But Steamer genes haven’t made that leap yet. They can’t make their own envelopes, so it’s anyone’s guess how they might first escape from their host cells.

Another alternative is that parasites might carry Steamer from one animal to another. On land, bloodsuckers like ticks and leeches have been implicated in the spread of another jumping gene—BovB—between cows, snakes, elephants, bats, horses, and more. Goff’s team couldn’t find Steamer-like genes in any parasites that might act as a suitable vehicle, but that might just be because marine parasites are poorly studied.

It’s also unclear whether Steamer genes do anything in their hosts. They might just be selfish parasites, spreading themselves through the marine world and causing no harm. Perhaps they are occasionally causing cancers on their journey—or perhaps something else is helping both Steamer and cancer cells proliferate in some species. But there are several cases where evolution has repurposed jumping genes—and the retroviruses that arose from them—into the basis for new adaptations.

The syncytin gene, which is essential for creating the placenta, came from the envelope of a retrovirus. Parts of our immune systems have been rewired by retroviruses. And the Arc gene, which helps us to learn from experience and make new memories, comes from the same group of retrotransposons to which Steamer belongs. It’s possible that we’re only able to learn about jumping genes, and their incredible biology, because of jumping genes.