The long scientific names of bacteria often tell a story of their discovery. Over time, however, the names can get a little awkward in light of new revelations.

Consider Elizabethkingia anophelis—a largely unremarkable microbe named after the pioneering microbiologist Elizabeth O. King, who discovered the genus, and the Anopheles mosquito, in whose gut it was first found in the Gambia.

In 2016, this bacterium was suddenly all over the news in Wisconsin. People were getting sick with Elizabethkingia anophelis blood infections and dying. The victims were often elderly and in poor health. But this microbe had never, to anyone’s knowledge, caused such a big outbreak before. And it was winter. In Wisconsin. There were no mosquitoes. Eventually 20 patients died: 18 in Wisconsin, one in Michigan, and another in Illinois.

What happened? Did something change in Elizabethkingia anophelis to make it so much more dangerous? Yes, according to new genetic analysis of deadly strains from those patients in the Midwest. This may be alarming but it should not be surprising: With a stroke of evolutionary luck, harmless microbes can often turn into pathogens.

The study came together, of all places, on Twitter. In an uncharacteristic act of transparency, the Centers for Disease Control and Prevention began posting the DNA sequences of Elizabethkingia anophelis from patients online in near real time. An @CDC_AMD tweet about this caught the attention of Kat Holt, a microbiologist in Australia. Sylvain Brisse, a microbiologist at the Institut Pasteur in Paris, France, who had the extraordinary timing to have just finished a global survey of Elizabethkingia anophelis diversity, eventually joined in, too.

“We call it the tweet heard around the world,” says John McQuiston, the team leader for the CDC’s Special Bacteriology Reference Laboratory. “I literally was sleeping while this was happening and everybody’s emailing me within 20 hours.” Scientists in France, Australia, Wisconsin, and the CDC all ended up collaborating on the study.

The first thing they noticed is that all of the sequences looked very different. In a typical outbreak with a single strain, the DNA sequences from bacteria in different patients might have a handful of variations between them. “We were seeing 40, 50, even 100 differences between some of them,” says McQuiston. But the sequences were definitely from the same strain, and the tell was one especially big and obvious mutation: a chunk of DNA more than 60,000 letters long inserted in the middle of a gene called mutY. The gene codes for a protein that repairs broken DNA, but the insertion makes the gene useless. Now it made sense: No mutY, no DNA repair system, a cascade of subsequent mutations. That’s why all the sequences were so different.

The vast majority of mutations will be bad, but a tiny percentage might be helpful. The authors think that one or several of those mutations may have then given Elizabethkingia anophelis a new foothold—like extracting more nutrition from its environment. Or they might have made the bacteria more virulent in humans.

It’s the first time researchers have identified this specific mutY mutation in a pathogen, to the authors’ knowledge. But one or a few mutations can tip a microbe from harmless to pathogen. “Many emerging pathogens have evolved from previous states that were less transmissible and less virulent,” says Brisse. It’s why we worry about new strains of flu evolving every year. And why some strains of E. coli live harmlessly in your gut and other E. coli make you violently ill. It’s not enough to simply know the species of bacteria—you have to know the specific strain.

The epidemiologists never did quite identify where the Elizabethkingia anophelis strain that killed so many people in the Midwest came from. You usually see this kind of bacteria in a single hospital or retirement home, where many people of poor ill health are living together. But the victims in the 2016 outbreak were spread across more than a dozen counties, pointing to some a source out in the community rather than a hospital.

Scientists tested tap water, drains, lotions, soaps, shampoos—all moist places bacteria like to hang out—but the strain didn’t come up in  the wild. The outbreak eventually died down. The Wisconsin Department of Health Services and the CDC are still testing the occasional Elizabethkingia sample they get from patients, so that they can catch any future outbreaks early. Lina Elbadawi, who worked on the new study as an epidemiologist with the CDC and Wisconsin Department of Health Services, noted the genetic analysis suggests Elizabethkingia anophelis had evolved its mutation a year before it started sickening people—so perhaps it was lurking in a product that had been stored for months and later distributed.

Mosquitoes, of course, were also discussed as a possible source during the outbreak. But again, this was winter in the Midwest. No one in the scientific literature had ever documented a case of mosquito transmission of Elizabethkingia anophelis before. No one knew the bacteria could be so deadly either. The name endures for now, a vestige to remind us how little scientists knew of the bacteria they thought they understood well enough to name.