Abstrakt

What’s in a Diagnosis? A Genetic Decomposition of Major Depression

Bradley S Jermy 1,2*, Kylie P Glanville1, Jonathan RI Coleman1,2, Cathryn M Lewis1,2,3, Evangelos Vassos1,2

Determining a diagnosis of Major Depressive Disorder (MDD) is complex, involving consideration and rating of a variety of different components. These include number of symptoms over an agreed threshold, symptom duration, functional impairment, persistence of symptoms within an episode, and symptom recurrence. While these components are generally accepted amongst physicians, it is unknown whether they reflect partly distinct biology between phenotypes. The aim of this study was to investigate how the genetic aetiology varies in the presence of different MDD components.

Thirty-two depression phenotypes which systematically incorporate the MDDbull Components were created using the mental health questionnaire data within the UK Biobank. SNP-based heritabilities and genetic correlations with three previously defined major depression phenotypes were calculated (broad depression, Psychiatric Genomics Consortium (PGC) defined depression and 23andMe, Inc. self-reported depression) and differences between estimates analyzed.

All phenotypes were heritabilities (h2 SNP range: 0.102–0.162) and showed substantial genetic correlations with other major depression phenotypes (Rg range: 0.651–0.894 (PGC); 0.652–0.837 (23andMe); 0.699–0.900 (broad depression)). The requirement for 5 or more symptoms and for long episode duration had the strongest effect on SNP-based heritability, in the positive and negative direction respectively (1.4% average increases; 2.7% average decrease). No significant differences were noted between genetic correlations.

While there is some variation, the two cardinal symptoms, depressed mood and anhedonia, largely reflect the genetic aetiology of phenotypes incorporating more MDD components. These components may appropriately index for severity; however, the genetic component between phenotypes incorporating none and all components is comparable.