Applying polygenic risk scores to postpartum depression

Abstract

Objective The etiology of major depressive disorder (MDD) is likely to be heterogeneous, but postpartum depression (PPD) is hypothesized to represent a more homogenous subset of MDD. We use genome-wide SNP data to explore this hypothesis. Method We assembled a total cohort of 1,420 self-report cases of PPD and 9,473 controls with genome-wide genotypes from Australia, the Netherlands, Sweden and the United Kingdom. We estimated the total variance attributable to genotyped variants. We used association results from the Psychiatric Genomics Consortia (PGC) of Bipolar Disorder (BPD) and MDD to create polygenic scores in PPD and related MDD data sets to estimate the genetic overlap between the disorders. Results We estimated that the percentage of variance on the liability scale explained by common genetic variants to be 0.22 with a standard error of 0.12, p = 0.02. The R2 from a logistic regression of PPD case-control status in all four cohorts on a SNP profile score weighted by PGC-BPD association results was small (0.1%) but significant (p= 0.004) indicating a genetic overlap between BPD and PPD. The results were highly significant in the Australian and Dutch cohorts (R2 > 1.1%, p < 0.008), where the majority of cases met criteria for MDD. The genetic overlap between BPD and MDD was not significant in larger Australian and Dutch MDD case-control cohorts after excluding PPD cases (R2 =0.06%, p= 0.08), despite the larger MDD group affording more power. Conclusions Our results suggest empirical genetic evidence for a more important shared genetic etiology between BPD and PPD, than between BPD and MDD.

Is Postpartum Depression a Distinct Disorder?

“…Second, a recent Swedish study conducted on 3427 female twins who completed a self-rated scale for screening PPD and a population-based cohort of 580,006 sisters with clinical diagnosis of PPD found that 14 % of the variance (one third of the total hereditability estimate) of PPD was explained by genetic factors not shared with non-puerperal major depression [20•]. Third, by applying polygenic risk scores of bipolar disorder and major depression from the Psychiatric Genomics Consortium datasets, Byrne et al found a stronger genetic overlap between bipolar disorder and self-reported PPD, than between bipolar disorder and major depression outside the postpartum period [21]. However, the relatively small sample size of this study (1420 cases and 9473 controls) and consequent large standard errors limited the interpretation of the findings.…”

Bipolar II disorder as a risk factor for postpartum depression

“…In support for a bipolar diathesis, in women with PDD the following interesting features were reported: early age at onset (similar to that observed in BD) (Akdeniz et al, 2003;Azorin et al, 2012;Garvey et al, 1983;Viguera et al, 2011), family history of BD (Azorin et al, 2012;Munk-Olsen et al, 2007), familial transmission of PDD within BD pedigrees (Jones and Craddock, 2001;Payne et al, 2008) and, possibly, a shared genetic etiology (Byrne et al, 2014;Dallaspezia et al, 2011;Jones and Craddock, 2007).…”

Polygenic scores via penalized regression on summary statistics

“…The vast majority of such genetic contribution is thus spread across the huge landscape of the genome, with many loci each contributing a small, almost undetectable effect on the phenotypes (Dudbridge, 2013(Dudbridge, , 2016. One important source of evidence towards this conclusion is from studies that examined the association of polygenic predictors of diseases/traits, where it has been repeatedly found that SNPs that are not themselves significantly associated with the phenotypes can, by being aggregated as a score, be very significantly associated with the phenotypes in different samples (Agerbo et al, 2015;Byrne et al, 2014;Evans et al, 2009;Wei et al, 2009;Purcell et al, 2009;Ripke et al, 2013;Speliotes et al, 2010;Machiela et al, 2011;Stahl et al, 2012;Martin et al, 2015;Chang et al, 2014). A particular remarkable demonstration is that persons with such polygenic scores for schizophrenia at the top 10 percentile of the population can be at more than 10 times the risk of having the disease than those at the bottom 10 percentile (Ripke et al, 2014;Agerbo et al, 2015), raising hope that one day a person's risk for many common disease can be accurately assessed simply by the examination of one's genome.…”

“…10, 2016; of SNPs to be included, the optimal p-value cutoff is generally unknown (Wray et al, 2014). As a result, in many studies, PGS are constructed using a number of thresholds (Purcell et al, 2009;Ripke et al, 2014;Byrne et al, 2014;Martin et al, 2015;Chang et al, 2014), and there is at least one piece of software developed to facilitate this . Generally, we focus on the p-value threshold that achieves the highest correlation/association with the phenotypes in a validation dataset that contains a measure of the phenotype under study.…”

“…Euesden et al, 2015;Byrne et al, 2014;Agerbo et al, 2015;Dudbridge, 2013), and allelic scores (Burgess and Thompson, 2013;Evans et al, 2013). In a typical application, a unique PGS is assigned to each individual based on the person's genotype.…”