Examining the relevance of candidate genes for major depression

Major Depressive Disorder (MDD) is substantially heritable, but its genetic architecture is complex, and identifying specific molecular genetic polymorphisms underlying MDD susceptibility has been difficult, even compared to... [ view full abstract ]

Major Depressive Disorder (MDD) is substantially heritable, but its genetic architecture is complex, and identifying specific molecular genetic polymorphisms underlying MDD susceptibility has been difficult, even compared to other complex psychiatric disorders. Whether or not the candidate gene approach has aided our understanding of MDD, particularly in the context of genome-wide association study results, is controversial and polarizing among behavioral scientists. The current research, comprising the most comprehensive and well-powered investigation of candidate polymorphism, candidate gene, and candidate gene-by-environment interaction hypotheses in MDD to date, is in direct response to this ongoing controversy. We focus on three lines of inquiry concerning how these polymorphisms may impact MDD liability:

  1. additive effects of the most commonly studied polymorphisms;

  2. moderation of polymorphism effects by environmental exposure;

  3. additive effects at the whole gene level.

We first empirically identified the 18 most commonly studied candidate genes in MDD research between 1991 and 2016 from the corpus of scientific publications indexed in the PubMed database. Within these regions, we determined the most commonly studied variants and their canonical risk alleles. Using data from the UK Biobank initial touchscreen interview and online mental health follow-up, we examined multiple measures of MDD (e.g., lifetime diagnostic status, symptom severity among individuals reporting mood disturbances, lifetime number of depressive episodes), employing multiple statistical frameworks and, in interaction analyses, considering multiple indices of environmental exposure. Our results aim to directly address the following question: do the large datasets of the whole-genome data era lend any support to previous canonical candidate gene hypotheses?