Exploring how a dramatic shift in diet has altered our health

Soda, salty snacks, and processed foods are among the top groceries purchased by Americans – a diet far removed from the one consumed for most of human history.

Now researchers, health professionals, and the public want to know: how has this change in diet affected us and what, if anything, should we do about it?

Hundreds of bacterial species and trillions of bacterial cells in the human gut allow us to regulate and maintain metabolic and immune system functions. Keeping this microbial community stable is crucial for human health and well-being – unbalanced gut microbes have been linked to diseases such as diabetes and inflammatory bowel disease.

And due to a widespread increase in the use of antibiotics and a higher-sugar, higher-fat diet, the composition of human gut microbiome has shifted.

Scientists do not currently know how to manipulate microbiomes to explore how this shift has impacted our health. This would require an understanding of how species interact within the microbial network and what impacts its stability. Presidential Fellows in Data Science Lu Tian and Yingnan Gao, supported by the University of Virginia Data Science Institute, are using microbial genomics and machine learning to develop new tools that will allow us to understand microbial stability and improve human health.

“This research helps biologists better understand the dynamics of floras of microbes,” Tian said, “such as the microbes inside human gut.”

Tian and Gao obtained microbiome data from the American Gut Project to identify which bacterial species are present in different people and their abundance. The researchers then analyzed microbial DNA sequences from more than 15,000 people, giving them a large-scale data set.

Tian and Gao are using this data to investigate how microbial species interact through both integrated statistical analyses and stratified analyses, in order to explore the complex data set. The team also developed a high dimensional graphical model to identify the distribution of each microbial species.

This novel research project extends beyond current projects reconstructing the microbial network using overly simplified models that only account for some species relationship, while ignoring the effects of the whole network, and aims to construct a holistic account of the gut microbiome.

“This research can help us improve overall physical health,” Tian said, “In other applications it can also help us better protect the environment and maintain the ecological balance.”

Completed in:

Data Science Institute
Office of Graduate and Postdoctoral Affairs