Genetic stratification of depression in UK Biobank

Polygenic prediction of major depressive disorder and related traits in African ancestries UK Biobank participants

Genome-Wide Association Studies (GWAS) over-represent European ancestries, neglecting all other ancestry groups and low-income nations. Consequently, polygenic risk scores (PRS) more accurately predict complex traits in Europeans than African Ancestries groups. Very few studies have looked at the transferability of European-derived PRS for behavioural and mental health phenotypes to Africans. We assessed the comparative accuracy of depression PRS trained on European and African Ancestries GWAS studies to predict major depressive disorder (MDD) and related traits in African ancestry participants from the UK Biobank. UK Biobank participants were selected based on Principal component analysis clustering with an African genetic similarity reference population, MDD was assessed with the Composite International Diagnostic Interview (CIDI). PRS were computed using PRSice2 software using either European or African Ancestries GWAS summary statistics. PRS trained on European ancestry samples (246,363 cases) predicted case control status in Africans of the UK Biobank with similar accuracies (R2 = 2%, β = 0.32, empirical p-value = 0.002) to PRS trained on far much smaller samples of African Ancestries participants from 23andMe, Inc. (5045 cases, R² = 1.8%, β = 0.28, empirical p-value = 0.008). This suggests that prediction of MDD status from Africans to Africans had greater efficiency relative to discovery sample size than prediction of MDD from Europeans to Africans. Prediction of MDD status in African UK Biobank participants using GWAS findings of likely causal risk factors from European ancestries was non-significant. GWAS of MDD in European ancestries are inefficient for improving polygenic prediction in African samples; urgent MDD studies in Africa are needed.

Distinct genetic liability profiles define clinically relevant patient strata across common diseases

Stratified medicine holds great promise to tailor treatment to the needs of individual patients. While genetics holds great potential to aid patient stratification, it remains a major challenge to operationalize complex genetic risk factor profiles to deconstruct clinical heterogeneity. Contemporary approaches to this problem rely on polygenic risk scores (PRS), which provide only limited clinical utility and lack a clear biological foundation. To overcome these limitations, we develop the CASTom-iGEx approach to stratify individuals based on the aggregated impact of their genetic risk factor profiles on tissue specific gene expression levels. The paradigmatic application of this approach to coronary artery disease or schizophrenia patient cohorts identified diverse strata or biotypes. These biotypes are characterized by distinct endophenotype profiles as well as clinical parameters and are fundamentally distinct from PRS based groupings. In stark contrast to the latter, the CASTom-iGEx strategy discovers biologically meaningful and clinically actionable patient subgroups, where complex genetic liabilities are not randomly distributed across individuals but rather converge onto distinct disease relevant biological processes. These results support the notion of different patient biotypes characterized by partially distinct pathomechanisms. Thus, the universally applicable approach presented here has the potential to constitute an important component of future personalized medicine paradigms.

Night shift work, genetic risk, and the risk of depression: A prospective cohort study

BACKGROUND

Genetic factors and night shift work both contribute to the risk of depression, but whether the association of night shift work with depression varies by genetic predisposition remains unclear.

OBJECTIVES

To assess whether night shift work is associated with a higher risk of depression regardless of genetic predisposition.

METHODS

We used data from the UK biobank of 247,828 adults aged 38-71 free of depression at baseline from March 13, 2006, to October 1, 2010. Genetic predisposition to depression was assessed using polygenic risk scores (PRS) weighted sums of genetic variant indicator variables and classified as low (lowest tertile), intermediate (tertile 2), and high (highest tertile). Night shift work exposures were collected using a touchscreen questionnaire and were divided into four categories.

RESULTS

After a median follow-up of 12.7 years, 7315 participants developed depression. Compared with day workers, HRs (95 % CIs) of depression were 1.28 (1.19-1.38) for shift work, but never or rarely night shifts, 1.32 (1.20-1.45) for irregular night shifts, and 1.20 (1.07-1.34) for permanent night shifts. Considering lifetime employment and compared with never shift workers, >8 nights/month (HR: 1.40; 95 % CI: 1.19-1.66) and <10 years (HR: 1.30; 95 % CI: 1.09-1.54) of night shift work were associated with a higher risk of depression. In joint effect analyses, compared to participants with low genetic predisposition and day workers, the HRs (95 % CIs) of depression were 1.49 (1.32-1.69) in those with high genetic predisposition and shift work, but never or rarely night shifts, and 1.36 (1.20-1.55) for those with high genetic predisposition and irregular/permanent night shifts. In addition, there was neither multiplicative nor additive interaction between genetic predisposition and night shift work on the risk of depression.

CONCLUSIONS

Night shift work was associated with an increased risk of depression regardless of genetic risk.

Polygenic prediction of major depressive disorder and related traits in African ancestries UK Biobank participants

Introduction

Genome-Wide Association Studies (GWAS) over-represent European ancestries compared to the global population, neglecting all other ancestry groups and low-income nations. Consequently, polygenic risk scores (PRS) more accurately predict complex traits in Europeans than African Ancestries groups. Very few studies have looked at the transferability of European-derived PRS for behavioural and mental health phenotypes to non-Europeans. We assessed the comparative accuracy of PRS for Major Depressive Disorder (MDD) trained on European and African Ancestries GWAS studies to predict MDD and related traits in African Ancestries participants from the UK Biobank.

Methods

UK Biobank participants were selected based on Principal component analysis (PCA) clustering with an African genetic similarity reference population and MDD was assessed with the Composite International Diagnostic Interview (CIDI). Polygenic Risk Scores (PRS) were computed using PRSice2 using either European or African Ancestries GWAS summary statistics.

Results

PRS trained on European ancestry samples (246,363 cases) predicted case control status in Africans of the UK Biobank with similar accuracies (190 cases, R2=2%) to PRS trained on far much smaller samples of African Ancestries participants from 23andMe, Inc. (5045 cases, R2=1.8%). This suggests that prediction of MDD status from Africans to Africans had greater efficiency per unit increase in the discovery sample size than prediction of MDD from Europeans to Africans. Prediction of MDD status in African UK Biobank participants using GWAS findings of causal risk factors from European ancestries was non-significant.

Conclusion

GWAS studies of MDD in European ancestries are an inefficient means of improving polygenic prediction accuracy in African samples.

Trustworthy Data and AI Environments for Clinical Prediction: Application to Crisis-Risk in People With Depression

Depression is a common mental health condition that often occurs in association with other chronic illnesses, and varies considerably in severity. Electronic Health Records (EHRs) contain rich information about a patient's medical history and can be used to train, test and maintain predictive models to support and improve patient care. This work evaluated the feasibility of implementing an environment for predicting mental health crisis among people living with depression based on both structured and unstructured EHRs. A large EHR from a mental health provider, Mersey Care, was pseudonymised and ingested into the Natural Language Processing (NLP) platform CogStack, allowing text content in binary clinical notes to be extracted. All unstructured clinical notes and summaries were semantically annotated by MedCAT and BioYODIE NLP services. Cases of crisis in patients with depression were then identified. Random forest models, gradient boosting trees, and Long Short-Term Memory (LSTM) networks, with varying feature arrangement, were trained to predict the occurrence of crisis. The results showed that all the prediction models can use a combination of structured and unstructured EHR information to predict crisis in patients with depression with good and useful accuracy. The LSTM network that was trained on a modified dataset with only 1000 most-important features from the random forest model with temporality showed the best performance with a mean AUC of 0.901 and a standard deviation of 0.006 using a training dataset and a mean AUC of 0.810 and 0.01 using a hold-out test dataset. Comparing the results from the technical evaluation with the views of psychiatrists shows that there are now opportunities to refine and integrate such prediction models into pragmatic point-of-care clinical decision support tools for supporting mental healthcare delivery.

The genetic basis of major depressive disorder

The genetic dissection of major depressive disorder (MDD) ranks as one of the success stories of psychiatric genetics, with genome-wide association studies (GWAS) identifying 178 genetic risk loci and proposing more than 200 candidate genes. However, the GWAS results derive from the analysis of cohorts in which most cases are diagnosed by minimal phenotyping, a method that has low specificity. I review data indicating that there is a large genetic component unique to MDD that remains inaccessible to minimal phenotyping strategies and that the majority of genetic risk loci identified with minimal phenotyping approaches are unlikely to be MDD risk loci. I show that inventive uses of biobank data, novel imputation methods, combined with more interviewer diagnosed cases, can identify loci that contribute to the episodic severe shifts of mood, and neurovegetative and cognitive changes that are central to MDD. Furthermore, new theories about the nature and causes of MDD, drawing upon advances in neuroscience and psychology, can provide handles on how best to interpret and exploit genetic mapping results.

Distinct genetic liability profiles define clinically relevant patient strata across common diseases

Genome-wide association studies have unearthed a wealth of genetic associations across many complex diseases. However, translating these associations into biological mechanisms contributing to disease etiology and heterogeneity has been challenging. Here, we hypothesize that the effects of disease-associated genetic variants converge onto distinct cell type specific molecular pathways within distinct subgroups of patients. In order to test this hypothesis, we develop the CASTom-iGEx pipeline to operationalize individual level genotype data to interpret personal polygenic risk and identify the genetic basis of clinical heterogeneity. The paradigmatic application of this approach to coronary artery disease and schizophrenia reveals a convergence of disease associated variant effects onto known and novel genes, pathways, and biological processes. The biological process specific genetic liabilities are not equally distributed across patients. Instead, they defined genetically distinct groups of patients, characterized by different profiles across pathways, endophenotypes, and disease severity. These results provide further evidence for a genetic contribution to clinical heterogeneity and point to the existence of partially distinct pathomechanisms across patient subgroups. Thus, the universally applicable approach presented here has the potential to constitute an important component of future personalized medicine concepts.

Accelerated biological aging and risk of depression and anxiety: evidence from 424,299 UK Biobank participants

Theory predicts that biological processes of aging may contribute to poor mental health in late life. To test this hypothesis, we evaluated prospective associations between biological age and incident depression and anxiety in 424,299 UK Biobank participants. We measured biological age from clinical traits using the KDM-BA and PhenoAge algorithms. At baseline, participants who were biologically older more often experienced depression/anxiety. During a median of 8.7 years of follow-up, participants with older biological age were at increased risk of incident depression/anxiety (5.9% increase per standard deviation [SD] of KDM-BA acceleration, 95% confidence intervals [CI]: 3.3%-8.5%; 11.3% increase per SD of PhenoAge acceleration, 95% CI: 9.%-13.0%). Biological-aging-associated risk of depression/anxiety was independent of and additive to genetic risk measured by genome-wide-association-study-based polygenic scores. Advanced biological aging may represent a potential risk factor for incident depression/anxiety in midlife and older adults and a potential target for risk assessment and intervention.

Symptom dimensions of major depression in a large community-based cohort

BACKGROUND

Our understanding of major depression is complicated by substantial heterogeneity in disease presentation, which can be disentangled by data-driven analyses of depressive symptom dimensions. We aimed to determine the clinical portrait of such symptom dimensions among individuals in the community.

METHODS

This cross-sectional study consisted of 25 261 self-reported White UK Biobank participants with major depression. Nine questions from the UK Biobank Mental Health Questionnaire encompassing depressive symptoms were decomposed into underlying factors or 'symptom dimensions' via factor analysis, which were then tested for association with psychiatric diagnoses and polygenic risk scores for major depressive disorder (MDD), bipolar disorder and schizophrenia. Replication was performed among 655 self-reported non-White participants, across sexes, and among 7190 individuals with an ICD-10 code for MDD from linked inpatient or primary care records.

RESULTS

Four broad symptom dimensions were identified, encompassing negative cognition, functional impairment, insomnia and atypical symptoms. These dimensions replicated across ancestries, sexes and individuals with inpatient or primary care MDD diagnoses, and were also consistent among 43 090 self-reported White participants with undiagnosed self-reported depression. Every dimension was associated with increased risk of nearly every psychiatric diagnosis and polygenic risk score. However, while certain psychiatric diagnoses were disproportionately associated with specific symptom dimensions, the three polygenic risk scores did not show the same specificity of associations.

CONCLUSIONS

An analysis of questionnaire data from a large community-based cohort reveals four replicable symptom dimensions of depression with distinct clinical, but not genetic, correlates.

Association Between Polygenic Risk Scores and Outcome of ECT

OBJECTIVE

Identifying biomarkers associated with response to electroconvulsive therapy (ECT) may aid clinical decisions. The authors examined whether greater polygenic liabilities for major depressive disorder, bipolar disorder, and schizophrenia are associated with improvement following ECT for a major depressive episode.

METHODS

Between 2013 and 2017, patients who had at least one treatment series recorded in the Swedish National Quality Register for ECT were invited to provide a blood sample for genotyping. The present study included 2,320 participants (median age, 51 years; 62.8% women) who had received an ECT series for a major depressive episode (77.1% unipolar depression), who had a registered treatment outcome, and whose polygenic risk scores (PRSs) could be calculated. Ordinal logistic regression was used to estimate the effect of PRS on Clinical Global Impressions improvement scale (CGI-I) score after each ECT series.

RESULTS

Greater PRS for major depressive disorder was significantly associated with less improvement on the CGI-I (odds ratio per standard deviation, 0.89, 95% CI=0.82, 0.96; R²=0.004), and greater PRS for bipolar disorder was associated with greater improvement on the CGI-I (odds ratio per standard deviation, 1.14, 95% CI=1.05, 1.23; R²=0.005) after ECT. PRS for schizophrenia was not associated with improvement. In an overlapping sample (N=1,207) with data on response and remission derived from the self-rated version of the Montgomery-Åsberg Depression Rating Scale, results were similar except that schizophrenia PRS was also associated with remission.

CONCLUSIONS

Improvement after ECT is associated with polygenic liability for major depressive disorder and bipolar disorder, providing evidence of a genetic component for ECT clinical response. These liabilities may be considered along with clinical predictors in future prediction models of ECT outcomes.

Joint exposure to positive affect, life satisfaction, broad depression, and neuroticism and risk of cardiovascular diseases: A prospective cohort study

BACKGROUND AND AIMS

Psychologic wellbeing can impact cardiovascular health. We aimed to evaluate the joint association of multiple psychologic wellbeing factors with cardiovascular diseases (CVD) and examine whether this association was modified by genetic susceptibility.

METHODS

In the UK Biobank, 126,255 participants free of CVD (coronary heart disease [CHD], stroke, and heart failure [HF]) at baseline, who completed a questionnaire on psychological factors, were included. The psychological wellbeing score was calculated by four factors: happiness, life satisfaction, broad depression, and neuroticism. Cox proportional hazard models were used to assess the association between the psychological wellbeing score and CVD risk.

RESULTS

During the median follow-up of 11.5 years, 10,815 participants had newly diagnosed CVDs. Low life satisfaction, the presence of depression, and neuroticism score ≥1 were significantly associated with an increased risk of CVD in the multivariable-adjusted model. Through decreasing the psychological wellbeing score, there were significant increasing linear trends in the risk of CVD, CHD, stroke, and HF (all p for trend < 0.001). Participants with the lowest psychological wellbeing score had the highest risk for CVD (HR 1.51, 95% CI 1.42-1.61). Women were more susceptible to worse psychological wellbeing status for CVD than men (p for interaction = 0.009). The associations of the psychological wellbeing score with CVD were consistent across genetic risk (p for interaction >0.05). When considered jointly, participants exposed to high-risk psychological wellbeing and genetic status had a 2.70-fold (95% CI 2.25-3.24) risk for CHD.

CONCLUSIONS

Joint exposure to multiple psychological wellbeing factors was associated with increased risks of incident CVD in an additive manner, regardless of genetic susceptibility.

Joint Exposure to Positive Affect, Life Satisfaction, Depressive Symptoms, and Neuroticism and Incident Type 2 Diabetes

CONTEXT

Whether the psychological wellbeing status could be a risk factor for type 2 diabetes is unclear.

OBJECTIVE

We aimed to measure the association between combined psychological wellbeing factors and type 2 diabetes and investigate whether this association was modified by genetic predisposition.

METHODS

Prospective cohort study from the UK Biobank. In total, 127 496 participants who completed a psychological wellbeing questionnaire and did not have type 2 diabetes at baseline (2006-2010) were included; among them, 88 584 (69.5%) were analyzed to determine their genetic predisposition. The main outcome measure was incident type 2 diabetes.

RESULTS

During the median follow-up of 10.0 years, 2547 incident type 2 diabetes cases were documented. Moderate to extreme unhappiness, satisfaction score ≤3, presence of broad depression, and a neuroticism score ≥3 were all significantly and independently associated with an increased risk of diabetes. When considered as a combination indicator, compared with individuals in the highest quartile of the psychological wellbeing score, the fully adjusted hazard ratios (95% CI) of type 2 diabetes were 1.41 (1.21-1.65) in the third quartile, 1.45 (1.24-1.69) in the second quartile, and 1.73 (1.48-2.01) in the lowest quartile. In the stratified analysis, we observed significant interactions between age and physical activity, and type 2 diabetes (Pinteraction < .001 and 0.049, respectively). However, there was no significant interaction between the psychological wellbeing score and genetic susceptibility to diabetes (Pinteraction = .980).

CONCLUSION

Worse overall psychological wellbeing was associated with a significantly increased risk of type 2 diabetes in a dose-response fashion regardless of genetic predisposition.

Identification of transdiagnostic psychiatric disorder subtypes using unsupervised learning

Psychiatric disorders show heterogeneous symptoms and trajectories, with current nosology not accurately reflecting their molecular etiology and the variability and symptomatic overlap within and between diagnostic classes. This heterogeneity impedes timely and targeted treatment. Our study aimed to identify psychiatric patient clusters that share clinical and genetic features and may profit from similar therapies. We used high-dimensional data clustering on deep clinical data to identify transdiagnostic groups in a discovery sample (N = 1250) of healthy controls and patients diagnosed with depression, bipolar disorder, schizophrenia, schizoaffective disorder, and other psychiatric disorders. We observed five diagnostically mixed clusters and ordered them based on severity. The least impaired cluster 0, containing most healthy controls, showed general well-being. Clusters 1-3 differed predominantly regarding levels of maltreatment, depression, daily functioning, and parental bonding. Cluster 4 contained most patients diagnosed with psychotic disorders and exhibited the highest severity in many dimensions, including medication load. Depressed patients were present in all clusters, indicating that we captured different disease stages or subtypes. We replicated all but the smallest cluster 1 in an independent sample (N = 622). Next, we analyzed genetic differences between clusters using polygenic scores (PGS) and the psychiatric family history. These genetic variables differed mainly between clusters 0 and 4 (prediction area under the receiver operating characteristic curve (AUC) = 81%; significant PGS: cross-disorder psychiatric risk, schizophrenia, and educational attainment). Our results confirm that psychiatric disorders consist of heterogeneous subtypes sharing molecular factors and symptoms. The identification of transdiagnostic clusters advances our understanding of the heterogeneity of psychiatric disorders and may support the development of personalized treatments.

Shared genetic architecture across psychiatric disorders

Psychiatric disorders overlap substantially at the genetic level, with family-based methods long pointing toward transdiagnostic risk pathways. Psychiatric genomics has progressed rapidly in the last decade, shedding light on the biological makeup of cross-disorder risk at multiple levels of analysis. Over a hundred genetic variants have been identified that affect multiple disorders, with many more to be uncovered as sample sizes continue to grow. Cross-disorder mechanistic studies build on these findings to cluster transdiagnostic variants into meaningful categories, including in what tissues or when in development these variants are expressed. At the upper-most level, methods have been developed to estimate the overall shared genetic signal across pairs of traits (i.e. single-nucleotide polymorphism-based genetic correlations) and subsequently model these relationships to identify overarching, genomic risk factors. These factors can subsequently be associated with external traits (e.g. functional imaging phenotypes) to begin to understand the makeup of these transdiagnostic risk factors. As psychiatric genomic efforts continue to expand, we can begin to gain even greater insight by including more fine-grained phenotypes (i.e. symptom-level data) and explicitly considering the environment. The culmination of these efforts will help to inform bottom-up revisions of our current nosology.

Causal Effects of Positive Affect, Life Satisfaction, Depressive Symptoms, and Neuroticism on Kidney Function: A Mendelian Randomization Study

BACKGROUND

Further investigation of the causal effects of psychologic wellbeing on kidney function is warranted.

METHODS

In this Mendelian randomization (MR) study, genetic instruments for positive affect, life satisfaction, depressive symptoms, and neuroticism were introduced from a previous genome-wide association study meta-analysis of European individuals. Summary-level MR was performed using the CKDGen data of European ancestry (n=567,460), and additional allele score-based MR was performed in the individual-level data of White British UK Biobank participants (n=321,024).

RESULTS

In summary-level MR with the CKDGen data, depressive symptoms were a significant causative factor for kidney function impairment (CKD OR, 1.45; 95% confidence interval, 1.07 to 1.96; eGFR change [%] beta -2.18; 95% confidence interval, -3.61 to -0.72) and pleiotropy-robust sensitivity analysis results supported the causal estimates. A genetic predisposition for positive affect was significantly associated with better kidney function (CKD OR, 0.69; 95% confidence interval, 0.52 to 0.91), eGFR change [%] beta 1.50; 95% confidence interval, 0.09 to 2.93) and sensitivity MR analysis results supported the finding for CKD outcome, but was nonsignificant for eGFR. Life satisfaction and neuroticism exposures showed nonsignificant causal estimates. In the UK Biobank with covariate-adjusted allele score MR analysis, allele scores for positive affect and life satisfaction were causally associated with reduced risk of CKD and higher eGFR. In contrast, neuroticism allele score was associated with increased risk of CKD and lower eGFR, and depressive symptoms allele score was associated with lower eGFR, but showed nonsignificant association with CKD.

CONCLUSIONS

Health care providers in the nephrology field should be aware of the causal linkage between psychologic wellbeing and kidney function.

Polygenic risk score, healthy lifestyles, and risk of incident depression

Genetic factors increase the risk of depression, but the extent to which this can be offset by modifiable lifestyle factors is unknown. We investigated whether a combination of healthy lifestyles is associated with lower risk of depression regardless of genetic risk. Data were obtained from the UK Biobank and consisted of 339,767 participants (37-73 years old) without depression between 2006 and 2010. Genetic risk was categorized as low, intermediate, or high according to polygenic risk score for depression. A combination of healthy lifestyles factors-including no current smoking, regular physical activity, a healthy diet, moderate alcohol intake and a body mass index <30 kg/m²-was categorized into favorable, intermediate, and unfavorable lifestyles. The risk of depression was 22% higher among those at high genetic risk compared with those at low genetic risk (HR = 1.22, 95% CI: 1.14-1.30). Participants with high genetic risk and unfavorable lifestyle had a more than two-fold risk of incident depression compared with low genetic risk and favorable lifestyle (HR = 2.18, 95% CI: 1.84-2.58). There was no significant interaction between genetic risk and lifestyle factors (P for interaction = 0.69). Among participants at high genetic risk, a favorable lifestyle was associated with nearly 50% lower relative risk of depression than an unfavorable lifestyle (HR = 0.51, 95% CI: 0.43-0.60). We concluded that genetic and lifestyle factors were independently associated with risk of incident depression. Adherence to healthy lifestyles may lower the risk of depression regardless of genetic risk.

Multiple measures of depression to enhance validity of major depressive disorder in the UK Biobank

BACKGROUND

The UK Biobank contains data with varying degrees of reliability and completeness for assessing depression. A third of participants completed a Mental Health Questionnaire (MHQ) containing the gold-standard Composite International Diagnostic Interview (CIDI) criteria for assessing mental health disorders.

AIMS

To investigate whether multiple observations of depression from sources other than the MHQ can enhance the validity of major depressive disorder (MDD).

METHOD

In participants who did not complete the MHQ, we calculated the number of other depression measures endorsed, for example from hospital episode statistics and interview data. We compared cases defined this way with CIDI-defined cases for several estimates: the variance explained by polygenic risk scores (PRS), area under the curve attributable to PRS, single nucleotide polymorphisms (SNPs)-based heritability and genetic correlations with summary statistics from the Psychiatric Genomics Consortium MDD genome-wide association study.

RESULTS

The strength of the genetic contribution increased with the number of measures endorsed. For example, SNP-based heritability increased from 7% in participants who endorsed only one measure of depression, to 21% in those who endorsed four or five measures of depression. The strength of the genetic contribution to cases defined by at least two measures approximated that for CIDI-defined cases. Most genetic correlations between UK Biobank and the Psychiatric Genomics Consortium MDD study exceeded 0.7, but there was variability between pairwise comparisons.

CONCLUSIONS

Multiple measures of depression can serve as a reliable approximation for case status where the CIDI measure is not available, indicating sample size can be optimised using the entire suite of UK Biobank data.

Reviewing the genetics of heterogeneity in depression: operationalizations, manifestations and etiologies

With progress in genome-wide association studies of depression, from identifying zero hits in ~16 000 individuals in 2013 to 223 hits in more than a million individuals in 2020, understanding the genetic architecture of this debilitating condition no longer appears to be an impossible task. The pressing question now is whether recently discovered variants describe the etiology of a single disease entity. There are a myriad of ways to measure and operationalize depression severity, and major depressive disorder as defined in the Diagnostic and Statistical Manual of Mental Disorders-5 can manifest in more than 10 000 ways based on symptom profiles alone. Variations in developmental timing, comorbidity and environmental contexts across individuals and samples further add to the heterogeneity. With big data increasingly enabling genomic discovery in psychiatry, it is more timely than ever to explicitly disentangle genetic contributions to what is likely 'depressions' rather than depression. Here, we introduce three sources of heterogeneity: operationalization, manifestation and etiology. We review recent efforts to identify depression subtypes using clinical and data-driven approaches, examine differences in genetic architecture of depression across contexts, and argue that heterogeneity in operationalizations of depression is likely a considerable source of inconsistency. Finally, we offer recommendations and considerations for the field going forward.