Wan-Jen Hong

Wan-Jen Hong

Medical Student

Background:

BS in Chemical Engineering and Biology, MIT, 2001

Publications:

Rieger K, Hong WJ, Tusher VG, Tang J, Tibshirani R, Chu G: Toxicity from radiation therapy associated with abnormal transcriptional responses to DNA damage. Proc Natl Acad Sci USA 101: 6635-6640, 2004
Hong WJ, Warnke R, Chu G: Immune Signatures in Follicular Lymphoma (Corres). N Engl J Med. 352: 1496-1497, 2005 More Information
Hong WJ, Tibshirani R, Chu G: Local false discovery rate facilitates comparison of different microarray experiments. Nuc Acids Res, 2009

Research:

Currently, I am working with Kerri Rieger on developing a method to predict radiation toxicity in cancer patients using gene expression profiles. Ionizing radiation (IR) is often used in cancer therapy, but causes adverse reactions in some patients. Therefore, predicting radiation toxicity would allow doctors to individualize treatment and possibly prevent these acute reactions. We are using several statistical methods to identify genes whose responses to DNA damage are predictive for radiation toxicity. In collaboration with Rob Tibshirani in the Department of Statistics, our lab previously developed a supervised method called Significance Analysis of Microarrays (SAM) (Tusher et al., 2001). SAM identifies statistically significant changes in gene expression relative to the standard deviation of repeated microarray experiments. In addition, SAM provides an estimate of the false discovery rate (FDR), the percentage of genes identified by chance. Recently, a method called Heterogeneity Analysis of Microrarrays (HAM) was developed in our lab to account for different abnormal responses among the radiation sensitive patients. Another statistical method we use on our data is Prediction Analysis of Microarrays (PAM) which applies the nearest shrunken centroid method (Tibshirani et al., 2002). PAM provides a list of genes that distinguishes each class and estimates the prediction error using cross-validation. So far, we have produced a list of predictive genes that have transcriptional responses to either IR or UV that are different in radiation sensitive patients than in control patients. Also, these genes will help us understand the biological pathways involved in response to IR and UV damage. Using a combination of the statistical tools available to us, we have been able to separate most of the radiation sensitive patients from the control patients. We are continually improving our statistical methods to determine an improved list of predictive genes that can distinguish as many radiation sensitive patients from control patients as possible.

Future Directions:

Compare transcriptional response profiles of patients with known mutations in damage response pathways, such as ATM, NBS, BRCA1, or BRCA2, to response profiles of radiation sensitive patients and healthy patients.
Develop new statistical methods to reduce noise in microarray probe data

Hobbies: Playing tennis, watching TV, shopping, sleeping