placeholder image
  • Contact Info
  • Websites

Jiang Gui, PhD
Associate Professor of Biomedical Data Science

Positions

My research involves the development of statistical and computational methods for relating high-dimensional microarray gene expression data to censored survival data. Since there is usually a large variation in time to certain clinical event (e.g. tumor recurrence), analysis focusing on time to response is clinically more relevant than simple classification. I have developed dimension reduction and penalized regression methods to overcome high-dimensionality and built prediction models for patients’ survival probability. I also developed a Gaussian Graphical method to infer gene network for isoprenoid biosynthesis in Arabidopsis thaliana.

Another component of my research program involves identifying gene-gene interaction and gene-environment interactions. I have developed several novel non-parametric machine learning algorithms to detect and characterize gene-gene and gene-environment interactions in the absence of statistically significant main effects. We have successfully applied it to a population-based bladder cancer study in New Hampshire and identified several biologically meaningful gene-gene interactions that have a stronger prognostic effect than smoking status that is a well known cancer risk factor. We also implemented these algorithms in an open-source software package (MDR 3.0) for all researchers. In addition, we proposed to use Weighted False Discovery Rate (WFDR) method to address the multiple comparison issue arise from genome-wide association studies. We plan to incorporate expert biology knowledge about gene function, gene location and biochemical pathways into a weighting scheme to improve power for our analysis.

The DNA methylation and microRNA expression data collect by Dr. Andrew’s project is high-dimensional and requires proper statistical modeling and inference in order to draw conclusions. This is consistent with my research interest. With my strong background and training in methods germaine to the proposed project, I will lead the cleaning, quality control and analysis of the microRNA expression data and classification analysis.