Is it possible to use "missing genetic data" to estimate environmental and genetic effects upon disease?

Kelly Burkett, in the Department of Statistics and Actuarial Science, successfully completed her thesis on "Logistic regression with missing haplotypes".

She developed methods for medical researchers studying common disorders such as diabetes, asthma, cancer, and heart disease. These disorders are caused by a complex combination of genes and environment and represent a significant health-care burden. To better understand them, medical researchers have turned to association studies which consider the relationship between the disease, the environment, and blocks of genetic material called haplotypes. In some instances, haplotypes cannot be observed and it may be cost-prohibitive to determine them through laboratory techniques or other means. Therefore, researchers often make "guesses" for missing haplotypes based on partial information but do not account for the uncertainty in these guesses, leading to potential errors in interpretation. In her method, the effects of haplotypes can be studied even when only partial information is available. Her method is shown to detect the genetic signal better than an analysis which does not include environmental factors.

This type of interdisciplinary work is a hallmark of our program in Applied Statistics at Simon Fraser University. For more information, please contact Kelly Burkett (kellyb (at) stat.sfu.ca) or her supervisor Jinko Graham (jgraham (at) stat.sfu.ca).