Sign Up for Fundamentals

Stay up-to-date with the latest research findings from the Institute for Basic Biomedical Sciences.

Please enter a valid email address.
Fundamentals Topics+

Alexis Battle on Teaching Machines to Make Health Predictions

Profiles

More Profiles

Alexis Battle on Teaching Machines to Make Health Predictions

Interviewed by Catherine Kolf

Alexis Battle on Teaching Machines to Make Health Predictions

Alexis Battle is an assistant professor of computer science. She uses machine learning to better understand how human genetics influence health. 

How did you become interested in science?

BATTLE: My family was very encouraging. My parents have both been in academia, and my older brother is a doctor. When I was a little kid, I was interested in math and science, and I figured out that engineering was a good way to incorporate both.

After high school, I did a summer internship in computer science at the University of Texas. That project was my first exposure to computational work, and it got me interested. Then, I went to Stanford and started working with Daphne Koller. She introduced me to computational biology, which allowed me to combine several of my interests, with the goal of improving human health.

What are the goals of your research?

BATTLE: We use “machine learning” to better understand how human genetics influence health. Machine learning involves writing complex computer programs that use a mathematical framework and large sets of data to “learn” to make predictions on their own. The specific machine learning methods we use are also able to incorporate what we already know about biology to make better predictions for a new data set or disease.

What type of predictions are you trying to make?

BATTLE: We want to learn how genetic variants affect a person’s biological traits, like height and weight, and how they influence the risk of getting particular illnesses. Some diseases are genetically simple—we know that they are caused by changes in a single gene, like cystic fibrosis. But many diseases, like autism and heart disease, are much more complex. They seem to be affected by alterations in many genes and by environmental factors, so causes are even harder to identify. Computers are able to look for patterns in huge amounts of data that a human being could never handle.

What are some of the challenges you face?

BATTLE: The more data, the better our predictions, but we never have as much data as we’d like. We want to minimize finding false correlations and missing real ones. Even a few thousand participants in a study are not always sufficient for drawing strong conclusions, given the size and complexity of the human genome.

What do you hope to learn from your research?

BATTLE: For individuals, we hope to get to a point where we can use genetic information to determine which diseases someone is most susceptible to so that he or she can seek treatment early or make helpful lifestyle changes. Also, if we learn what genes and cellular components are most influential for complex diseases, these can be targeted by researchers developing new therapies.

What do you enjoy doing outside of science?

BATTLE: Since my son was born recently, mostly spending time with my family, but I also enjoy camping and biking.

Alexis Battle on machine learning as a biomedical discovery tool

Alexis Battle, assistant professor of computer science at The Johns Hopkins University, describes how computer programs can "learn" to better predict which genes contribute to various health problems.