Making Data Count
Balasubramanian uses statistical models to address a range of public health needs, from improving early detection of HIV infection in infants to identifying cardiovascular risk in adults.
Balasubramanian’s expertise lies in the development and application of statistical models for problems stemming from pediatric HIV clinical trials, metabolomics (the study of complete sets of metabolites), andsystems biology studies. She has spent years designing and analyzing outcomes in HIV-related clinical studies in order to help improve early detection of HIV infection in infants. Along with collaborators at Harvard School of Public Health, Johns Hopkins School of Medicine and the Centers for Disease Control and Prevention, she has participated in a World Health Organization working group aimed at revising guidelines for early diagnosis of HIV in infants born to HIV positive mothers. In addition to evaluating the properties of HIV diagnostic assays in infants, she is analyzing the feasibility of less expensive diagnostics that might be used in resource-limited settings—where 90 percent of the approximately 800,000 annual cases of transmission occur, according to the United Nations program UNAIDS.
Balasubramanian notes that while research in the 1990’s revealed how treating HIV-positive pregnant women with antiretroviral therapies dramatically reduces the risk of transmission from mother to child, these highly potent antiretroviral therapies make it difficult to accurately diagnose HIV infection in infants as they suppress viral replication. Because it is important to treat infected infants as early as possible and it can be difficult (and costly) to repeat testing, Balasubramanian and her colleagues are analyzing the data in hopes of streamlining the diagnostic process. They have developed a highly diverse and complex data set comprised of several cohorts, including approximately 5,800 HIV-positive infants born to HIV-positive mothers in the U.S., Europe, Africa, and Asia. She and her collaborators are also evaluating the performance of other less resource-intensive procedures, such as using dried blood spots for HIV testing in infants.
In the area of biomarker discovery, Balasubramanian is collaborating with colleagues at Brigham and Women’s Hospital to identify metabolomic markers that could be used to predict cardiovascular disease in women. Work with metabolomes, or complete sets of metabolites, is a new and very active area of life sciences research. Whereas genes are the “rulebook” in biochemical reactions, metabolites are the key actors. The team is measuring 300 metabolites in approximately 1,500 women subjects drawn from the Women’s Health Initiative, a large-scale, epidemiological study that followed about 160,000 postmenopausal women for 15 years. In the subset of women who have given blood samples, Balasubramanian and her colleagues are searching for markers that can signal increased risk of a cardiovascular event.
Balasubramanian, who is funded largely through the National Institutes of Health, hopes to expand upon her metabolome research to study chronic stress, post-traumatic stress disorder and neurological disorders.
Closer to home, Balasubramanian is affiliated with the campus’s Institute for Computational Biology, Biostatistics, and Bioinformatics and utilizes the computing power of the Massachusetts Green High-Performance Computing Center in nearby Holyoke—a research computing data center operated by the University of Massachusetts, Boston University, Harvard University, MIT, and Northeastern University.
Balasubramanian’s work simultaneously fulfills her love for mathematics, her desire to positively impact human health and to work on interdisciplinary research.
“We’re all about working hands-on with data,” Balasubramanian says. “The nature of our work is inherently very collaborative.”
Amanda Drane '12