Student Profiles
Julia Kozlitina, 32, Statistics
Julia Kozlitina, who grew up in Ryazan, Russia, during the fall of communism in that country, an era of great economic turmoil, earned her undergraduate degree in economics from a college in Russia. She came to the United States in 2000 for graduate school, earning a master's degree in agricultural economics in 2002. She began her pursuit of a Ph.D. in statistics at Southern Methodist University in 2003 and will complete her work this year.
Kozlitina says students working on Ph.D.s in statistics often apply statistical methodology to data from other disciplines. "There are applications in all kinds of sciences and even business. You might use it in labor law to assess discrimination or in engineering to assess reliability of products," she says.
Her Ph.D. work involves developing new methodology for the analysis of data gathered by the Center for Human Genetics at UT-Southwestern Medical Center. One of their major projects is the Dallas Heart Study, a large, multiethnic population sample of people in Dallas County. The project has gathered data on 10,000 genetic markers on 3,000 individuals. Kozlitina's work involves analyzing this mass of data and teasing out correlations between certain genetic variations and risk factors that are related to heart disease, such has high levels of LDL cholesterol or a high body mass index.
In particular, her task is to develop statistical methods for analyzing genetic data. Each gene has a contribution from both the male and female parent, so there are three possible variants, or alleles, for each gene: AA (homozygous dominant), Aa (heterozygous), and aa (homozygous recessive). The homozygous dominant allele for one gene might contribute strongly to high cholesterol, while the heterozygous allele for that gene has a weaker link to high cholesterol and the homozygous recessive allele has no correlation with high blood pressure. Because of this stair-step nature of genetics, standard statistical tools such as analysis of variance, are not the optimal way to analyze genetic data. Kozlitina has developed a method of statistical analysis that is particularly suited to genetic data, and is applying that method to the set of data from the Dallas Heart Study.
