Theory and study design
-
Q1:
What problems can best be solved by “discovery-driven” approaches? What problems are best addressed by “hypothesis driven” approaches and what theoretical approaches would be most helpful in guiding hypothesis generation?
-
Q2:
Should G-E studies be targeted, that is, focused on a particular gene, exposure, phenotype, or disease? Or should studies be broad, designed to encompass as many factors as possible?
-
Q3:
To what extent can existing studies be adapted to investigate G-E interplay? Which questions will require the development of new cohorts?
-
Q4:
Are there research designs that allow us to investigate the complexity (on G and E sides) without infinitely large sample sizes? Conversely, how do we design studies to avoid major pitfalls?
|
Methods and data analysis
-
Q5:
What analytic strategies might be most useful at this point in investigating G-E interplay? Can multiple strategies be combined in a single “proof-of-principle” study?
-
Q6:
How do we integrate more complex environmental measures into our models? How do we approach incorporating different non-discrete environmental variables? What statistical/computational methods are needed to integrate these disparate data streams?
-
Q7:
What level of mechanistic understanding is needed to verify G or E ‘hits’ before follow up in G × E studies?
-
Q8:
What statistical tools and resources are needed?
|
Phenotypes, endophenotypes and other variables
-
Q9:
What are the characteristics of end-points or variables that are “ready to go” for G × E studies? Are there specific diseases, traits, biological phenotypes, or environmental exposures that currently meet these characteristics?
-
Q10:
Should we focus on complex phenotypes, or search for associations to the underlying mechanisms or intermediate/endophenotypes?
-
Q11:
How do we integrate variables in G–E studies, many of which are interdependent, that incorporate a comprehensive view of “environment”?
-
Q12:
What are the best strategies to measure environmental variables and exposures in large cohorts? What is needed to incorporate next-generation tools to scale up to large epidemiological studies?
|
