Microbial load is a crucial but often overlooked factor in understanding the associations between the gut microbiome and various diseases. Researchers in the group of Peer Bork, PhD, at EMBL Heidelberg developed a machine learning model that estimates microbial load, a measure of the density of microbes in the gut, without additional experimental procedures, potentially transforming the way scientists analyze microbiomes.
The study highlights the importance of microbial load, as opposed to the more often used microbial composition, as a metric in microbiome research. The paper published in Cell is titled, “Fecal microbial load is a major determinant of gut microbiome variation and a confounder for disease associations.”
While microbiome research typically focuses on microbial composition—the proportional variety of bacteria, viruses, and other microorganisms within a sample—the total amount of microbes can provide a different view of microbiome data. Composition research approaches have provided insights into how specific microbes are associated with some diseases, but the total amount of these same microbes may have differing associations or impacts on diseases.
“We wanted to develop a new method that required no additional experimental methods to quantify microbial load,” said Suguru Nishijima, PhD, first author of the Cell report. This new model allows researchers to estimate microbial load based solely on microbial composition data, bypassing challenges to time-consuming and costly experimental methods for measuring microbial loads directly.
Nishijima added, “We had access to large datasets with both microbial composition and experimentally measured microbial load data. We wanted to see if we could use these to train a machine-learning model to estimate microbial load given microbial composition alone.”
To build and validate their machine-learning model, the researchers used data from two large-scale EU-funded projects: the GALAXY/MicrobLiver and Metacardis consortia, including data from more than 3,700 individuals. When tested on a separate dataset, the model successfully and accurately predicted microbial loads. They then applied the model to a much larger dataset comprising more than 27,000 individuals from 159 studies in 45 countries.
“Importantly, many microbial species previously thought to be associated with disease were more strongly explained by variations in microbial load,” Nishijima noted.
The analysis revealed new insights into microbial density variations stemming from various factors besides disease. Medications can significantly impact microbial loads. Digestive distress causes fluctuations in microbial load—diarrhea and constipation can reduce or increase microbial load, respectively. Furthermore, demographic factors including age and sex can impact microbial loads. Young people tend to have lower microbial loads than older individuals and women tend to have higher microbial loads than men, although there might be a link to constipation as women report constipation more frequently than men.
“These findings suggest that changes in microbial load, rather than the disease itself, may be the driver of shifts in the microbiome in patients. However, certain disease-microbe associations remained, and this shows that these are truly robust.” Nishijima concluded, “This further confirms the importance of including microbial load in microbiome association studies to avoid false positives or false negatives.”
The machine-learning model developed by Nishijima and colleagues is the first to predict microbial loads from composition data. This new model, available for other researchers to use for free, is expected to have additional uses outside of the initial gut microbiome studies.
Bork, who is director at EMBL Heidelberg, shared his thoughts on the broader impacts of this model: “Our oceans, soils, rivers are all teeming with microbes, and understanding these microbiomes could yield valuable insights to help preserve our planetary health. This study shows us that microbial load is an important measure that must be taken into account in such studies. Thus, we will work towards translating the knowledge on the gut microbiome to other habitats.”
