The communities formed by human gut microbes can now be predicted more accurately thanks to a new computer model developed in a collaboration between biologists and engineers, led by the University of Michigan and the University of Wisconsin.
Model making also suggests a path to scaling up the 25 species of microbes explored to the thousands that may be present in human digestive systems.
“Every time we increase the number of species, we get an exponential increase in the number of possible communities,” said Alfred Hero, John H. Holland Professor Emeritus of Electrical Engineering and Computer Science at the University of Michigan and co. -corresponding author. of the study in the journal eLife.
“That’s why it’s so important that we can extrapolate from data collected on a few hundred communities to predict the behaviors of the millions of communities we haven’t seen.”
As research continues to uncover the many facets of how microbial communities influence human health, probiotics often don’t live up to the hype. We don’t have a good way to predict how introducing a strain will affect the existing community. But machine learning, an approach to artificial intelligence in which algorithms learn to make predictions based on sets of data, could help change that.
“Problems of this magnitude have required a complete overhaul of the way we model community behavior,” said Mayank Baranwal, assistant professor of systems and control engineering at the Indian Institute of Technology, Bombay and co -first author of the study.
He explained that the new algorithm could map the entire landscape of 33 million possible communities in minutes, compared to the days or months required for conventional ecological models.
Microbial Sim Cities
Ophelia Venturelli, assistant professor of biochemistry at the University of Wisconsin and co-corresponding author of the study, was integral to this major milestone. Venturelli’s lab conducts experiments with microbial communities, keeping them in low-oxygen environments that mimic the mammalian gut environment.
His team created hundreds of different communities with microbes prevalent in the human large intestine, mimicking the healthy state of the gut microbiome. They then measured how these communities changed over time and the concentrations of health-important metabolites, or chemicals produced when microbes break down food.
“Metabolites are produced in very high concentrations in the intestines,” Venturelli said. “Some are beneficial to the host, such as butyrate. Others have more complex interactions with the host and gut community.
The machine learning model allowed the team to design communities with the desired metabolite profiles. This type of control may eventually help doctors discover ways to treat or protect against disease by introducing the right microbes.
Feedback for faster model building
Although research on the human gut microbiome still has a long way to go before it can come up with this type of intervention, the approach developed by the team could help get there faster. Machine learning algorithms are often produced with a two-step process: accumulate the training data, then train the algorithm. But the feedback step added by the Hero and Venturelli team provides a template for quickly improving future designs.
Hero’s team initially trained the machine learning algorithm on an existing dataset from the Venturelli lab. The team then used the algorithm to predict the evolution and metabolite profiles of new communities that Venturelli’s team built and tested in the lab. Although the model performed very well overall, some of the predictions identified weaknesses in the model’s performance, which Venturelli’s team reinforced with a second round of experiments, closing the feedback loop.
“This new modeling approach, coupled with the speed at which we could test new communities in the Venturelli lab, could enable the design of useful microbial communities,” said study co-first author Ryan Clark, who was postdoctoral researcher in Venturelli’s lab when he conducted the microbial experiments. “It was much easier to optimize the production of several metabolites at once.”
The group opted for a short-term memory neural network for the machine learning algorithm, which is good for sequence prediction problems. However, like most machine learning models, the model itself is a “black box”. To determine which factors went into its predictions, the team used the mathematical map produced by the trained algorithm. It revealed how each type of microbe affected the abundance of the others and what types of metabolites it supported. They could then use these relationships to design communities worth exploring through the model and in follow-up experiments.
The model can also be applied to different microbial communities beyond medicine, including accelerating the degradation of plastics and other materials for environmental cleaning, the production of valuable compounds for bioenergy applications, or improving the plant growth.
This study was supported by the Army Research Office and the National Institutes of Health.
Hero is also the R. Jamison and Betty Williams Professor of Engineering, and Professor of Biomedical Engineering and Statistics. Venturelli is also a professor of bacteriology and chemical and biological engineering. Clark is now a principal investigator at Nimble Therapeutics. Baranwal is also a scientist in the Data and Decision Science Division at Tata Consultancy Services Research and Innovation.