| My general interest lies in the area of statistical genetics, genetic epidemiology and bioinformatics. My current interests can be divided into two parts: 1) The development and validation of statistical approaches and software for personalized cancer risk prediction and 2) the development of methods and software for accurate estimation of copy number and genotype calls on SNP arrays. I have been working on various projects related to these subjects as listed below. |
| 1. Documenting, testing risk prediction functions for breast and ovarian cancer syndrome and Lynch syndrome. Developing risk prediction functions for pancreatic cancer and cutaneous melanoma. These functions are included in the R package: BayesMendel, which is available from BayesMendel website. |
| 2. Developing statistical framework for validating risk prediction models using family registries (article in preparation, see publications). |
| 3. Developing statistical approaches for cancer syndrome overlap by 1) direct multi-allelic carrier probability model, which extends to arbitrary number of genes and pedigree members, and 2) using multi-dimensional classification models to determine probabilities of each joint genetic mutation status. |
| 4. Estimating allele-specific copy number and therefore making genotype calls for SNP arrays using robust linear mixture models (extended abstract accepted to RECOMB2007, see working paper) as well as developing an R package CN-RLMM for this application (under development). |