Sourdough starter, a fermented mix of flour and water, is a staple for bakers. It’s also reportedly a rich experimental testing ground for microbiologists. The bread’s chewy texture and tangy taste arise from the mix of microbes that leaven the dough, and past studies have identified more than 60 genera of bacteria and 80 species of yeast in sourdoughs from around the world. “We can use sourdough as an experimental evolution framework, to see what happens over time,” says evolutionary biologist Caiti Heil, Ph.D.

For a study published recently in Microbiology Spectrum, Heil and her collaborators at North Carolina State University, in Raleigh, analyzed sourdough starters to understand how the type of flour shaped the microbial community. They found that strains in the genus Kazachstania, a common sourdough yeast, to be most abundant in all the starters, but the bacterial composition varied by flour varieties.

For bakers, the study suggests that changing the type of flour could change the microbial composition. “And because the microbial composition affects different traits, by altering the flour you could potentially alter how your bread tastes,” says Heil, senior author on the study. More generally, she says, the findings highlight the sensitivity of the sourdough microbiome to environmental variables.

Previous studies have revealed a mix of variables that shape the sourdough microbial community, including the raw materials, the surrounding air and surfaces and the hands of the bakers. Plus, bakers may use wheat, rye, barley, teff, millet, or other grains in starters, each of which provide a different mix of nutrients for the microbial community.

The sourdough study began when Enrique Schwarzkopf, Ph.D., a sourdough enthusiast and postdoctoral researcher in Heil’s lab, launched a program at a local middle school aimed at teaching students about fermentation and evolution. For Schwarzkopf, who keeps a sourdough starter named Seth, the sticky stuff became a natural teaching tool. He challenged the students to experiment with different mixtures of flour and feeding schedules to see who could produce the fastest-growing starter.

Heil and her colleagues analyzed microbial communities in sourdough starters using metabarcoding, an efficient way to scan the genetic makeup of a material and get a snapshot of the microbes present. The sourdough samples had started with one of three substrates: all-purpose flour, bread flour, or whole wheat flour. At the outset, she said, the raw flours all had roughly the same composition of bacteria and a mix of yeasts.

After weeks of passaging, though, the researchers found the opposite: the same dominant yeast, and a mix of bacterial genera. She says she expected that their analysis would reveal strains of Saccharomyces cerevisiae, or brewer’s yeast, which are at the center of many of her group’s projects and are often used in baking.

Instead, the most dominant yeast was Kazachstania, a yeast that often shows up in fermented foods, no matter the kind of flour or feeding schedule. The genetic search also showed that while all the starters contained the same general mix of bacterial genera, starters made with whole wheat flour had a higher abundance of Companilactobacillus. Starters made with bread flour, on the other hand, showed a higher abundance of Levilactobacillus.

Heil, who studies how organisms adapt to new environments and how genetic mechanisms allow competition among microbes, notes that each type of flour provides different nutritional opportunities. Connecting those substrates to the environmental conditions of the microbial communities, she says, offers new opportunities for better understanding how diverse microbes compete and thrive.

Related: Dave's Killer Bread introduces Supreme Sourdough