University of Michigan
News for Faculty and Staff

October 22, 2017

Projects use Big Data to predict diseases, advance genomics analysis

January 24, 2017

Projects use Big Data to predict diseases, advance genomics analysis

Researchers at the University of Michigan will use Big Data and mobile technology to learn how to predict when individuals will get diseases including depression and hepatitis C, and to unlock the potential of single-cell gene sequencing under three recently funded projects.

The Michigan Institute for Data Science awarded the three interdisciplinary projects a combined $3 million under the second round of its Challenge Initiative program. The program is part of U-M's $100 million Data Science Initiative, which was announced in September 2015.

MIDAS co-director Brian Athey, professor and chair of computational medicine and bioinformatics, said taken together, the projects show U-M researchers' ability to advance translational science — from pure research to wide application — using Big Data.

"These projects have the potential to improve the lives of millions of people and to enhance our understanding of the basic elements of cell biology," he said. "Plus, the data science tools and methodologies being developed by the U-M research teams will be applicable for many other fields of inquiry."

One of the awards is for the Michigan Center for Health Analytics and Medical Prediction, which aims to improve researchers' ability to diagnose and predict acute respiratory distress syndrome and hepatitis C.

The project is led by Brahmajee Nallamothu, professor of internal medicine, and will bring together an interdisciplinary team to find patterns over time in the massive amounts of data produced in the health care industry.

Although health care produces an extraordinary volume of information on patients as they receive care in hospitals and clinics, temporal patterns in data from individual patients or   groups of patients with the same condition are frequently overlooked.

"For data science to become a regular part of health research, it must be socialized into the day-to-day workflow of medical domain experts — especially as clinicians are frequently the rate-limiting step in implementing new advances due to their general unfamiliarity with these tools," said Nallamothu, who is a member of the Institute for Healthcare Policy and Innovation.

"M-CHAMP is specifically designed to integrate medical domain experts and leaders with method and informatics experts in statistics, computer science and engineering to tackle problems that directly impact patient care."

The M-CHAMP research team includes members from LSA, the College of Engineering, School of Nursing, School of Public Health and Medical School.

Another project, the Michigan Center for Single-Cell Genomic Data Analytics, will tackle the challenge of analyzing single-cell genomics data — that is, the detailed sequence of all the RNA contained in a single cell. Such information can be very useful in studies of cancer and cell development, but is often noisy and distorted.

"Our goal is to develop new tools to properly deal with such data. We want to build strong support in mathematical theory and promote data science practices that improve reproducibility," said project leaders Jun Li, associate professor of human genetics, and computational medicine and bioinformatics, and Anna Gilbert, Herman H. Goldstine Collegiate Professor of Mathematics, and professor of mathematics, and electrical engineering and computer science.

"Our team includes experts in statistics, machine learning and algorithms, as well as biomedical researchers. The methods we develop will also have an impact in the study of electronic health records, consumer data, mobile device data and many other types of noisy, heterogeneous data."

The research team includes members from the Medical School, SPH, LSA, CoE, Computational Medicine and Bioinformatics, and the Comprehensive Cancer Center.

A third project will use data collected from the cell phones and wearable sensors from over 1,000 medical interns to chart the relationships between mood, sleep and circadian rhythms as they relate to the onset of depression. The goal is to recognize patterns that predict depression in order to identify at-risk individuals and provide preventive treatment.

Srijan Sen, the Frances and Kenneth Eisenberg Professor of Depression and Neurosciences in the Medical School's Department of Psychiatry and Molecular and Behavioral Neuroscience Institute, is leading the project. Sen is a member of the U-M Depression Center and the Institute for Healthcare Policy and Innovation.

"This work has the potential to transform our ability to predict the development of depression under stress and to get personalized interventions to patients when they most need them," he said.

"These interdisciplinary projects will advance innovation and cross-disciplinary collaboration across data science and health research," said MIDAS co-director Alfred Hero, professor of electrical engineering and computer science. "Combining engineering, medicine, health, nursing, and science faculty, these research projects will lead to exciting developments impacting precision medicine, personalized health and associated biotechnology."

The goal of the multiyear MIDAS Challenge Initiatives program is to foster data science projects that have the potential to prompt new partnerships between U-M, federal research agencies and industry. The challenges are focused on four areas: transportation, learning analytics, social science and health science.

Leave a Comment

Please read our comment guidelines.

Will not be published.