A recent Massive Science article by Adriana Romero-Olivares highlighted the fact that many fungal and protist species are missed in microbiome studies, which tend to focus heavily on just bacteria. If this doesn't change, we might miss out on a lot of important discoveries. But now, a new paper in Nature, led by Marcus de Goffau and Gordon Smith of the University of Cambridge, U.K., shows that the opposite is also true: some microbiome studies are finding microbes that aren't actually there at all!
Ed Yong explained this finding in his excellent piece for The Atlantic. A 2014 study claimed to find a defined microbiome in the human placenta, an environment previously thought to be sterile. This surprised scientists, who had long thought that we are colonized by our microbial friends soon after birth, based on the fact that babies delivered by C-section take a lot longer to develop a stable and healthy microbiome than babies born vaginally. The paper made waves back then because, if the placenta itself were also colonized, this would certainly change our view on when and how the human body becomes host to a complex microbial ecosystem.
But the recent Nature paper by de Goffau and colleagues strongly challenges the published observation of a placental microbiome. They showed that there were reproducible microbial contaminants in the equipment and reagents used in the 2014 study, a nightmare scenario for lab-based scientists. Knowing that this type and scale of microbial contamination is possible in microbiome studies is hugely important, and encourages a healthy level of skepticism among readers of such studies.
Although the whole thing might seem like an embarrassing oversight to some, rest assured that this is how science should work: we test our colleagues' work and refine our findings based on the data. Mistakes and misinterpretations happen, but we are always in search of the truth.