Microbiota—the ecosystems of bacteria, fungi, and other microbes in humans’ and other animals’ guts, genitals, mouths, and elsewhere—are astoundingly powerful organisms. They can make us sick or help keep us well. They can even influence our behavior. And now, scientists are realizing, they’re probably altering the outcomes of our experiments.

In recent years, the scientific community has wrestled with what some have called a “crisis of reproducibility”—the realization that the bulk of published experiments cannot be successfully replicated by other researchers. One of the most likely culprits is human error, or, more specifically, the very human inability to notice and consider variables that could affect an experiment’s results. And when it comes to animal studies, there are a lot of variables to consider. One experiment found that a parasite changed a commonly studied behavior in zebrafish, while another noted that uncomfortable temperatures in the lab can make mice more susceptible to cancer. Now the rodents’ belly bacteria are under the microscope.

Because as powerful as our microbes are, they’re also incredibly sensitive. Studies have shown that the communities in our bodies can be affected by the company we keep, whether or not we use deodorant, and even by what our grandparents ate. They’re essentially a variable stuffed with other variables. And when they change, so do we.

Michigan State University physiology researcher Laura McCabe learned this firsthand when she decided to repeat one of her own successful mouse experiments. “I was thinking, ‘Hey, great! Let’s repeat it one more time to be certain,’” McCabe told Kelly Servick in Science. But she couldn’t. McCabe and her team tried to reproduce their results in a new set of mice, to no avail. They tried a second time and failed again. At a loss, they decided to take a look at the rodents’ gut bacteria. They took poop samples from the control groups of all three experiments. There, under the microscope, they saw the difference: Tthe three groups had very different microbiota even before the experiments began.

McCabe and her colleagues are in good company. Last year, experts at the National Institutes of Health’s Mutant Mouse Resource and Research Center (MMRRC) set out to determine what a healthy, normal mouse microbiome looked like. They took poop samples from mice from two major scientific mouse suppliers, looking for a consensus. They found some overlap, but also some significant differences. One vendor’s mice were missing important immune-boosting microbes called segmented filamentous bacteria. Without them, the mice were more likely to fall ill.

Today, researchers concerned with matching their mouse guts can send samples of their mouse poop to the MMRRC for analysis. Some scientists predict that within the next two decades, published papers will have to include microbiome analysis. Until then, it’s up to researchers to keep track of the little things.

[h/t Science]

Know of something you think we should cover? Email us at tips@mentalfloss.com.