Projects and Publications

Comparability of exposure information is of central importance to the assessment of gene-environment interaction studies. Given that a central goal of the Public Population Project in Genomics (P3G) is the promotion of collaboration between researchers in the field of population genomics through harmonization of biobanks, the development of methods to calibrate measures of exposure is essential.

Principal Activities:

• Methodologic workshop to review empirical and theoretical work on combining data on exposures across studies, including the assumptions that have to be made in applying modeling approaches to combining data. Priorities will be set to develop the modeling framework that will be applied to calibrate measures of exposure used in the various biobanks.
• Systematic review and simulation studies to compare statistical models using simulated data. These data will reflect different scenarios about measurement error, its dependence on exposure frequency, gene-environment interaction and confounding structure. Design of the review, and selection of the scenarios, will be informed by the workshop discussions.
• Empirical calibration studies to compare multiple instruments relating to specific exposure in the same individuals in one population.

Julian Little, Muin Khoury, Paul Burton, Isabel Fortier

Members of P3G; HuGENet Network of Networks

Quantitative

This research project focuses on the investigation of the utility of germline genomic profiling in cancer screening, using a candidate gene approach.  Our aims include evaluation of the predictive power and clinical validity of genetic profiling as a component of a population based cancer screening program.  The focus of our work to this point has been a pilot (funded through a CIHR Pilot Project Grant) aimed at evaluating the utility of germline genetic information as an adjunct to population-based colorectal cancer (CRC) screening, through theoretical work, through statistical analysis of a population-based case-control study of CRC, and through focus groups conducted with health professionals and potential screenees.

As evidence is accruing from genome-wide association studies and other studies that at least some risk loci confer elevated risk for more than one type of cancer, we believe that it is timely to expand upon this pilot work in CRC, and extend our scope to include other common cancers.  In addition to simulation studies, extensive reviews of published evidence, as well as the qualitative components, we are currently collaborating with investigators from the ARCTIC case control study. The ARCTIC study looked at risk factors for colorectal cancer in population-based CRC cases and healthy controls from Ontario, with additional validation samples from Newfoundland and Seattle.  The ARCTIC study has genotyped >600,000 SNPs in the Ontario subjects, as well as collecting detailed questionnaire data.  We plan to perform additional genotyping where justified and feasible, in the ARCTIC data and in additional data from both prospective and retrospective studies, though our expanding network of collaborators.  Any additional genotyping will be guided largely through systematic reviews of evidence.

Qualitative

Julian Little

Nick Birkett, Brenda Wilson, John McLaughlin, Samual Lunenberg, Brent Zanke, Tom Hudson, Quanhe Yang, and Muin Khoury