The genetic and environmental relationship betweenmajor depression and the five-factor modelof personalityK. S. Kendler1,2,3* and J. Myers1,21Virginia Institute for Psychiatric and Behavioral Genetics and Departments of2Psychiatry and3Human and Molecular Genetics,Virginia Commonwealth University, Richmond, VA, USABackground. Certain personality traits have long been suspected to reflect an enduring vulnerability to major de-pression (MD) in part because of shared genetic risk factors. Although many have agreed that normative personalityis well captured by the ‘ Big-Five’ personality traits of Openness (O), Conscientiousness (C), Extraversion (E),Agreeableness (A) and Neuroticism (N), to date genetically informative studies have only examined the relationshipbetween MD and N and E.Method. Questionnaires were completed on a website, yielding a sample of 44 112 subjects including both membersof 542 same-sex twin pairs. Personality was measured by the Big Five Inventory. Structural modeling was performedby Mx.Results. Three of the big-five personality traits – O, E and A – had small phenotypic associations with risk for MDand small genetic correlations. Two traits – N and C – had stronger phenotypic associations (positive for N andnegative for C) with the following estimates of the genetic correlation with MD : +0.43 for N and x0.36 for C. N andC were moderately negatively correlated. Controlling for N reduced the genetic correlation between C and MD morethan controlling for C reduced the genetic correlation between N and MD.Conclusions. A large proportion of the genetic risk for MD that is expressed via personality is captured by N, with amodest amount due to C, and small amounts from O, E and A.Received 23 October 2008 ; Revised 16 June 2009 ; Accepted 24 June 2009 ; First published online 7 September 2009Key words : Five-factor model, genetics, major depression, personality, twin studies.IntroductionA long tradition of research has examined the inter-relationship between personality and major de-pression (MD) with a leading hypothesis being thatcertain personality traits reflect an enduring vulner-ability to MD (Kendler et al. 1993a; Klein et al. 1993;Bagby et al. 1995; Enns & Cox, 1997). Given the sub-stantial evidence that genetic factors contribute to bothrisk for MD (Kendler et al. 2006a; Sullivan et al. 2000)and to variation in personality (Loehlin, 1992 ; Loehlinet al. 1998), it is of particular interest to determinethe genetic contribution to the covariation betweenpersonality and MD risk.Some consensus has developed in the last 20 yearsthat human personality can be well accounted forby five personality traits most commonly termed :Openness (O), Conscientiousness (C), Extraversion (E),Agreeableness (A) and Neuroticism (N) (McCrae,1989; Digman, 1990). However, we are aware of ge-netically informative studies that have examined therelationship only between MD and neuroticism (N)(Kendler et al. 1993a, 2006b; Fanous et al. 2007) andextraversion (E) (Kendler et al. 2006b). The presentstudy examines, for the first time to our knowledge,the genetic relationship between MD and all of the‘Big-Five’ personality traits.MethodSampleAs outlined in detail elsewhere (Kendler et al. 2009),participants in this study were part of data collectedfrom ‘Twins: an interactive personality test ’ from 1July 2005 to 1 May 2008. This survey was designedas an interactive assessment tool for measures ofpersonality, psychopathology, and substance use anddependence. The website permits any two people,* Address for correspondence : K. S. Kendler, M.D., Virginia Institutefor Psychiatric and Behavioral Genetics, Virginia CommonwealthUniversity, Box 980126, Richmond, VA 23298-0126, USA.(Email : [email protected])Psychological Medicine (2010), 40, 801–806. f Cambridge University Press 2009doi:10.1017/S0033291709991140ORIGINAL ARTICLEregardless of whether they are twins or not, to com-pare their personalities and behaviors. Participantscould take the survey as individuals. All participantswere volunteers and were recruited over the worldwide web. Potential respondents found out aboutthe site via internet search engines, direct access to itsaddress (http://www.outofservice.com/twins/), orthrough links from other sites.Data collection was done with automated com-puterized administration, data entry and scoring.All participants received individualized feedbackafter completing the survey. The data presented inthis article were collected using a non-commercial,advertisement-free website (www.outofservice.com)that contains personality measures as well as severalgames, quizzes, and questionnaires for entertainmentpurposes. Participants did not provide any identifyinginformation and anonymity was assured. This re-search obtained exempt ethics approval at VirginiaCommonwealth University.We utilized a variety of quality control measures toassess the amount of duplicate or faked responses, orfalse twin pairs in the sample (Kendler et al. 2009). Thisincluded examining distributions of our personalitymeasures and finding no excess of extreme scores,examining similarity of reported year of birth, heightand weight in twin pairs, and asking and following upthe small number of positive responses to an item inour questionnaire about duplicate entries. Consistentwith other reports of internet samples (Gosling et al.2004), these investigations suggest quite low levels offaked or duplicate data.Zygosity was assessed by responses in both twinpairs to the three items found most discriminatingwhen tested against DNA results in the VirginiaAdult Twin Study of Psychiatric and Substance UseDisorders (VATSPSUD ; Kendler & Prescott, 2006).Pairs, when the responses of the two twins wereinconsistent, were eliminated from the study (n=9).AssessmentsPersonality was assessed by the Big Five InventoryPersonality Test (John & Srivastava, 1999), a 44-itemscale that assesses O, C, E, A, and N. Responsesto these items are recorded by a 5-point Likert scale :(1) strongly disagree, (2) disagree a little, (3) neitherdisagree nor agree, (4) agree a little, and (5) stronglyagree. In the entire sample, these variables were rela-tively normally distributed with estimates of skewnessranging from x0.50 for A, to+0.04 for N.Lifetime MD was assessed by self-report using aquestionnaire validated with a Swedish mailed survey(Kendler et al. 1993b). This questionnaire containedan expanded version, adapted to self-report, of thesection