A related issue in developing and testing cGxE hypotheses involves the coding of the genotype. Although cGxE analyses are statistical interactions, it is important that the coding of the genotype reflect our understanding of the biology of the underlying gene function; that is, genotypic groups should be collapsed only in instances where it is biologically justifiable. Schlomer et al. discuss the ambiguity of coding DRD4, and several of the issues that must be considered when deciding on a genetic model. They draw on functional and gene expression data and the precedent set from the previous literature on differential susceptibility theory in order to justify use of a 7+ versus 7− coding scheme for DRD4 genotype. This is a reasonable approach in the context of this gene, and probably the best possible strategy; however, the ambiguity surrounding the gene coding and function remains a limitation. For many genetic variants there is not clear functional or expression evidence. Thus, although preliminary analyses may suggest that certain genotypic groups can be collapsed in order to enhance power, this is not typically a justifiable approach because the statistical model may mismatch the underlying biology and lead to erroneous results.
