Fred Hutchinson Cancer Research Center biostatistician Dr. James Dai recently was awarded a $2.1 million grant from the National Cancer Institute to develop new analytical methods for identifying genetic and epigenetic alterations that drive cancers. He hopes this work will facilitate the discovery of molecular cancer drivers — biomarkers — and the development of new preventive, diagnostic and prognostic techniques.
Dai is using data from prostate cancer studies conducted by Fred Hutch epidemiologist Dr. Janet Stanford and SWOG biostatistician Dr. Catherine Tangen that measure how genetics, epigenetics, and environment and lifestyle factors can influence the risk and prognosis ofdisease. Dai hopes to assess the clinical implications of these alterations while developing statistical methods for monitoring:
Using statistical assessment tools, Dai is sifting through thousands — if not millions — of genes, wading through excess, confounding data, or “noise,” in hopes of finding genetic and epigenetic variants that in some way contribute to prostate cancer.
He first embarked on this project a year and a half ago by asking himself how molecular epidemiology studies could do better in terms of design and analysis.
“We should be able to better integrate molecular phenotypes [genetic or epigenetic alterations that could drive cancer development] into population studies, improving cancer biomarker discovery,” he said, highlighting the growing integration of epidemiology and molecular research.
“Epidemiology has now put a molecular side into its story,” Dai said. “We can get blood, tissue and urine specimens from the same population over an extended period of time and use them to do molecular profiling. We are trying to use these markers that have the potential to predict disease development and progression as an instrument to study the mechanisms of the disease itself.”
A key focus of Dai’s project is testing which epigenetic alterations may shepherd disease progression as mediators, and/or act as a lookout for future risk of disease — an area of research in need of effective statistical models.
“The methodology is motivated by prostate cancer, but is not limited to it,” he said. “The models proposed, if accomplished, can have a far-reaching impact on the genetic epidemiology research field and potentially influence other cancer sites when they come across similar topics.”
Science writing intern Colin Petersdorf is a junior at the University of Southern California, where he is majoring in biological sciences and minoring in screenwriting. Tweet him @colinpetersdorf to talk medicine, baseball, Star Wars or anything in between.