We are going over the fifth version(*) of a manuscript with a Ph.D. student and colleagues, polishing some details before submission, thinking about one of the results: basically there is little, if any, Genotype by Environment (GxE) interaction.
There is a long history of studying GxE interaction in radiata pine, with parallels in the development of statistical techniques. From basic models with a few sites with common parents, to massive multivariate, factor analytic, animal (tree) model BLUP with huge levels of imbalance, showing substantial GxE in some cases.
This time, however, we have a number of extremely well connected sites, clonal replication, SNP-based pedigree, factor analytic covariance matrices, etc. And there is almost no GxE interaction for stem diameter (a low heritability trait) and wood basic density (a high heritability trait).
Could it be that a large-proportion of reported GxE interaction relates to data structure? If that’s the case, Is there much point on trying to explain the “interaction” with environmental variables? 🤔 I’m not saying that this happens all the times, but I have seen the issues quite a few times already.
(*) One of those cases where the journey—learning how to pitch the problem, emphasising the most relevant parts, etc.—is almost more important than the destination. We are training researchers 😉.