Department of Microbiology and Immunology
Assistant Professor in Bioinformatics
While high-throughput genomics technologies are revolutionizing many aspects of modern biology, the lack of computational algorithms and resources for analyzing the massive data generated by these techniques has become a rate-limiting factor for scientific discoveries in biology research.
Our lab studies machine learning, bioinformatics and their applications to cancer informatics and metagenomics. Our work is based on solid mathematical and statistical theories, and its focus is twofold: 1) developing advanced algorithms and building computational infrastructures to help biologists keep pace with the unprecedented growth of genomics datasets available today, and 2) enabling them to make full use of their massive, high-dimensional data for various biological inquiries.
We are currently working on three major projects. The first project is funded by the National Science Foundation (NSF). Our goal is to develop an integrated suite of computational and statistical algorithms that enable researchers to process millions of 16S ribosomal RNA sequences in order to: 1) derive quantitative microbial signatures to characterize various infectious diseases, 2) interactively visualize the complex metagenomic structure of a microbial community, 3) study microbe-microbe interactions and community dynamics, and 4) identify novel species. We collaborate with researchers throughout the University at Buffalo and at the University of Florida to apply bioinformatics algorithms developed in this project to various applications.
The second project is funded by the National Institutes of Health. This is a joint project with Dr. James Jarvis in UB’s Department of Pediatrics. Our goal is to use high-throughput genomics technologies to identify molecular markers that characterize juvenile idiopathic arthritis.
In the third project, we use advanced machine learning algorithms to develop computational models for breast cancer prognosis by using all available information, including clinical variables and genetics information. We have conducted extensive computational studies using thousands of cancer tissue samples and have obtained solid evidence suggesting that cancer progression trajectories exist. I hope that our work can significantly advance our understanding of the underlying mechanisms of cancer growth and thus open new avenues for cancer research.
The algorithms and software related to metagenomics and feature selection developed in my lab have been used by more than 200 research institutes worldwide to process large, complex data sets that are core to a wide variety of biological and biomedical research.