jzwang@northwestern.edu
Assistant Professor of Statistics
Ph.D., 2003, Pennsylvania State University
My research focuses on two areas: (1) mixture model, computing algorithms and applications; (2) bioinformatics and computational biology. Mixture models provide a flexible and powerful modeling tool to a wide variety of applications. In a likelihood based approach to problems involving mixtures, one could find the global solution via nonparametric maximum likelihood estimation (NPMLE) of the mixture distribution. However the NPMLE solution in many important applications may suffer an instability issue. In my recent work, I have proposed a penalized NPMLE solution with a simple VDM/ECM algorithm, and illustrated its effectiveness in stabilizing estimators in species richness and population size estimation. My research in bioinformatics and computational biology aims to develop statistical methodologies to high throughput genomic and genetics data. The current working projects include Expressed Sequence Tag (EST) data analysis, nucleosome sequence alignment and prediction and human braining mapping. I have developed a statistical model and diagnosing tool called ESTstat for the expressed sequence tag (EST) data. Some tools we developed for nucleosome sequence alignment and prediction are also freely available at my personal website.
Some recent publications:
- A penalized nonparametric maximum likelilhood approach to species richness estimation (with Linday, B.), Journal of American Statistical Association, 2005,100(471):942-959.
- Improved alignment of nucleosome DNA sequences using a mixture model(with Widom, J.), Nucleic Acids Research, 2005 , 33(21):6743-6755 2005
- A genomic code for nucleosome positioning(with Segal, E.,Widom, J. et al) Nature, 2006, 442(7104):719-846
- Some optimization results and a VDM/ECM algorithm for penalized and constrained nonparametric maximum likelihood estimation for mixtures, Computational Statistics and Data Analysis , 2007, 51:2946-2957.
