E209: Microbiome Research and a Vision of Precision Medicine

The microbiome is fascinating and how diet affects it is something we all want to know. The work of today's guest, Dr. Abigail Johnson, tells us a lot. She's on the faculty and the School of Public Health at the University of Minnesota and is an expert on diet and the microbiome and their impact on health and disease.

Interview Summary

So there's an awful lot of interest in your work but let's start with kind of a broad question. So there's much talk in the news and obviously in the scientific world about the microbiome and one fundamental question is this, how much should we pay attention? How important is this all? Is it important enough for dietary recommendations to be modified? For policies to be affected? Give us the big picture from your perspective.

I think it's really important. I wouldn't be studying it if I didn't think it was important. I think that there's so much that we don't know about what all of the microbes that are in and on our bodies are doing that impacts our health. There's just so much that we don't know at the moment that I think it's very important that we continue to study it. From a nutrition perspective which is where I spend my time, I think that there is a lot of potential that we could eventually understand how we can use food to change the microbiome. From that perspective, we might get to the point that we can make recommendations that affect national nutritional guidance like the Dietary Guidelines for Americans, for example. I don't know that we are there yet, but I do think that we are learning more and more with every passing year that gets us closer to being able to say some things definitively.

That's consistent with the way I've been seeing it and it's nice to hear it affirmed by an expert. So, your research on diet and health uses, and now I'm quoting from your website, "novel computational methods," can you explain what the methods might involve and how they improve the assessment of what people eat?

I did my postdoc with Dr. Dan Knights here at the University of Minnesota. I joined his lab as a nutrition scientist with very minimal background in statistics but an interest in computer science. I told him when I joined his lab that I wanted to learn to code. I immersed myself in this world of people doing bioinformatics and using computer science and bioinformatics to answer questions about the microbiome. What I noticed while I was in my postdoc was that a lot of the questions, statistical problems, and computational problems that exist in the microbiome space where people are coming up with unique ways to solve those problems mathematically are really analogous to issues that the field of nutrition has also been grappling with - sort of since its existence. When I say novel computational methods, I'm often referring to really stealing methods from the microbiome space and applying them to nutritional data. Like microbiome data, nutritional data have these sort of strange qualities that make them interesting from a mathematical standpoint. To be more specific, I've developed an approach that uses trees. In microbiome space we use phylogenetic trees to share information in statistical approaches to understand how communities are different. For example if we both haves E. coli species and they're actually different species, we can still use the fact that they're both E. Coli to understand the similarities between our different microbiome compositions. In the nutrition world, we think about foods as they fit under umbrellas of food groups. I've basically stolen the method of using trees in the microbiome and applied that to nutrition. That's kind of the novel computational approach that I take when trying to understand diet and dietary patterns.

That sounds like a real advance. What are some of the highlights of your work on microbiome in the diet?

In my work, I've been able to find signatures. We can pick up the signature of what you ate in your microbiome and the way your microbiome changes. We can link that back to your diet on a daily timescale. So if we have information about what you ate in the three days prior to providing a microbiome sample, then we can make some prediction about how your microbiome is going to change the next time we sample you. So not really good at predicting how it's going to change yet, but at least if we have the data, we can link you up with your diet and understand that the changes that happen in your diet show up in your gut microbiome. Also, being able to identify to some level of specificity, how much of the variation in microbiome, be it across populations or within an individual, can be connected back to differences in dietary intake.

You know that work that you're doing on using the microbiome as a measure of what people eat is a real advance, isn't it? For so many years the nutrition field has relied on people's reports of what they eat to make judgements about the relationship between diet and health and even individual dietary recommendation