Helen,
Thanks for these comments. Measurement of childhood SES is a major challenge because you can either link back to administrative records (e.g., birthweight, school records, but generally limited options), use a physical measure such as anthropometrics that likely reflects childhood conditions (which is indicative but is quite crude), or ask people. You can ask people objective facts, like parent's education, which they may or may not be able to accurately report, or subjective perceptions, like standing in the community. There's reason to believe that both the objective and subjective measures of adult SES are relevant, so also plausible that both measures in childhood are relevant. But the concerns about retrospective reporting of subjective childhood SES seem more troubling to me than retrospective reporting of objective indicators.
Regarding GxE - I want to be very clear that GxE interactions can contribute to racial disparities in health even if there are no racial differences in minor allele frequencies whatsoever. If the frequency of a genetic variant that makes eating Big Macs and fries (call it FTO) is equally prevalent in whites and blacks, but residential segregation makes it much more likely that blacks have convenient access to McDonald's offerings, then the GxE will differentially affect obesity in blacks. The FTO variant is most relevant if you live in an obesogenic environment, and disadvantaged groups are more likely to live in obesogenic environments. You can link this back to fundamental cause theory: one way people use social advantages such as money, networks, or education is to overcome genetic vulnerabilities.
Here's an article you might find relevant.
Liu SY, Walter S, Marden J, Rehkopf DH, Kubzansky LD, Nguyen T, Glymour MM. Genetic vulnerability to diabetes and obesity: does education offset the risk?. Social science & medicine. 2015 Feb 28;127:150-8. https://ucsf.idm.oclc.org/login?url=http://dx.doi.org/10.1016/j.socscimed.2014.09.009
Maria