The predictive value of polygenic risk scores for depression in gene-environment interaction studies: a systematic review

Sabrina Illius; Julian Eder; Susanne Vogel; Nina Alexander

doi:10.1038/s41398-025-03793-7

. 2026 Feb 25;16:121. doi: 10.1038/s41398-025-03793-7

The predictive value of polygenic risk scores for depression in gene-environment interaction studies: a systematic review

Sabrina Illius ^1,^2,^✉, Julian Eder ³, Susanne Vogel ^1,², Nina Alexander ^4,^5,^✉

PMCID: PMC12960946 PMID: 41741414

Abstract

Background

According to the diathesis-stress model, genetic liability and environmental exposures interact in the pathogenesis of depression. Polygenic risk scores for depression (PRS_D) based on large-scale genome-wide association studies have opened new avenues for investigating gene-environment interaction (GxE) beyond candidate gene studies. To the best of our knowledge, this is the first systematic review of studies that have taken a polygenic score approach to study GxE interaction effects on depression phenotypes.

Methods

Based on a preregistered, systematic literature search according to PRISMA guidelines, 56 studies were considered for qualitative analysis. Respective studies investigated a broad range of adverse and protective environmental exposures across the lifespan, e.g., trauma, stressful life events, social environments and (un)healthy lifestyle factors, using cross-sectional and longitudinal designs.

Results

While most studies reported significant main effects of an individual’s PRS_D and different environmental exposures on depression phenotypes, the overall evidence for GxE interactions was considerably heterogeneous. Findings of significant PRS_DxE interactions mostly stem from large cohort studies comprising > 40000 participants, in particular, when recent environmental exposures were considered.

Conclusion

Two general conclusions can be drawn from this review. First, PRS_DxE interactions, if at all, add a small amount of explained variance in depression phenotypes to the corresponding additive model and may thus require large samples to be reliably detected. Second, in a considerable number of studies, different environmental exposures were found to depend on an individual’s PRS_D, indicating significant gene-environment correlation. We further discuss limitations, future directions and potential clinical relevance of PRSxE research in depression.

graphic file with name 41398_2025_3793_Figa_HTML.jpg

Subject terms: Clinical genetics, Predictive markers

Introduction

Major depressive disorder (MDD) is one of the most common mental health disorders, with a considerable heritability estimated at approximately 30–40% [1]. The genetic basis of MDD arises from joint effects of many thousands of loci with individual odds ratios of less than 1.2 [2]. According to the diathesis-stress model, genetic liability and environmental exposures interact in the pathogenesis of MDD, thereby explaining why individuals respond so differently to adverse experiences in terms of disease outcome [3]. Early studies investigating gene-environment interactions (GxE) have exclusively focused on candidate genes with known effects on putative biological pathways of MDD. The most prominent finding of a specific GxE interaction relates to the serotonin transporter polymorphism, which has been shown to moderate the influence of life stress on MDD in an influential study [4]. However, meta-analytic evaluations of follow-up studies have seriously questioned the robustness of GxE findings based on single candidate genes [5]. Within the last decade, large-scale genome-wide association studies (GWAS) on a broad spectrum of depression phenotypes have now opened new avenues for GxE research. In 2019, a meta-analysis drawing on 807553 individuals from the three largest GWAS of depression at that time [6–8] identified 102 independent variants associated with depression, of which 87 replicated in an independent sample [9]. Recently, an even larger meta-analysis (n > 1 million participants) yielded 233 genetic variants associated with depression [10]. Analogously, first genome-wide by environment interaction studies (GWEIS) of depression and stressful life events (SLE) have already successfully identified GxE interactions in a hypothesis-free approach (e.g., [11–13]). However, GWEIS require overly large sample sizes to achieve adequate statistical power which hampers an in-depth assessment of environmental exposures and (intermediate) phenotypes of depression. To this end, polygenic risk scores (PRSs) have been advocated as a promising strategy to integrate genetic liability into a single measure, thereby allowing a hypothesis-free study of GxE effects with higher statistical power as compared to GWEIS. PRSs are commonly calculated based on a set of top-ranking genetic markers identified by prior GWAS and represent the weighted sum of their trait-associated alleles [14]. The predictive power of polygenic risk scores for depression (PRS_D) has steadily increased in parallel with rising GWAS discovery sample sizes [15]. While early GWAS from the Psychiatric Genomics Consortium (PGC) explained only around 0.5% of the variance in depression liability (n = 18759) [16], PRS_D derived from more recent large-scale GWAS now explain approximately 1.5–3.2% (n = 807553) [9]. Consequently, PRS_D are increasingly recognized as a promising approach to test the diathesis–stress model of depression [3]. For example, recent studies have now begun to address PRS_D as a potential moderator of well-established environmental risk factors of depression, including different types of childhood trauma (CT [17–31]) and other forms of early [31–38] and prenatal [39–43] adversity, as well as adult trauma/stressful life events (SLE) [23, 28, 31, 44–52], social environments [53–57], (un)healthy lifestyle factors (e.g., [58–63]) and cumulative environmental adversity [64–70]. Such GxE studies searched for departures of PRS_D and environmental exposure effects from additivity (i.e., combined effects differ from their sum) or multiplicativity (i.e., combined effects differ from the product). At the same time, recent GxE studies leveraging PRS_D face several methodological challenges, including heterogeneity in the measurement of environmental exposures and depression outcomes, lack of consensus on the optimal threshold for PRS_D construction, and potential confounding. A key source of such confounding is gene–environment correlation (rGE), the non-random association between genetic predisposition and environmental exposures, which can arise through passive, evocative, or active mechanisms [71, 72]. For example, individuals with higher genetic liability for depression may be more likely to encounter environmental adversity, thereby mimicking or inflating GxE interaction effects [27, 73].

In light of both promises and challenges of this field of research, the aim of this systematic review was to summarize and critically discuss the rapidly growing number of studies using PRS_D to investigate GxE interaction effects on depression-related outcomes. Given the substantial methodological and conceptual heterogeneity outlined above, we opted for a narrative rather than a quantitative, meta-analytic synthesis.

Material and methods

Search strategy

The systematic literature search was based on PRISMA guidelines [74] using Excel and Zotero and conducted up to May 15th, 2024 using PubMed, Embase (Ovid), MEDLINE (Ovid), Scopus, Web of Science Core Collection, Medline – Web of Science and PsycINFO. The following search terms were used: (Depression OR “Depressive Disorder” OR “Major Depression”) AND (“Polygenic Score” OR “Polygenic Risk”). All peer-reviewed journal articles released up to this date were considered. Two authors independently reviewed all identified journal articles (SI, JE). Where conflicts occurred a third author was consulted (NA).

Inclusion and exclusion criteria

We considered all English language original studies published until May 15th, 2024, that investigated interactions of PRS_D and environmental exposures (both risky and protective) on depression-related phenotypes and behaviors in healthy and clinical human samples of any ethnic origin and age. Depression-related phenotypes included intermediate phenotypes such as neural, endocrine and immunological changes associated with major depression (MD). Conference abstracts, published abstracts, preprints, editorials, reviews as well as studies that had used PRS_D to predict GxE interactions on other mental diseases (e.g., schizophrenia) were not considered in this review. We refined this data set by selecting studies that met the following inclusion criteria: i) PRS_D calculation based on whole-genome summary statistics of GWAS using definitions from the spectrum of depressive phenotypes like lifetime MDD, depressive symptoms or alternative methods to measure MD, ii) no overlap between GWAS discovery sample used for PRS_D calculation and target samples of the GxE study for the same phenotype; studies with overlapping samples were only included if outcomes differed (e.g., biomarkers), iii) analysis of GxE interactions. The review was registered to PROSPERO in advance (registration number CRD42021276066).

Quality assessment

While no validated quality or risk-of-bias assessment tool exists for PRS studies, particularly those investigating GxE interactions, we considered it important to conduct a structured narrative quality assessment to provide at least some evaluation of methodological rigor. We adapted the methodological approach developed by Sharew et al. [75] for systematic reviews of PRS studies, comprising seven core criteria (Q1–Q7) covering: power calculation, clarity of in-/exclusion criteria, presence of an external validation cohort, adjustment for multiple testing, specification of PRS construction methods, type of association analysis, and consideration of covariates/confounders. To capture additional study characteristics relevant to quality, bias, and interpretability, we also extracted information on study design, study sample size, GWAS discovery sample size, assessment of ethnicity, p-value threshold (pT) used for PRS calculation, method of environmental exposure and phenotype assessment, and whether rGE was assessed. The quality assessment was performed by SI under the supervision of NA. Full definitions of Q1–Q7 and the coding schemes for these additional variables are provided in Supplementary Tables 2 and 3.

Data extraction

A data extraction form was developed through discussion between all authors. Items were included based on expert knowledge and relevance to the review topic. The form was piloted with two publications by each reviewer independently and updated before being applied to all studies. Extracted data were checked by a second reviewer. For each study, we extracted information about authors, year of publication, study design, sample characteristics (size, ancestry, age, sex), PRS_D calculation (GWAS that provided the basis for PRS_D calculation, p-value threshold), type of environmental exposure, predicted depression-related outcome and main study results, including GxE interaction effects, main effects and gene-environment correlations (rGE, i.e., whereby an individual’s PRS_D influences the exposure to specific environments), if reported. For the classification of results as significant or non-significant, we adopted the significance thresholds reported by the original studies. Reported effect size estimates (e.g., odds ratios (OR), hazard ratios (HR), incidence rate ratios (IRR), relative excess risk due to interaction (RERI), standardized regression coefficients (β), unstandardized regression coefficients (b), correlation coefficient (r)) and explained variance (R²) were extracted whenever available. If multiple models or effect estimates were reported within a study, we extracted the largest effect size. This could reflect either the strongest positive association for risk effects (e.g., positive β or OR/HR > 1) or the strongest negative association for protective effects (e.g., most negative β or OR/HR < 1). Only information provided in original articles and supplements was included.

Results

Study selection and characteristics

A detailed flow chart of the systematic literature search is depicted in Fig. 1. In total, the literature search identified 6721 records. After exclusion of duplicates, supplements, conference abstracts, books, dissertations and non-English language articles, 1434 records remained. In a next step, abstracts were screened to remove reviews, editorials and records with irrelevant topics. Remaining records (n = 621) were screened in full-text and those studies not investigating GxE interactions were excluded. This resulted in 92 records assessed for eligibility. Out of these studies, those using PRS_D that were not based on GWAS data or those not reporting GxE interaction analyses were excluded, resulting in a final number of n = 56 studies for qualitative analysis.

Study results were grouped into those addressing interactions between an individual’s PRS_D and i) early environmental exposures (25/56 studies, 44.6%) and ii) adult/cumulative environmental exposures (26 of 56 studies, 46.4%) on depression-related outcomes, with n = 5 studies reporting data on both (8.9%). Studies on early exposures focused on different types of CT, but also contributed data on other forms of adversity, e.g., adoption, peer victimization, socioeconomic status (SES), urbanization, and prenatal adversity. The majority of studies on adult exposures investigated different types of trauma or SLEs, whereas others addressed an individual’s social environment as well as other protective (e.g. healthy lifestyle) and risk (e.g. urbanicity) factors. In terms of depression-related outcomes, across the 56 studies a total of 88 distinct outcomes were reported. Most of these referred to depression risk or depressive symptoms (70/88, 79.5%), whereas others focused on intermediate phenotypes such as brain structure and functioning or low-grade inflammation (18/88, 20.5%). PRS_D calculations in individual studies were based on GWAS on depressive phenotypes, including lifetime MDD, depressive symptoms or alternative methods to measure MD [6, 7, 9, 10, 16, 76–81] (Supplementary Table 1). The study samples cover a broad age range from infancy to late adulthood.

Quality assessment results

Most studies (47/56; 83.9%) were conducted in samples of European ancestry, while nine studies (16.1%) included participants from other ancestries. With one exception, all discovery GWAS used for PRS_D calculation were based on European ancestry populations. Target sample sizes ranged from 105 to 490780 participants (median = 3428), and discovery GWAS sample sizes ranged from 15298–1306090. Studies employed cross-sectional (exposure and outcome assessed at the same time point in the same sample), retrospective longitudinal (exposure measured before the outcome in time, but both obtained from past records and analyzed after outcome occurrence), or prospective longitudinal designs (both environmental exposures and depressive outcomes were prospectively assessed).

All studies described inclusion and exclusion criteria. Only a minority (8/56; 14.3%) reported a formal power calculation, and all studies considered relevant covariates or confounders in their analyses. Correction for multiple testing was applied in 33/56 studies (58.9%), while external cohort validation was performed in only four studies (7.1%). More than half of the studies (31/56; 55.4%) examined rGE. In terms of methodological specification, all studies detailed their PRS_D construction approach, with considerable variability in the p-value thresholds applied. The type of analysis used for the reported results was always specified. Environmental and phenotype assessments included objective measurement, established interview, register-based assessment, established self-report questionnaire, self-constructed measurement, and biomarker-based assessment. A detailed overview of all quality assessment ratings (Q1–Q7 and additional coded variables) is provided in Supplementary Table 4, with definitions in Supplementary Tables 2 and 3. In the absence of a validated quality or risk-of-bias tool for PRS-based GxE studies, these ratings should be interpreted with caution, as the assessment may not encompass all bias-relevant aspects important to this field.

Studies investigating the interaction between PRS_D and early environmental exposures on depression-related outcomes

Childhood trauma

Most studies on early exposures investigated the interaction of PRS_D and different types of CT (Table 1), which are among the strongest risk factors for developing MDD [25]. In a first cross-sectional study on 1645 adult MDD patients and 340 controls from the Netherlands Study of Depression and Anxiety (NESDA), Peyrot et al. [26] identified a significant additive and multiplicative interaction of PRS_D and CT (different types of abuse) that explained 0.5% of the variance in MDD risk. Specifically, the effect of PRS_D in predicting MDD and severe (chronic or recurrent) MDD risk was found to be increased in the presence of CT as assessed by a DSM-based interview. In an attempt to replicate these findings, Mullins et al. [23] reported a significant, but opposing multiplicative interaction in 240 MDD adult patients and 272 controls of the RADIANT UK cohort, explaining 1.9% of the variance in MDD. In this cross-sectional study, a higher PRS_D was associated with an increased MDD risk in those individuals unexposed to CT (neglect and abuse). In order to resolve these inconsistencies, Peyrot et al. [27] reanalyzed data from both cohorts with an updated, more accurate PRS_D [7] and meta-analyzed their results together with seven additional cohorts. Drawing on a total of 3024 MDD patients and 2741 controls, this study revealed no meta-analytic evidence for a significant interaction between PRS_D and self-reported CT on MDD risk in adulthood, irrespective of the specific CT domain. Comparable findings stem from a study on 1359 Syrian refugees, where the associations of war events and self-reported internalizing symptoms were found to be unmoderated by the child’s PRS_D [29]. Likewise, another cross-sectional analysis of 5853 children from the Adolescent Brain and Cognitive Development (ABCD) study found no interaction between PRS_D and a composite score of parent-reported CT (and other life stressors) on suicidal ideation or attempts [21]. Only recently, however, significant PRS_D x CT interaction effects on interview-derived depressive symptoms were reported in 38945 participants of the Lifelines Cohort Study (Lifelines), explaining up to 0.09% of variance [31]. Here, different types of neglect and abuse (except sexual abuse) were found to amplify the effect of an individual’s PRS_D across multiple p-value thresholds. In contrast to the Lifelines findings, a similarly large study of 35633 UK Biobank (UKB) participants found no evidence for a PRS_D x CT interaction, although both PRS_D and CT showed significant main effects on depression risk [17].

Table 1.

Studies investigating the interaction between a PRS_D and early environmental exposures on depression-related outcomes.

study	study design¹	population (ancestry)	age (mean)	female (%)	PRS discovery sample	PRS p-value threshold (pT)	environmental exposure	predicted outcome	main results (effect size estimate [CI])
Prenatal adversity
Chen et al., [43]	prospective longitudinal	cohort 1: (European) 5546	t1 = 4 yrs (6 follow-ups until age 16)	49.1	MTAG-2018	range: p= 0.01–1.00 at intervals of 0.01	prenatal maternal depression (EPDS) and anxiety (CCEI)	internalizing symptoms (SDQ)	• no sign. multiplicative interaction of PRS_D and prenatal maternal depression (β = 0.004 [–0.031, 0.039]) and anxiety (β = 0.036 [0.002, 0.071]) on internalizing symptoms • sign. main effects of PRS_D (β = 0.035 [0.015, 0.055]), prenatal maternal depression (β = 0.079 [0.052, 0.106]) and anxiety (β = 0.062 [0.036, 0.089]) • sign. correlation of PRS_D and prenatal maternal depression (r = 0.061) and anxiety (r = 0.069)
Chen et al., [43]	prospective longitudinal	cohort 2: (European) 514	t1 = 3.4 yrs t2 = 8.7 yrs	49.1	MTAG-2018	range: p= 0.01–1.00 at intervals of 0.01	prenatal maternal depression (CES-D)	internalizing symptoms (CBCL)
Wang et al., [42]	prospective longitudinal	2938 mother-child dyads (European)	t1 = 10.6 yrs t2 = 13.8 yrs	49.8	PGC-MDD-2018 (excluding 23andMe)	<0.05	prenatal maternal vitamin D status	depression (SMFQ)	• no sign. interaction of PRS_D and prenatal maternal vitamin D status on childhood or adolescent depression (largest OR = 4.09 [1.26, 13.30]) • sign. main effect of PRS_D for childhood depression (largest OR = 1.96 [1.24, 3.03]), no sign. main effect of vitamin D status (largest OR = 1.32 [0.98, 1.79])
Childhood trauma and aversive experiences
Chen et al., [17]	cross-sectional	35633 (European)	55.7 yrs	62.6	UKB-2020	<5 x 10^-8	adverse exposures (TFMHQ)	depression age of onset (self-report)	• no sign. GxE interaction with CT (β = −0.550) • sign. main effects of PRS_D (β = –1.601) and adverse exposures (felt hated: β = –7.385)
Halldorsdottir et al., [19]	cross-sectional	cohort 1 (European): 279 MDD 187 HC	14.8 yrs 14.7 yrs	68.0 63.0	PGC-MDD-2018	7 pTs tested (range: <5 x 10^-8 to <0.10; final pT: <0.05)²	CT (LES, MEL)	depression (DISC, DIJK, BDI-II)	• no sign. multiplicative interaction of PRS_D and CT on MDD status (OR = 0.838 [0.502, 1.401]), depression severity (β = –0.187 [–0.444, 0.069]), age of onset (β = –0.321 [–0.934, 0.292]) • sign. main effect of PRS_D on case-control status (OR = 1.560 [1.230, 1.980], R² = 7.9%), depression severity (β = 0.177 [0.021, 0.288], R² = 7.7%), and age at onset (β =−0.375 [−0.688, −0.062], R² = 45.9%), main effect of CT NR
Halldorsdottir et al., [19]	prospective longitudinal	cohort 2 (European): 756 depressed 694 non-depressed	t1 = 14.0 yrs t2 ≈ 14.6 yrs t3 ≈ 15.0 yrs t4 ≈ 16.0 yrs	63.0	PGC-MDD-2018	7 pTs tested (range: <5 x 10^-8 to <0.10; final pT: <0.05)²	CT (CTQ)	depression(CDI)	• no sign. multiplicative interaction of PRS_D and CT on (prospective) MDD status (HR = 0.877 [0.592, 1.300]), depressive symptoms (β = 1.010 [–0.301, 2.320]) • sign. main effect of PRS_D on (prospective) MDD status (HR = 1.202 [1.045, 1.383], R² = 4.4%), depressive symptoms (β = 0.557 [0.167, 0.947], R² = 7.5%), main effect of CT (NR)
Iob et al., [20]	retrospective longitudinal	3428 (British)	t1 ≈ 56.7 yrs (8 follow-ups over 14 yrs)	55.4	PGC-MDD & 23andMe 2019	=1.0	CT (ELSA’s Life History interview), parental bonding (PBI)	depressive symptoms (CESD-8)	• sign. additive and multiplicative interaction of PRS_D and CT on depressive symptom trajectories (largest OR = 1.13 [1.04, 1.23]) • sign. additive and multiplicative interaction of PRS_D and low parental bonding on depressive symptom trajectories (largest OR = 1.47 [1.15, 1.89]) • sign. main effects of PRS_D (largest OR = 1.47 [1.28, 1.70]), CT (largest OR = 1.44 [1.30,1.60]) and low parental bonding (largest OR = 2.36 [1.70, 3.28])
Joo et al., [21]	cross-sectional	5853 (multi-ethnic)	9.9 yrs	52.8	MDD-2019	=1.0	early life stress (FES & KSADS-5, PTSD module)	suicidal ideation, suicide attempt (KSADS-Comp)	• no sign. interaction of PRS_D and early life stress on suicidal ideation (NR) or suicide attempts (NR) • sign. main effect of PRS_D in a subsample of European ancestry (largest OR = 1.12 [1.04, 1.21], Pseudo R² = 2.9%), main effect of early life stress NR
Mullins et al., [23]	cross-sectional	240 MDD 272 HC (European)	44.7 yrs NR	73.5 NR	PGC-MDD-2013 (excluding RADIANT UK data)	9 pTs tested (range: <0.0001 to <0.5)	CT (CTQ)	depression (SCAN)	• sign. multiplicative interaction of PRS_D and CT on MDD status (threshold specific, largest R² = 1.9%), no sign. additive interaction (largest OR = 1.00 [0.99, 1.00]) • no sign. main effect of PRS_D (OR = 1.18 [0.98, 1.42], Pseudo R² = 0.7%), sign. main effect of CT (R² = 30.2%) • no sign. correlation of PRS_D and CT (largest OR = 1.76 [0.49, 6.28])
Musci et al., [24]	retrospective longitudinal	488 (mostly African American)	t1 ≈ 11.9 yrs t2 ≈ 12.9 yrs t3 ≈ 13.9 yrs t4 ≈ 14.9 yrs t5 ≈ 15.9 yrs t6 ≈ 16.9 yrs t7 = 17.8 yrs	45.6	PGC-MDD-2013	<0.05	parental death and/or divorce by age 6 (HDLF)	internalizing symptoms (BHIF-A)	• no sign. interaction of PRS_D and parental death/divorce on internalizing symptoms (β = 0.011) • sign. main effects of PRS_D (β = 0.109 [−0.141, 0.085]) and parental death/divorce (occasion-specific; largest β = −0.352 [−0.573, −0.005])
Peyrot et al., [26]³	cross-sectional	1645 MDD 340 HC (European)	42.2 yrs 43.3 yrs	68.0 57.0	PGC-MDD-2013 (excluding NESDA data)	8 pTs tested (range: <0.001 to <0.5)	CT (CTI)	lifetime MDD (CIDI)	• sign. additive (RERI = 0.37 [0.14, 0.71]) and multiplicative (largest OR = 1.16) interaction of PRS_D and CT (for emotional neglect and physical abuse) on MDD risk • sign. main effects of PRS_D (largest OR = 1.22) and CT (OR = 1.69) • no sign. correlation between PRSD and CT (largest β = 0.02)
Peyrot et al., [27]	meta-analysis on mixed design studies	9 PGC cohorts (European): 3024 MDD 2741 HC	36.3-53.6 yrs	50.0- 70.0	PGC-MDD-2018 (excluding data from cohorts of this study)	5 pTs tested (range: <0.01 to <1)	CT (CTQ and other self-report measures)	MDD (DSM-criteria based diagnostic interviews)	• no sign. additive (largest RERI = 0.84 [−0,52, 22.18]) or multiplicative (largest OR = 1.05 [0.91, 1.20]) interaction of PRS_D and CT on MDD risk. • sign. main effects of PRS_D (largest OR = 1.35 [1.22, 1.48], R² = 1.70%) and CT (largest OR = 2.63) • sign. correlation of PRS_D and CT (β = 0.76)
Schür et al., [28]	retrospective longitudinal	516 soldiers (European)	t1 = 29.6 yrs t2 ≈ 29.7 yrs t3 ≈ 30.2 yrs t4 ≈ 31.2 yrs t5 ≈ 32.2 yrs	8.5	PGC-MDD-2018	13 pTs tested (range: <5x10^-8 to <1)	CT (ETI)	depressive symptoms (SCL-90)	• no sign. interaction of PRS_D and CT on baseline (largest β = −0.332) and post-deployment depressive symptoms (largest β = –1.035) • no sign. main effect of PRS_D (largest β = 1.154), main effect of CT NR • nominally sign. correlation of PRS_D and CT (β = 0.09)
Smeeth et al., [29]	cross-sectional	1359 Syrian refugees (Middle Eastern ancestry)	11.3 yrs	52.8	UKB-portability-2022	NR	war events (WEQ)	depressive symptoms (CES-DC) trauma symptoms (CPSS)	• no sign. interaction of PRS_D and war events on composite score of internalizing symptoms (OR = 0.86 [0.68, 1.09]) • no sign. main effect of PRS_D (OR = 0.67 [0.24, 1.88]), sign. main effect of war events (OR = 0.93 [0.90, 0.95])
Taylor et al. [30]	cross-sectional	2050 (British)	75.2 yrs	65.8	PGC-MDD & 23andMe 2019	=1.0	CT (ELSA’s Life History interview)	depressive symptoms (CESD-8)	• no sign. interaction of CT and PRS_D on depressive symptoms during the Covid-19 pandemic (OR = 0.944 [0.849, 1.049]) • no sign. main effects of PRS_D (OR = 1.135 [0.934, 1.386])_, sign. main effect of CT (OR = 1.250 [1.124, 1.386])
Wang et al., [31]	cross-sectional	38945 (European)	44.1 yrs	59.6	PGC-MDD & 23andMe 2019	11 pTs tested (range: <5 x 10^-8 to <1)	CT (CTQ)	depressive symptoms (MINI)	• sign. interaction of PRS_D and CT on depressive symptoms (largest β = 0.0384, R² = 0.09%) • sign. main effects of PRS_D (largest β = 0.1149, R² = 0.69%) and CT (β = 0.3191, R² = 5.91%) • sign. correlation of PRS_D and CT (r = 0.08)
Other forms of childhood adversity and protective factors
Agerbo et al., [32]	retrospective longitudinal	17098 MDD 18582 HC (Danish)	t1 = 10 yrs (continuous follow-up up to 21 yrs)	68.7 49.2	MDD-2019 (excluding iPSYCH)	10 pTs tested (range: <5x10^-8 to <1)	childhood SES (register-based)	early-onset depression (clinical records)	• no sign. interaction of PRS_D and SES (t1) on early-onset depression (NR) • sign. main effects of PRS_D (HR=1.32 [1.29, 1.35]) and SES (largest HR = 1.65 [1.45, 1.87]) • sign. correlation of PRS_D and childhood SES (largest r = –0.058)
Armitage et al., [33]	prospective longitudinal	2268 (European)	t1 = 13.0 yrs t2 = 23.0 yrs	63.9	PGC-MDD & 23andMe 2019	11 pTs tested (range: <5x10^-8 to <0.5)	peer victimization (BFIS)	depressive symptoms (SMFQ), well-being (WEMWBS)	• no sign. interaction of PRS_D and peer victimization on prospective depressive symptom development (largest β = 0.024 [–0.022, 0.070]) • no sign. interaction of PRS_D and peer victimization on well-being (largest β = –0.085 [–0.554, 0.384]) • sign. main effects of PRS_D (largest β = 0.100 [0.017, 0.184]) and peer victimization (largest β = 0.188 [0.084, 0.293]) on prospective depressive symptoms • no sign. main effects of PRS_D (largest β = 0.091 [–0.702, 0.885]) and peer victimization (largest β = –0.014 [–0.979, 1.010]) on well-being • sign. correlation of PRS_D and peer victimization (largest β = 0.163 [0.059, 0.266], R² = 0.42% at pT = 0.3)
Chen et al., [17]	cross-sectional	35633 (European)	55.7 yrs	62.6	UKB-2020	<5x10^-8	protective exposures (TFMHQ)	depression age of onset (self-report)	• sign. interaction of PRS_D and felt loved as a child on depression age of onset (β = 0.708), effect driven by males (β = 1.421) • sign. main effects of PRS_D (β = –1.601) and adverse/protective exposures (felt loved: β = 5.534)
Kosciuszko et al., [34]	retrospective longitudinal	6202 (British)	65.2 yrs (5 follow-ups over 14 yrs)	52.2	PIR-2021	NR	SES (self-report educational attainment)	depressive symptoms (CES-D)	• sign. interaction of PRS_D and educational attainment on baseline depressive symptoms (β = −0.01 [−0.02, −0.01]), but not on rate of change in depressive symptoms (β = −0.00 [00.00, 0.00]) • sign. main effect of PRS_D (β = 0.35 [0.27, 0.44]) and educational attainment (β = −0.06 [−0.07, −0.05]) on baseline depressive symptoms • no sign. main effects of PRS_D (β = 0.02 [−0.04, 0.01]) and educational attainment (β = 0.001 [−0.001, 0.004]) on rate of change in depressive symptoms
Lehto et al., [35]	cross-sectional	3151 adopted 240329 non-adopted (European)	56.9 yrs	54.4	SSGAC-2016 (excluding UK Biobank data)	7 pTs tested (range: <5x10^-8 to <0.5; final pT: <0.5)³	adoption (self-report)	depressive symptoms (self-report)	• no sign. interaction of PRS_D and adoption on depressive symptoms (β = −0.01 [−0.05, 0.03], R² = 1.97%) • sign. main effects of PRS_D (β = 0.05 [0.05, 0.06], R²= 1.93%) and adoption (β = 0.10 [0.05, 0.14], R²= 1.66%) • sign. correlation of PRS_D and adoption (OR = 1.066 [1.03–1.11])
Misztal et al., [82]	cross-sectional, prospective longitudinal	4975 (European)	9-10 yrs (followed-up longitudinally)	47.0	PGC-MDD & 23andMe 2019	<0.05	sport frequency and sport type (parent report)	anxious/ depressed, withdrawn/ depressed symptoms (CBCL)	• no sign. interaction of PRS_D and sport type/frequency on anxious/depressed and withdrawn/depressed symptoms (NR) • sign. main effects of PRS_D (NR) and sport type/frequency (e.g., team sports: β = −0.09 [−0.11, −0.06])
Nelemans et al., [36]	prospective longitudinal	327 mother-child dyads with low SES (European)	t1=13.0 yrs t2=14.0 yrs t3=15.0 yrs t4=16.0 yrs t5=17.0 yrs t6=18.0 yrs	43.4	PGC- 23andMe-2016	<5x10^-8; sensitivity analysis: 12 pTs tested (range: <5x10^-7 to <0.5)	critical parenting t1-t6 (LEE, multi-informant)	depression (RADS-2)	• sign. interaction of PRS_D (threshold specific) and critical parenting on intercept levels of adolescent depressive symptoms (largest β = 0.15) and changes over time (largest β = 0.25) • sign. main effect of PRS_D (different thresholds) on mean depressive symptoms (largest r = 0.20), but not on changes over time (NR), sign. main effect of critical parenting (largest r = 0.43) • sign. correlation of PRS_D and critical parenting (largest r = 0.21)
Perret et al., [37]	prospective longitudinal	748 (Canadian)	t1≈12-13 yrs t2≈15-17 yrs	55.5	MDD-2019	4 pTs tested: (<0.01; <0.10; <0.50; <1.00)	peer victimization (multi-informant: SRVS, Behavior Questionnaire)	depressive symptoms (MAI)	• no sign. interaction of PRS_D and peer victimization (t1) on depressive symptoms (t2) (largest β = 0.039) • sign. main effects of PRS_D (largest β = 0.10) and peer victimization (largest β = 0.34) • sign. correlation of PRS_D and self-reported (but not teacher reported) peer victimization (r = 0.11)
Wang et al., [31]	cross-sectional	overall sample: 38945 (European)	44.1 yrs	59.6	PGC-MDD & 23andMe 2019	11 pTs tested (range: <5x10^-8 to <1)	long-term difficulties; SLE; social support (parent-report)	depressive symptoms (MINI)	overall sample: • sign. interaction of PRS_D and long-term difficulties (largest β = 0.0541, R² = 0.17%), SLE (largest β = 0.0385, R² = 0.05%) and social support (largest β = –0.0417, R² = 0.09%) on depressive symptoms • sign. main effects of PRS_D (β = 0.1105, R² = 0.66%), long-term difficulties, SLE and social support (largest β = 0.4195, R² = 9.54%) • sign. correlation of PRS_D and long-term difficulties, SLEs, social support (largest r = 0.08)
Wang et al., [31]	cross-sectional	child subsample: 2865 (European)	12.4 yrs	50.3	PGC-MDD & 23andMe 2019	11 pTs tested (range: <5x10^-8 to <1)		depressive symptoms (CBCL, YSR)	child subsample: • no sign. interaction of PRS_D and long-term difficulties (largest β = 0.0862, R² = 0.15%), SLE (largest β = 0.0352, R² = 0.12%) and social support (largest β = 0.0229, R² = 0.12%) on depressive symptoms • sign. main effects of PRS_D (in children: β = 0.0403, R² = 0.21%), long-term difficulties, SLE and social support (largest β = 0.6235, R² = 7.59%)

Open in a new tab

b unstandardized regression coefficient, BDI-II Becks-Depression-Inventar Revision, BFIS Bullying and Friendship Interview Schedule, BHIF-A Baltimore How I Feel questionnaire – Adolescent Report, CBCL Child Behavior Checklist, CCEI Crown Crisp Experiential Index, CDI Children’s Depression Inventory, CES-D Center for Epidemiologic Studies Depression Scale, CESD-8 8-item version of the Centre for Epidemiological Studies Depression, CES-DC Center for Epidemiological Studies Depression Scale for Children, CI confidence interval, CIDI Composite International Diagnostic Interview for DSM-IV, CPSS Trauma Symptoms Checklist, CT childhood trauma, CTI Childhood Trauma Interview, CTQ Childhood Trauma Questionnaire, DIJK Depressions Inventar für Kinder und Jugendliche, DISC Depression Intensity Scale Circles, DSM Diagnostic and Statistical Manual of Mental Disorders, EPDS Edinburgh Postnatal Depression Scale, ELSA ELSA ’s (English Longitudinal Study of Ageing) Life History interview, ETI Early Trauma Inventory, FES Youth Family Environmental Scale, HC healthy controls, HDLF Life Change Events subscale from the Health and Daily Living Form, HR hazard ratio, KSADS(-Comp) Kiddie Schedule for Affective Disorders and Schizophrenia, LEE Level of Expressed Emotions Scale, LES Life Event Survey, MAI Mental Health and Social Inadaptation Assessment (MIA), MDD Major depressive disorder, MEL Munich Life Event List, MINI International Neuropsychiatric Interview, NR not reported, OR odds ratio, PBI Parental Bonding Instrument, PRS_D polygenic risk score for depression, pT p-value threshold, Pseudo R² goodness-of-fit measure for non-linear regression models, analogous to R² but not directly interpretable as explained variance, r correlation coefficient; R² proportion of variance in the outcome explained by the predictors, RADS-2 Reynolds Adolescent Depression Scale, 2nd edition, RERI Relative Excess Risk due to Interaction, SCL-90 Symptom-Checklist, SDQ Strengths and Difficulties Questionnaire, SES socioeconomic status, SLE Stressful life events, sign. significant, SMFQ The Short Mood and Feelings Questionnaire, SRVS Self-Report Victimization Scale, TFMHQ Thoughts and Feelings Mental Health Questionnaire, WEMWBS Warwick-Edinburgh Mental Well-Being Scale, WEQ War Events Questionnaire, yrs years, YSR Youth Self-Report questionnaires, β standardized regression coefficient, ≈ age not reported, inferred from the longitudinal design.

¹a study was classified as “cross-sectional” when exposure and outcome were assessed at the same time point in the same sample, “retrospective longitudinal” when the exposure was measured before the outcome in time, but both were obtained from past records and analyzed after outcome occurrence, or as “prospective longitudinal” when both environmental exposures and depressive outcomes were prospectively assessed;

²data meta-analyzed in Peyrot et al. [27];

³final pT: p-value threshold which explained the highest variance in the respective study used in main analysis.

Moving beyond cross-sectional analyses, studies also investigated whether the interaction between PRS_D and CT prospectively predicts the onset of depressive symptoms. In one of the few studies on adolescents, Musci et al. [24] observed no interaction between PRS_D and early adversity (parental death and/or divorce by age 6) on state or trait internalizing symptoms in 488 participants, who were annually assessed over a seven-year period from late childhood ( ≈ 12 years) to late adolescence ( ≈ 18 years). Likewise, Halldorsdottir et al. [19] reported no PRS_D × CT interaction on MDD status, depressive symptoms and age of onset in both cross-sectional and longitudinal analyses across two independent cohorts. Instead, additive effects of PRS_D and CT (different types of abuse) on prospective depressive symptoms were found in an epidemiological cohort of 1450 adolescents that underwent follow-ups after 6, 12, and 24 months. In line with this finding, a cross-sectional analysis of data from a second adolescent sample comprising 279 MDD patients and 187 controls revealed comparable additive effects on MDD status and depression severity.

Complementing longitudinal GxE analyses in adolescent cohorts, three other studies investigated how PRS_D and retrospectively assessed CT interact to predict depressive symptoms in adulthood. In a sample of 516 soldiers who completed five assessments prior and up to 5 years after deployment, Schür et al. [28] reported no interaction between PRS_D and a broad spectrum of CT on baseline and post-deployment depressive symptoms. In contrast, significant additive and multiplicative interactions between PRS_D and CT (with an OR of up to 1.13) and between PRS_D and low parental bonding on depressive symptom trajectories over a fourteen-year period in adulthood were found in 3428 participants from the English Longitudinal Study of Ageing (ELSA, [20]). More precisely, GxE analysis revealed that the positive association between CT and depressive symptoms was stronger among those individuals with higher PRS_D. In a subsequent wave of the ELSA study, the interaction between PRS_D and CT on depressive symptoms during the COVID-19 pandemic was not significant [30].

To conclude, while most studies suggest main effects of an individual’s PRS_D [19, 21], CT [23, 28–30] or both [20, 24, 26, 27, 31] on depression phenotypes, evidence for GxE interaction remains considerably heterogeneous. Notably, small but significant interactions between PRS_D and CT (including abuse, neglect, and early adversity) have been identified in the largest study to date [31] as well as in some smaller individual studies [20, 23, 26], with significant effects reported for both diagnosis of MDD [23, 26] and depressive symptoms [20, 31]. Overall, explained variance for significant GxE interactions was consistently very small (R² = 0.09–1.9%), standardized regression coefficients were modest (β ≈ 0.04), and observed odds ratios indicated only minor risk increases (OR = 1.13–1.16).

Other forms of childhood adversity and protective factors

A number of GxE studies have focused on other forms of childhood adversity associated with an increased risk for MDD. In a comprehensive study, Lehto et al. [35] analyzed the interaction of PRS_D and self-reported childhood adoption on depressive symptoms and other mental health outcomes in 243480 adult UKB participants. While the authors found no evidence for GxE interactions, each standard deviation increase in PRS_D was associated with an increase of 6% in the odds of being adopted, underlining the importance to account for potential genetic confounding in GxE studies. In contrast, a joint analysis of 2865 children and 38945 adults from the Lifelines Cohort Study revealed significant interactions between PRS_D and various stress-related exposures (including long-term difficulties, stressful life events, and lack of social support) on parent- and self-reported depressive symptoms [31], although explained variance was small (R² = 0.17%). When restricting the analyses to the child subsample, the GxE interaction effects vanished, while significant main effects of PRS_D and stress-related exposures remained.

Two longitudinal studies further investigated whether an individual’s PRS_D moderates the detrimental consequences of peer victimization. Among 748 adolescents from the Quebec Longitudinal Study of Children who were monitored twice over a two-year period, Perret et al. [37] reported no interaction between PRS_D and a multi-informant index of peer victimization on (prospective) depressive symptoms. Similar findings were obtained in the Avon Longitudinal Study of Parents and Children (ALSPAC) that provides data from 2268 participants assessed twice at age 13 and 23 years. In this study, an individual’s PRS_D and peer victimization were found to be independent predictors of depressive symptoms in early adulthood [33]. Interestingly, evidence for rGE was observed for self-reported, but not teacher-reported peer victimization [33, 37].

In contrast, findings from a six-year longitudinal study on 327 adolescents from the Research on Adolescent Development and Relationships (RADAR–Young) cohort suggested significant interactions between PRS_D and a multi-informant longitudinal index of critical parenting in predicting depressive symptom trajectories across adolescence [36]. Consistent with a diathesis–stress framework, latent growth curve modelling indicated that higher PRS_D was associated with elevated baseline depressive symptoms (β up to 0.15) and a stronger increase in symptoms over time (β up to 0.25), particularly among adolescents exposed to high parental criticism. Moreover, GxE effects varied depending on the p-value threshold applied for PRS_D construction, highlighting the sensitivity of results to whether the score included only genome-wide significant variants or also more liberal sets of variants with weaker associations to depression. Finally, this study also reported a significant gene–environment correlation, as higher PRS_D was positively associated with the likelihood of experiencing critical parenting.

Other studies focused on broader indicators of childhood adversity, namely (low) SES and educational attainment. In a register-based study comprising 17098 MDD patients and 18582 individuals randomly selected from the population, PRS_D and (low) childhood SES were again identified as independent predictors of early-onset depression [32]. Conflicting data stem from the ELSA study (n = 6202), which followed older adults over 14-year period [34]. Here, a weak but significant interaction between PRS_D and educational attainment was observed for baseline depressive symptoms in late life (β = –0.01), indicating that each additional year of completed schooling predicted a slight reduction in depressive symptoms among those with higher genetic liability for depression. No interaction effects were found for changes in depressive symptoms over time.

Further research into protective factors in addition to educational attainment, was performed in one cross-sectional analysis of 35633 participants from the UKB [17]. As reported above, this study found no significant GxE effect for CT. However, a significant interaction emerged between PRS_D and the experience of feeling loved as a child on age of depression onset (β = 0.708), which was particularly pronounced in males (β = 1.421). These findings highlight the importance of incorporating positive as well as negative environmental factors in GxE research on depression. A subsequent GWEIS identified a novel candidate locus interacting with felt loved as a child (GSAP) as well as several genes with suggestive significant associations (CMYA5, KIRREL3) in males. Another study exploring protective factors in 4975 children participating in the ABCD study, however, reported independent effects of (lack of) sport activity and PRS_D on anxious/depressed, withdrawn/depressed symptoms, with no evidence for GxE [82].

Consistent with findings on CT, a higher PRS_D and different forms of childhood adversity were identified as independent predictors of depressive phenotypes in most studies [32, 33, 35, 37], with substantial evidence for gene–environment correlation [32, 33, 35–37]. Significant, albeit weak GxE effects have been reported with critical parenting (β up to 0.25) and low educational attainment (β = –0.01) [34, 36]. With respect to protective factors, only few studies exist [17, 31] but first evidence suggests significant interactions, such as with social support (β = –0.0417; R² = 0.09%) and the experience of feeling loved as a child (β = 0.708; in males up to β = 1.421).

Prenatal adversity

While most studies investigated adversity in early childhood, others have begun to examine prenatal exposures given the extensive literature on developmental origins of psychiatric symptoms [83]. In a prospective longitudinal analysis within ALSPAC (n = 5546), Chen et al. [43] tested whether a child’s PRS_D moderates effects of prenatal maternal depression/anxiety on internalizing symptoms from ages 4–16. While maternal depression and anxiety as well as a child’s PRS_D independently and additively predicted internalizing symptoms, the authors found no evidence for a significant GxE interaction. However, weak correlations between a child’s PRS_D and maternal psychopathology were observed, indicating significant rGE. Respective findings were replicated in the Prevention of Preeclampsia and Intrauterine Growth Restriction cohort (n = 514 children) that provided complementary measures. Another study focused on maternal vitamin D status during gestation, defined by circulating levels of 25-hydroxyvitamin D [25(OH)D], that has been previously linked to the child’s neurodevelopment and mental health status [84]. Analyzing data from 2938 mother-child dyads from the ALSPAC, Wang et al. [42] found no evidence for an interaction of PRS_D and prenatal vitamin D status on depression risk in childhood and adolescence.

So far, only few studies have examined whether PRS_D interacts with heterogeneous prenatal exposures. Available evidence points to independent main effects of PRS_D, maternal depressive and anxiety symptoms, and (to a lesser extent) maternal vitamin D status on child mental health, but no robust GxE interactions [42, 43].

Studies investigating the interaction between PRS_D and adult or cumulative environmental exposures on depression-related outcomes

Trauma and stressful life events in adulthood

Most studies on environmental exposures in adulthood (Table 2) focused on the interaction of PRS_D and different types of trauma or SLEs implicated in the pathogenesis of MDD [85]. In a first cross-sectional study on 8761 older adults from the Health and Retirement Study (HRS), Musliner et al. [49] reported additive main effects of an individual’s PRS_D and SLEs in the past two years on depressive symptoms, but no GxE interaction. Comparable findings were obtained from analyzing cross-sectional data of 1605 MDD patients and 1064 controls from the RADIANT UK cohort [23]. In order to account for a significant rGE of PRS_D and adult SLEs observed in this study, the authors conducted separate GxE analyses for those SLEs dependent on (e.g., unemployment) or independent of (e.g., death or illness of a family member) the individual’s behavior, both yielding no additive or multiplicative interaction with PRS_D on MDD status. By contrast, in another cross-sectional study comprising 4919 individuals from the Generation Scotland Cohort, a higher PRS_D was found to increase depressive symptoms in those participants reporting high numbers of SLE over the past 6 months [44]. The respective GxE interaction was nominally significant in the full cohort, explaining 0.08% of the variance in depression. Stratified analyses revealed significance in women, but not in men. Convergent evidence for significant interactions of PRS_D and a broad range of stress-related exposures on interview-derived depressive symptoms stems from the Lifelines Cohort Study (n = 41810). Here, the effect of an individual’s PRS_D was significantly amplified by self-reported SLEs, long-term difficulties and lack of social support, with GxE effects explaining up to 0.17% of the variance [31]. GxE effects attenuated after adjusting for the interactions between SES and stress-related exposures in order to account for small rGEs observed in this study. When restricted to adults, GxE effects remained significant and explained up to 0.15% of the variance in depressive symptoms. Extending this line of research, Colodro-Conde et al. analyzed cross-sectional data from 5221 twins [54] and explicitly differentiated between personal SLEs and those experienced by an individual’s close network in the past 12 months. Results suggested a multiplicative interaction of PRS_D and personal SLEs (but not network SLEs) that contributed positively to the risk of depression, yet explaining only a small amount of variance (0.12%). Respective GxE interaction was predominantly driven by women and by those personal SLEs that depend on the individual’s behavior (that were also found to be more heritable), raising concerns that obtained findings could be partly attributed to rGE.

Table 2.

Studies investigating the interaction between a PRS_D and adult or cumulative environmental adversity on depression-related outcomes.

author	study design¹	population (ancestry)	age (mean)	female (%)	PRS discovery sample	PRS p-value threshold (pT)	environmental exposure	predicted outcome	main results (effect size estimate [CI])
Trauma and stressful life events
Arnau-Soler et al., [11]	cross-sectional	4919 (European)	57.2 yrs	60.8	PGC-MDD-2013, July 2016* (excluding 23andMe & Generation Scotland data)	8 pTs tested (range: <5x10^-8 to <=1)	SLEs in the past 6 months (TSLE)	depressive symptoms (GHQ)	• nominally sign. interaction of PRS_D and SLEs (best-fit pT: 0.01) on depressive symptoms (largest β = 0.028, R² = 0.08%), sign. in women (β = 0.044, R² = 0.19%), not men (β = 0.039, R² = 0.15%) • sign. GxE interactions for both active (largest β = 0.039, R² = 0.15%) and passive SLEs (largest β = 0.040, R² = 0.16 %) in women • sign. main effects of PRS_D (β = 0.080, largest R² = 0.64%) and SLEs (β = 0.222, R² = 4.91%) • sign. correlation of PRS_D and SLEs (largest R² = 0.40-0.45%^v)
Choi, & Chen et al., [45]	prospective longitudinal	3079 soldiers (European)	pre-deployment t1 = 25.9 yrs post-deployment t2 ≈ 26.3 yrs t3 ≈ 27.2 yrs	4.0	PGC-MDD-2018 (excluding 23andMe data)	8 pTs tested (range: 5x10^-8 to 1.0; final pT: 0.01)	combat stress exposure (self-report); unit cohesion (WRAIR MCS)	MDD post- deployment (CIDI-SC based self-report)	• no sign. interaction of PRS_D and unit cohesion or combat stress on incident MDD (largest OR = 1.32 [0.96,1.76]) • sign. main effects of PRS_D (largest OR = 1.58 [1.12, 2.25], Pseudo R² = 0.4%), unit cohesion (largest OR = 0.67 [0.60, 0.74]) and combat stress (largest OR = 1.34 [1.21, 1.50]) • no sign. correlation of PRS_D and unit cohesion (r = −0.02)
Colodro-Conde et al., [54]	cross-sectional	5221 twins (European)	35.7 yrs	65.6	PGC-MDD-2013, July 2016* (excluding 23andMe & QIMR data)	8 tested (range: <5x10^-8 to p<1; final pT: p<1)²	personal and network SLEs in the past 12 months (HLQ)	depressive symptoms (DSSI & SCL-90 based)	• sign. interaction of PRS_D and personal SLEs on depressive symptoms in women (largest R² = 0.12%), not men • sign. main effects of PRS_D (largest R² = 0.46%), personal SLEs (R² = 12.9%) and network SLEs (R² = 0.3%) • sign. correlation of PRS_D and SLEs (NR)
Domingue et al., [46]	prospective longitudinal	8588 (non-Hispanic white)	t1 = 59 yrs t2 ≈ 60 yrs	58.0	PGC-MDD-2013 & SSGAC-2016	NR	death of spouse (self-reported)	depressive symptoms (CES-D)	• sign. interaction of PRS_D (SSGAC & PGC based) and death of spouse on depressive symptoms (largest β = 0.69 [0.24, 1.14]) • sign. main effects of PRS_D (SSGAC & PGC based, largest β = 0.13) and death of spouse (β = 1.97 [1.50, 2.40])
Fang et al., [47]	prospective longitudinal	5227 medical interns (European)	pre-internship t1 = 27.6 yrs during internship t2 ≈ 27.9 yrs t3 ≈ 28.1 yrs t4 ≈ 28.4 yrs t5 ≈ 28.7 yrs	50.3	MDD-2019 (excluding 23andMe data)	9 pTs tested (range: <5 ×10⁻⁸ to <0.5)	medical internship stress	MDD, depressive symptoms (PHQ-9)	• sign. interaction of PRS_D and internship stress on depressive symptoms (largest β = 0.036) and MDD diagnosis (OR = 1.40; RERI = 31.75 [5.47, 58.03]) • sign. main effects of PRS_D (β = 0.095; OR = 1.21) and internship stress (NR)
Mullins et al., [23]	cross-sectional	1605 MDD 1064 HC (European)	36.7 yrs 41.5 yrs	70.7 59.0	PGC-MDD-2013 (excluding RADIANT UK data)	9 pTs tested (range: <0.0001 to <0.5)	adult SLEs (LEQ-B)	depression (SCAN)	• no sign. additive (largest OR = 1.01 [0.99, 1.02], R² = 0.07%) or multiplicative (largest OR = 1.05 [0.98, 1.12], R² = 0.1%) interaction of PRS_D and SLEs (for both dependent and independent SLEs) on MDD status • sign. main effects of PRS_D (OR = 1.22 [1.12, 1.32], R² = 1.1%) and SLEs (dependent: R² = 6.6%, independent: R² = 1.9%, total: R² = 0.7%) • sign. correlation of PRS_D and dependent SLEs in MDD patients (OR = 1.08 [1.03, 1.14])
Musliner et al., [49]	cross-sectional	8761 (mostly European)	63.9 yrs	62.0	PGC-MDD-2013	10 pTs tested (range: <0.10 to <1.0; final pT: <0.4)²	SLEs in the past 2 yrs (self-report)	depressive symptoms (CES-D)	• no sign. interaction of PRS_D and SLEs on depressive symptoms (CES-D score ≥ 4, β = −0.06) and number of symptoms (β = −0.0003, Pseudo R² = 2.01%) • sign. main effects of PRS_D (β = 0.21, Pseudo R² = 0.95%) and SLEs (OR = 2.19 [1.86, 2.58], Pseudo R² = 1.08%) • no sign. correlation of PRS_D and SLEs (β = 0.02)
Peter et al., [52]	prospective longitudinal	218 stressed students (European)	22.9 yrs (6 assessments within 13 months)	73.0	PGC-MDD & 23andMe 2019	≤0.05	exam period	depressive symptoms (HADS), perceived stress (self-report)	• no sign. interaction of PRS_D and exam stress on depressive symptoms (b = −0.05) or perceived stress (b = 0.04) • no sign. 3-way interactions with time on depressive symptoms (b = 0.20) or perceived stress (b = 0.02) • no main effects of PRS_D (largest b = 0.07) and exam stress (largest b = 0.09), but sign. interaction of exam stress x time on depressive symptoms (b = 0.50) and perceived stress (b = 0.18)
Peter et al., [52]	prospective longitudinal	214 controls (European)	21.0 yrs (6 assessments within 13 months)	77.0	PGC-MDD & 23andMe 2019	≤0.05	exam period	depressive symptoms (HADS), perceived stress (self-report)
Schür et al., [28]	prospective longitudinal	516 soldiers (European)	t1 = 29.6 yrs t2 ≈ 29.7 yrs t3 ≈ 30.2 yrs t4 ≈ 31.2 yrs t5 ≈ 32.2 yrs	8.5	PGC-MDD-2018	13 pTs tested (range: <5x10^-8 to <1)	deployment-related trauma (self-report)	post-deployment depressive symptoms (SCL-90)	• no sign. interaction of PRS_D and deployment-related trauma on post-deployment depressive symptoms (largest β = −0.3087) • no sign. main effects of PRS_D (largest β = −4.4056) and deployment-related trauma (NR) • nominally sign. correlation of PRS_D and deployment-related trauma (b = −0.49)
Turner et al., [50]	prospective longitudinal	331 college freshmen (European)	18-19 yrs (up to 13 assessments within 1 yr)	58.4	PGC-MDD & 23andMe 2019	=1.0	Covid-19 pandemic stress	depressive symptoms (PHQ-9)	• sign. interaction of PRS_D and Covid-19 pandemic stress on depressive symptoms throughout the first freshman year (NR) • sign. main effects of PRS_D (β = 0.521) and Covid-19 pandemic stress (NR)
Wang et al., [31]	cross-sectional	overall sample: 41810 (European)	42.0 yrs	58.9	PGC-MDD & 23andMe 2019	11 pTs tested (range: <5 x 10^-8 to <1)	long-term difficulties (LDI); SLE (LTE); social support (SPF-IL)	depressive symptoms (MINI)	overall sample: • sign. interaction of PRS_D and long-term difficulties (largest β = 0.0541, R² = 0.17%), SLE (largest β = 0.0385, R² = 0.05%) and social support (largest β = −0.0417, R² = 0.09%) on depressive symptoms • sign. main effects of PRS_D (β = 0.1105, R² = 0.66%), long-term difficulties, SLE and social support (largest β = 0.4195, R² = 9.54%) • sign. correlation of PRS_D and long-term difficulties, SLEs, social support (largest r = 0.08)
Wang et al., [31]	cross-sectional	adult subsample: 38945 (European)	44.1 yrs	59.6	PGC-MDD & 23andMe 2019	11 pTs tested (range: <5 x 10^-8 to <1)		depressive symptoms (MINI)	adult subsample: • sign. interaction of PRS_D and long-term difficulties (largest β = 0.0515, R² = 0.15%), SLE (largest β = 0.0324, R² = 0.06%) and social support (largest β = −0.0423, R² = 0.07%) on depressive symptoms • sign. main effects of PRS_D (β = 0.1149, R² = 0.69%), long-term difficulties, SLE and social support (largest β = 0.4350, R² = 10.19)
Adverse and protective social environments in adulthood
Cleary et al., [51]	prospective longitudinal	1011 medical interns (European)	pre-internship t1 = 27.6 yrs during internship t2 ≈ 27.9 yrs t3 ≈ 28.1 yrs t4 ≈ 28.4 yrs t5 ≈ 28.7 yrs	47.6	PGC-MDD & 23andMe 2019 (MDD-PGC2 + 23andMe)	NR	changes in social support during internship (MSPSS)	depressive symptoms (PHQ-9)	• sign. interaction of PRS_D and changes in social support on depressive symptoms during internship (IRR = 0.96 [0.93, 0.98]) • sign. main effects of PRS_D (IRR = 1.14 [1.04,1.24]) and changes in social support (IRR = 0.88 [0.86–0.90])
Cleary et al., [51]	prospective longitudinal	435 widows (European)	t1 = 75.6 yrs (follow-up after loss of a spouse)	70.8	PGC-MDD & 23andMe 2019 (MDD-PGC2 + 23andMe)	NR	changes in social support after loss of a spouse (LBQ)	depressive symptoms (CES-D)	• sign. interaction of PRS_D and changes in social support on depressive symptoms after loss of a spouse (IRR = 0.78 [0.66, 0.92]) • no sign. main effects of PRS_D (IRR = 1.08 [0.96, 1.22]) and changes in social support (IRR = 1.04 [0.89, 1.22])
Colodro-Conde et al., [54]	cross-sectional	5221 twins (European)	35.7 yrs	65.6	PGC-MDD-2013, July 2016* (excluding 23andMe & QIMR data)	8 tested (range: <5 x 10^-8 to p < 1; final pT: p < 1)²	perceived social support (KPSS)	depressive symptoms (DSSI & SCL-90 based)	• sign. interaction of PRS_D and (lack) of social support (largest R² = 0.07%^v) • sign. main effects of PRS_D (largest R² = 0.46%) and (lack) of social support (R² = 3%) • sign. correlation of PRS_D and SLEs (NR)
Giannelis et al., [55]	cross-sectional	52078 (European)	63.6 yrs	52.0	PGC-MDD-2018 (excluding 23andMe and UK Biobank data)	11 pTs tested (range: <5 x 10^-8 to <1; final pT: 0.3)²	family status: living with spouse/partner, number of children (self-report)	lifetime depression (CIDI based self-report)	• no sign. interaction of PRS_D and family status on lifetime depression (NR) • sign. main effects PRS_D (NR) and family status (largest OR = 1.27 [1.14, 1.42]) • no sign. correlation between PRS_D and family status (largest OR = 1.006 [0.99, 1.01])
Steen et al., [56]	prospective longitudinal	19128 (mostly European)	52.4 yrs (16 assessments within 1 yr)	62.3	MDD-2019	<0.05	loneliness (UCLA-L)	MDD (DSM-criteria based diagnostic interviews), depressive symptoms (MINI)	• no sign. interaction between PRS_D and loneliness on MDD (OR = 0.978 [0.950, 1.006]) and depressive symptoms (IRR = 0.994 [0.985, 1.002]) • sign. main effect of PRS_D on MDD (OR = 1.265 [1.151, 1.391]) and on depressive symptoms (IRR = 1.116 [1.087, 1.145]) • sign. main effect of loneliness on MDD (OR = 1.539 [1.482, 1.598]) and depressive symptoms (IRR = 1.271 [1.225, 1.320])
Stringa et al., [57]	prospective longitudinal	cohort 1 (European): 590	t1 = 70.5 yrs (+23 yrs follow-up every 3 yrs)	50.7	PGC-MDD-2018 (excluding 23andMe data)	<0.05 (main analysis) & <0.2 (sensitivity check)	partner status, network size, emotional support (self-report)	depressive symptoms (CES-D)	Meta-analysis across cohorts: • no sign. interaction of PRS_D and partner status (b = −0.014 [−0.077, 0.049]), network size (b = −0.001 [−0.003, 0.002]), and emotional support (b = 0.001 [−0.002, 0.003]) on depressive symptoms over time • sign. main effects of PRS_D (b = 0.053 [0.023, 0.083]), partner status (b = −0.325 [−0.483, −0.166]), network size (b = −0.008 [−0.010, −0.005]), no sign. main effect of social support (b = 0.001 [−0.002, 0.003]) • no sign. correlation of PRS_D and a) partner status, b) network size, c) emotional support (NR)
		cohort 2 (European): 491	t1 = 68.8 yrs (+23 yrs follow-up every 3 yrs)	56.2
		cohort 3 (European): 631	t1 = 60.0 yrs (+13 yrs follow-up every 3 yrs)	52.8
		cohort 4 (European): 567	t1 = 60.5 (+3 yrs follow-up)	51.5
Other adverse and protective environmental exposures in adulthood
Cao et al., [58]	prospective longitudinal	339767 (European)	t1: 56.6 yrs (median follow-up of 8.1 yrs)	52.5	PGC-MDD-2013	none (PRS-CS)	healthy lifestyle factors (self-report)	depression (clinical records)	• no sign. interaction of PRS_D and healthy lifestyle (t1) factors on the incident of depression (largest HR = 1.03 [0.81, 1.32]) • sign. main effects of PRS_D (HR = 1.22 [1.14, 1.30]) and healthy lifestyle factors (largest HR = 3.52 [3.22, 3.84])
Chen & Yang et al., [60]	retrospective longitudinal	247828 (European)	t1: 53.0 yrs (median follow-up of 12.7 yrs)	50.9	GERA-2018	none (PRS-CS)	night shift work (self-report)	MDD (clinical records, self-report)	• no sign. interaction of PRS_D and night shift work on risk of MDD during follow-up (additive: RERI: −0.017 [−0.326, 0.292], multiplicative: NR) • sign. main effects of PRS_D (HR = 1.14 [1.081.21]) and night shift work (largest HR = 1.32 [1.20, 1.45])
Choi, & Zheutlin et al., [59]	prospective longitudinal	7968 Health Care patients (European)	t1 ≈ 57.9 yrs t2 = 59.9 yrs	57.0	MDD-2019 (excluding 23andMe data)	10 pTs tested (range: 5 x 10^-8 to 1.0; final pT: 1.0)²	physical activity in the past year (self-report)	MDD (ICD-9/10 code retrieved from patient data registry)	• no sign. interaction of PRS_D and physical activity (t1) on incident MDD within a two-year time period (OR = 1.05 [0.96, 1.15]) • sign. main effects of PRS_D (OR = 1.20 [1.11, 1.31]) and (lack of) physical activity (OR = 0.83 [0.75, 0.90]) • no sign. correlation of PRS_D and physical activity (r = 0.006)
Fu et al., [62]	prospective longitudinal	14189 MDD 476591 controls (European)	56.4 yrs 56.6 yrs (median follow-up of 8.8 yrs)	63.8 54.1	PGC-MDD & 23andMe 2019	NR (97 SNPs included)	air pollution (geographic information systems)	MDD (clinical records)	• sign. multiplicative (NR) and additive interaction of PRS_D and air pollution on MDD (largest RERI = 0.15 [0.07, 0.23]) • sign. main effects PRS_D (HR = 2.75 [2.43, 3.12]) and air pollution (largest HR = 2.12 [1.82, 2.47])
Kosciuszko et al., [34]	retrospective longitudinal	6202 (British)	65.2 yrs (5 follow-ups over 14 yrs)	52.2	PIR-2021	NR	SES (wealth in adulthood)	depressive symptoms (CES-D)	• sign. interaction of PRS_D and (low) wealth on baseline depressive symptoms (largest β = 0.08 [0.03, 0.13]), but not on rate of change in depressive symptoms (largest β = −0.002 [−0.01, 0.00]) • sign. main effects of PRS_D (β = 0.12 [0.08, 0.17]) and (low) wealth on baseline depressive symptoms (β = 0.77 [0.71, 0.83]) • no sign. main effects of PRS_D (β = −0.004 [−0.02, 0.12]) and low wealth (β = −0.02 [−0.05, 0.001]) on rate of change in depressive symptoms
Lin et al., [61]	retrospective longitudinal	13636 MDD 367340 controls (European)	57.1 yrs 56.8 yrs (continuously followed for up to ≈ 12.6 yrs)	61.7 52.7	GERA-2018	none (PRS-CS)	time spent in outdoor light (self-report)	MDD (clinical records)	• sign. interaction of PRS_D and time spent in outdoor light on MDD (largest HR = 1.05 [0.90, 1.22]) • sign. main effects of PRS_D (HR: 1.08 [1.06, 1.09]) and time spent in outdoor light (largest HR = 1.13 [1.07–1.20])
Sund et al., [88]	cross-sectional	41198 (Norwegian)	54.4 yrs	56.2	MDD-2019	<5 x 10^-8	urbanicity (secondary ecological data)	depressive and anxious symptoms (HADS), mental distress (MHI)	• no sign. interaction of PRS_D and urbanicity on depressive symptoms (largest OR = 0.91 [0.79, 1.05]) and other mental health outcomes • no sign. main effects of PRS_D (largest OR = 1.09 [1.00, 1.18]), sign. main effect of urbanicity (largest OR = 1.34 [1.15, 1.56]) • sign. correlation of PRS_D and urbanicity (NR)
Wesseldijk et al., [63]	cross-sectional	5648 (European)	40.4 yrs	59.1	PGC-MDD & 23andMe 2019	NR	music engagement & achievement (CAQ); lifetime amount of music practice (self-reported)	depression (registry-based ICD-10 diagnosis), depressive symptoms (HSC)	• no sign. interaction of PRS_D and music engagement or music practice on depression/depressive symptoms (NR) • sign. main effects of PRS_D (HR = 1.43) on depression and depressive symptoms (β = 0.08) and of music engagement on depression (HR = 1.25) and depressive symptoms (β = 0.08) • no sign. main effects of music practice or achievement on depression (largest HR = 1.09), but sign. main effect of music practice on depressive symptoms (β = 0.04) • sign. correlation of PRS_D and music engagement (OR = 1.08) and music practice (β = 0.04)
Cumulative lifetime adversity
Coleman et al., [64]	cross-sectional	35269 MDD 63451 HC (European)	64.1 yrs	56.3	PGC-MDD-2018 (excluding UK Biobank data, including 23andMe)	7 pTs tested (range: <0.001 to <0.5; final pT: 0.5)²	lifetime trauma exposure (self-report)	MDD (CIDI based self-report)	• sign. additive (β = 0.011 [0.008, 0.014]), but not multiplicative (OR = 1.01 [1.00, 1.03]) interaction of PRS_D and lifetime trauma on MDD risk • sign. main effect of PRS_D (largest OR = 1.26 [1.24, 1.28]), main effect of lifetime trauma NR • sign. correlation of PRS_D and lifetime trauma (NR)
Lipsky et al., [70]	cross-sectional	1389 veterans	36.4 yrs	17.4	PGC-MDD-2018	1001 PTs tested (range p= 0.0001 to 1); final pT: NR	lifetime trauma (TLEQ)	MDD (SCID-I)	• no sign. interaction of PRS_D and lifetime trauma on MDD (largest OR = 1.03 [0.90, 1.19]) • sign. main effects of PRS_D (largest OR = 1.19 [1.01, 1.39]) and lifetime trauma (OR = 1.22 [1.06, 1.41])
Musliner et al., [68]	retrospective longitudinal	cohort 1 (European) 18532 MDD	10 yrs (continuously followed for up to 21 yrs)	68.5	MDD-2019	=1.0	SLEs (population-based registers)	early-onset depression (clinical records)	overall sample: • sign. multiplicative interaction of PRS_D and SLEs on early-onset depression (HR = 0.96 [0.94, 0.99]) • sign. additive interaction of PRS_D and SLEs on early-onset depression (opposite effect, RERI = 0.09 [0.06, 0.12]) • sign. main effects of PRS_D (HR = 1.35 [1.31, 1.38]) and SLEs (HR = 1.36 [1.33, 1.39]) • sign. correlation of PRS_D and SLEs (HR = 1.09 [1.07, 1.11])
Musliner et al., [68]	retrospective longitudinal	cohort 2 (European) 20184 controls	10 yrs (continuously followed for up to 21 yrs)	49.2	MDD-2019	=1.0	SLEs (population-based registers)	early-onset depression (clinical records)
Peterson et al., [66]	cross-sectional	4785 MDD 4814 controls (Han Chinese)	45.3 yrs 47.7 yrs	100.0	PGC-MDD-2013	<0.2	lifetime trauma/SLE (self-report)	MDD (CIDI based self-report)	• no sign. additive (b = 0.002 [0.000, 0.005], Pseudo R² = 0.03%) and multiplicative (OR = 1.009 [1.000, 1.019], Pseudo R² = 0.06%) interaction between PRS_D and lifetime trauma/SLE on MDD risk • no sign. main effect of PRS_D (OR = 1.004 [1.000, 1.009], largest Pseudo R² = 0.15%), sign. main effect of lifetime trauma/SLE (OR = 6.811 [2.091, 22.293], largest Pseudo R² = 3.13%) • no sign. correlation between PRS_D and lifetime trauma/SLE (largest OR = 1.005 [0.999, 1.011], Pseudo R² = 0.07%)
Thorp et al., [69]	cross-sectional	102182 (European)	56.1 yrs	56.0	PGC-MDD-2018	NR	lifetime trauma (self-report)	current MDD, depressive symptoms (PHQ-9); lifetime MDD (CIDI based self-report)	• sign. additive interaction of PRS_D and lifetime trauma on current MDD (b = 0.158 [0.111, 0.205]), lifetime MDD risk (b = 0.015 [0.008, 0.022]), and six out of nine current depressive symptoms (largest b = 0.026 [0.018, 0.034]) • no sign. multiplicative interaction of PRS_D and lifetime trauma on current MDD (OR = 0.996 [0.964, 1.029]), lifetime MDD risk (OR = 1.009 [0.973, 1.047]), and six out of nine current depressive symptoms (largest OR = 1.010 [0.983, 1.036]) • no sign. interaction of PRS_D and lifetime trauma on specific lifetime MDD symptoms (additive: b = 0.009 [−0.003, 0.022]; multiplicative: largest OR = 1.113 [0.973, 1.273]) • sign. main effects of PRS_D (largest OR = 1.247 [1.226, 1.269]) and lifetime trauma (OR = 4.627 [4.465, 4.795])

Open in a new tab

b unstandardized regression coefficient, CAQ Creative Achievement Questionnaire, CES-D Center for Epidemiologic Studies Depression Scale, CI confidence interval, CIDI Composite International Diagnostic Interview, CIDI-SC Composite International Diagnostic Interview Screening Scale, DSM-criteria Diagnostic and Statistical Manual of Mental Disorders, DSSI Delusions-Symptoms-States Inventory, GHQ General Health Questionnaire, HADS Hospital Anxiety and Depression Scale, HC healthy controls, HLQ Health and Lifestyle questionnaire, HR hazard ratio, HSC Hopkins Symptom Checklist, ICD-9/10 International Classification of Diseases, 9th and 10th revision, IRR incidence rate ratio, KPSS Kessler Perceived Social Support Measure, LBQ Leave-Behind Questionnaire, LDI Long-term Difficulties Inventory, LEQ-B Brief Life Event Questionnaire, LTE List of Threatening Events, MDD major depressive disorder, MHI Mental Health Index, MINI Mini-International Neuropsychiatric Interview, MSPSS Multidimensional Scale of Perceived Social Support, NR not reported, OR odds ratio, PHQ-9 depression subscale of the Patient Health Questionnaire (PHQ), PRS_D polygenic risk score for depression, PRS-CS polygenic prediction method using continuous shrinkage (CS) priors, pT p-value threshold, Pseudo R² goodness-of-fit measure for non-linear regression models, analogous to R² but not directly interpretable as explained variance, r correlation coefficient, R² proportion of variance in the outcome explained by the predictors, RERI Relative Excess Risk due to Interaction, SCAN Schedules for Clinical Assessment in Neuropsychiatry Interview, SCID-I Structured Clinical Interview for DSM-IV-TR Axis I Disorders, SCL-90 Symptom Checklist-90, SES socioeconomic status, sign. significant, SLE stressful life events, SPF-IL Social Production Function Instrument for the Level of Well-being, t timepoint, TLEQ Traumatic Life Events Questionnaire, TSLE total number of stressful life events, UCLA-L short version of Revised UCLA (University of California, Los Angeles) Loneliness Scale, WRAIR MCS Walter Reed Army Institute of Research (WRAIR) Military Cohesion Scales, yr year, yrs years, β standardized regression coefficient.

²final pT: p-value threshold which explained the highest variance and is therefore used in the main analysis;

^vvisual inspection.

A number of studies further applied a pre-post design to investigate GxE interactions on prospective depressive symptom development after exposure to traumatic/stressful events. A first study on 516 soldiers reported no evidence for an interaction of PRS_D and deployment-related trauma on post-deployment depressive symptoms [28]. Comparable results stem from 3079 soldiers participating in the Army Study of Risk and Resilience in Servicemembers (STARRS) that focused on combat stress but also on protective effects of unit cohesion, where no interactions with PRS_D on the incidence of MDD were observed [45]. Besides war stress, other prospective longitudinal studies focused on students during periods of chronic stress exposure in an academic setting. In a sample of 432 law students preparing for their first state examination and controls, an individual’s PRS_D was found to be unrelated to trajectories of depressive symptoms and perceived stress over a 13-month period [52]. In another study on 311 college freshmen, however, Turner et al. reported associations between PRS_D and depressive symptoms throughout the first freshman year during the COVID-19 pandemic as compared to a pre-pandemic cohort [50]. While lower PRS_D predicted a reduced risk for depression during a typical freshman year, no associations were observed during the Covid-19 pandemic, indicating that this genetic advantage may vanish under specific, e.g., unfamiliar stress conditions. Partly conflicting results were obtained in the Intern Health Study comprising 5227 medical interns, where an individual’s PRS_D was more strongly associated with depressive symptoms (β up to 0.036) and MDD diagnosis (OR = 1.40) in the aftermath of physician training stress than at baseline [47]. Likewise, in another longitudinal analysis from the HRS cohort on 8588 older adults, Domingue et al. [46] found that individuals with lower PRS_D were less likely to develop depressive symptoms following the death of their spouse (β up to 0.69).

In summary, we observed substantial evidence for interactions of PRS_D and adult trauma/SLEs on both depressive symptoms [31, 44, 46, 47, 50, 54] and, to a lesser extent, depression diagnosis [47], while other studies suggest their main effects to be additive [23, 45, 49]. Across studies, significant GxE effects were consistently small in magnitude, with standardized regression coefficients ranging from β = 0.028 to 0.69, explained variance from R² = 0.05–0.15%, and a single OR of 1.40.

Adverse and protective social environments in adulthood

We further retrieved studies that investigated the interplay of PRS_D and an individual’s social environment, including negative (e.g., network stressors) and positive (e.g., social support) aspects. Colodro-Conde et al. [54], whose cross-sectional study of 5221 twins has been described above, examined not only personal and network SLEs but also perceived social support, and observed a multiplicative interaction of PRS_D with social support (R² ≈ 0.07%). Merging data from the HRS (n = 435) and the Intern Health Study (n = 1011), Cleary et al. further investigated whether depressive symptoms following developmentally normative life stressors are moderated by self-reported changes in social support [51]. In both cohorts, an individual’s PRS_D was found to interact with changes in social support on depressive symptoms following death of a spouse (IRR = 0.78) or physician training stress (IRR = 0.96). Precisely, individuals with highest PRS_D were most sensitive to a loss of social support in terms of depressive symptoms, but also benefited the most from gaining social support. These findings align with a differential susceptibility framework, where factors that confer risk in negative environments also confer benefits in positive environments [86]. Two other longitudinal studies also showed no evidence of GxE interactions with protective social factors. Stringa et al. [57] meta-analyzed prospective data from four cohorts of the Longitudinal Aging Study of Amsterdam (n = 2279) that included repeated assessments of depressive symptoms over 3–23 year periods and reported independent, but no interaction effects of PRS_D and partner status, social network size or emotional support. Likewise, analyzing data from 52078 UKB participants revealed no moderating effect of an individual’s family status (living with spouse/partner, number of children) on the association of PRS_D and lifetime depression [55]. Finally, in the Lifelines COVID-19 study, Steen et al. [56] performed a high-resolution assessment of subjective feelings of loneliness, a well-known risk factor for mental health problems [53], and depressive symptoms within one year during the pandemic. While both PRS_D and loneliness significantly predicted depressive symptoms and MDD diagnosis during the pandemic, again, no GxE interaction occurred.

To conclude, two studies reported significant interactions of PRS_D with aspects of the social environment (perceived social support: [54]; changes in social support: [51]), yet the majority of studies found additive rather than interactive effects [55–57]. Where significant, interaction effects were consistently small in magnitude (R² = 0.07%; IRRs ≈ 0.78–0.96).

Other adverse and protective environmental exposures in adulthood

A number of GxE studies have focused on other forms of environmental exposures implicated in the pathogenesis of MDD, including both protective (e.g., healthy lifestyle) and risk (e.g., urbanicity) factors.

A first longitudinal study utilizing electronic health records from 7968 older adults revealed no interaction between physical activity in the past year, a likely causal factor in reducing MDD risk, and an individual’s PRS_D on the diagnosis of MDD within a two-year period [59]. Comparable results were obtained in a prospective analysis of 339767 UKB participants who provided information on multiple healthy lifestyle factors, such as regular physical activity, healthy diet and moderate alcohol intake [58]. In this study, a combined healthy lifestyle score and an individual’s PRS_D were independently associated with incident depression during a follow-up of 8 years, with no evidence for GxE interaction. Likewise, an individual’s PRS_D and working night shifts were found to be independent predictors of a higher risk of depression in 247828 UKB participants after a median follow-up of 12.7 years [60]. In contrast, an average of 1.5 h/day spent in outdoor light predicted lower MDD risk in 380976 UKB participants during a 12.6 year time interval independent of PRS_D [61]. Here, evidence for GxE interaction was restricted to intermediate PRS_D levels (HR up to 1.05), indicating a small effect size. Another study on 5648 Swedish twins further reported significant main effects of music engagement and PRS_D on depression/depressive symptoms, with no evidence for GxE interaction [63].

Based on an ongoing discussion regarding urban-rural differences in mental health conditions [87], two other cross-sectional studies analyzed PRS_D interactions with urbanicity and air pollution. Utilizing secondary ecological data from 41198 participants of the Nord-Trøndelag Health study (HUNT), Sund et al. [88] observed that rural residents had increased odds for reporting depressive symptoms, independent of an individual’s PRS_D. Consistent with previous reports of rGE, urbanicity was found to be positively correlated with an individual’s PRS_D. In another longitudinal UKB study (n = 490780), participants with a high PRS_D had a 15% higher risk of developing MDD during a median follow-up of 8.8 years when exposed to high levels of air pollution, with evidence for both additive (RERI up to 0.15) and multiplicative GxE interaction effects.

Regarding SES during adulthood, data from the ELSA study revealed a weak, multiplicative interaction between an individual’s PRS_D and level of wealth for baseline depressive symptoms (β up to 0.08), but not for trajectories of depressive symptom change during a 14-year follow-up period [34]. Here, a 1-SD increase in PRS_D was associated with an increase in the number of symptoms by 0.08 points in n = 6202 older adults with low wealth.

To conclude, data from large cohort studies suggest significant, yet small, interactions of an individual’s PRS_D with specific environmental exposures in adulthood, particularly air pollution (RERI up to 0.15), outdoor light exposure (HR up to 1.05), and socioeconomic status (β up to 0.08) [34, 61, 62]. In contrast, for other exposures such as healthy lifestyle factors, physical activity, night shift work, and music engagement, only additive effects were reported [58–60, 63].

Cumulative lifetime adversity

We further retrieved studies investigating the interplay of PRS_D and cumulative lifetime adversity on depressive phenotypes. Drawing on 9599 participants of the CONVERGE study, Peterson et al. [66] reported no interaction between an individual’s PRS_D and self-reported lifetime trauma/SLE on MDD risk. Notably however, a GWAS stratified by lifetime trauma/SLE within the CONVERGE sample identified three new loci in participants with no history of adversity, where significant GxE effects were observed. In the largest study to date, Coleman et al. [64] contrasted genetic influences on MDD stratified by self-reported lifetime trauma exposure in 35269 MDD patients and 63451 controls from the UKB. Consistent with the stress-diathesis hypothesis, the authors observed a small additive interaction of PRS_D and lifetime trauma on MDD risk (β = 0.011), that was further reflected in a greater SNP-based heritability of MDD in trauma-exposed (24%) compared to non-exposed (12%) individuals. Subsequent simulation studies indicated that this effect was not confounded by a strong, positive rGE between PRS_D and lifetime trauma observed in this sample. A subsequent UKB study (n = 102182) conducted GxE analyses at the level of individual symptoms to account for the large, clinical heterogeneity observed in MDD patients [69]. Here, self-reported lifetime trauma and PRS_D were associated with specific patterns of current depressive symptoms and a positive, additive GxE interaction emerged for six out of nine current depressive symptoms (b up to 0.026), as well as for current (b = 0.158) and lifetime MDD risk (b = 0.015). The authors concluded that respective interactions between PRS_D and lifetime trauma in the UKB cohort are largely driven by overall current depression and not symptom-specific factors. Respective findings were partly confirmed by a longitudinal study comprising 18532 MDD patients and 20184 controls from the iPSYCH2012 cohort [68]. In this study, risk of early-onset depression was found to increase depending on an individual’s PRS_D and self-reported lifetime trauma/SLE load, with evidence for significant GxE interactions on the multiplicative (HR = 0.96) and additive (RERI = 0.09) scales. However, the effect sizes were small and in opposite directions. In contrast, no such GxE interaction between PRS_D and self-reported lifetime trauma on MDD risk was observed in a recent study on 1389 military service Veterans [70].

In summary, large cohort studies ( ~ 100000 participants) provide evidence for significant interactions of PRS_D and lifetime trauma on current and lifetime MDD as well as depressive symptoms, consistent with the stress-diathesis model [64, 69]. Similar findings of small but inconsistent GxE effects were reported in iPSYCH [68], while smaller studies found no evidence [66, 70]. Reported effect sizes were consistently small (β up to 0.158; b up to 0.026; HR = 0.96; RERI = 0.09), and some associations may be partly confounded by rGE [64].

Studies investigating the interaction between PRS_D and environmental exposures on depression-related intermediate phenotypes

Beyond clinical outcomes, several studies have examined PRS_D x environment interactions (PRS_DxE) on intermediate phenotypes relevant to depression pathophysiology, including cognitive and neuropsychological functions, inflammatory and neuroendocrine markers, and brain imaging measures. Although these phenotypes are not specific to depression, they can provide valuable insights into potential pathways and mechanisms underlying GxE effects. Overall, evidence for PRS_DxE interactions across the intermediate phenotypes investigated is limited and inconsistent. While some studies report significant associations such as effects on neonatal brain structure, prefrontal activation under social stress, brain network connectivity, chronic inflammation, and cognitive biases, most large-scale investigations have not found supporting evidence. A detailed overview of these studies is provided in Supplement 3, both in narrative form and as a summary table.

Discussion

Here, we provide a systematic review of studies that have taken a polygenic risk score approach to study GxE interaction on depressive phenotypes. Respective studies investigated a broad range of adverse and protective environmental exposures across the lifespan with a particular focus on trauma, SLEs, social environments and (un)healthy lifestyle factors. While the majority of individual studies reported significant main effects of an individual’s PRS_D and different environmental influences on depressive phenotypes, the overall evidence for PRS_DxE interactions was considerably heterogeneous (as illustrated in a Sankey plot, Supplement 6). Findings of significant PRS_DxE interactions mostly stem from large cohort studies, in particular, when recent environmental and lifetime exposures were considered. Among studies reporting significant GxE effects (29/69), the target sample size had a median of 5224 with an interquartile range (IQR) of 37534 and the corresponding discovery GWAS sample size had a median of 479299 (IQR = 650754). In studies without significant interaction effects, the target samples tended to be smaller (median = 3009, IQR = 7647), as were the discovery GWAS samples (median = 459014, IQR = 453815). These nominal differences should be interpreted with caution given the heterogeneity of environmental exposures and study designs.

Two general conclusions can be drawn from this review. First, PRS_DxE interactions, if at all, add a small amount of explained variance in depressive phenotypes to the corresponding additive model and may thus require large samples to be reliably detected. Bearing in mind that studies summarized in this review rely on PRS_D that at best explained 3.2% of the variance in depression, even moderately-sized samples may lack sufficient power to detect GxE interactions of small effect size. Notably, some of the largest cohort studies on PRS_DxE to date indeed revealed significant interactions between an individual’s PRS_D and CT as well as adult trauma/SLE (n = 38945, [31]), lifetime trauma (n > 100000, [64, 69]) and more specific exposures during adulthood such as time spent in outdoor light (n = 380976, [61]) and air pollution (n = 490780, [62]) on depression. Moreover, data from the UKB indicated that the SNP-based heritability of MDD was twice as large in trauma-exposed as compared to non-exposed individuals (24% vs 12%, [64]). Besides differences in statistical power, another crucial source of heterogeneity between studies derives from different strategies on how to select relevant SNPs for PRS_D calculation, ranging from very restrictive to more liberal p-value thresholds. Importantly, most studies that conducted sensitivity analyses across multiple p-value thresholds, found GxE interaction effects to be highly dependent on the specific threshold used for PRS_D construction (e.g., [36]), highlighting the need to establish a clear consensus. An apparently reasonable strategy applied by most studies was to analyze GxE effects with a PRS_D based on the specific threshold that provided highest accuracy for predicting depressive outcomes (e.g., [49, 55]). However, while significant main effects (of PRS_D) are generally proposed as desirable when testing for GxE [89], there is also evidence that this approach may miss key interaction effects. Intriguingly, a first GWEIS across 25 environments on neuroticism, a trait with strong genetic and phenotypic overlap with depression, suggested that none of the interacting SNPs, genes or gene sets showed any evidence of significant main effects [90].

Second, in a considerable number of studies summarized in this review and beyond [91, 92], different environmental exposures were found to depend on an individual’s PRS_D, indicating significant rGE, with small effect sizes across different metrics ( |r| = 0.058–0.21, |b| = 0.49, |β| = 0.04–0.163, OR = 1.07–1.76, R² = 0.40–0.45% and HR = 1.09). The assumption that genetic predispositions influence an individual’s environment via passive, evocative and active mechanisms is not new [71], but has received new support from GWAS indicating that traditional environmental measures such as self-reported CT are confounded by substantial heritable components [93, 94]. For example, a high PRS_D could inflate the perception and reporting of adverse environments or may even increase the individual’s risk of exposing themselves [71]. As detailed elsewhere, rGE thus constitutes a potential source of spurious GxE interactions [73] and has been found to inflate type I errors in simulation studies, albeit to a very modest extent [27]. Some studies pursued this problem by incorporating SLEs that are expected to be less dependent on the participant’s own behavior in GxE analyses, which were also found to be less heritable in some (e.g., [44, 95]), but not all [92] studies. To minimize rGE due to perception biases, studies may further rely on multi-informant or more objective assessments of environmental exposures, where heritability estimates are suggested to be slightly lower (29% for self-report vs. 26% for informant report [72]). As a general conclusion, reporting of rGE should be mandatory in future GxE studies in order to evaluate the extent of potential confounding.

A limitation of this review refers to the substantial heterogeneity of individual studies regarding sample size, study design, PRS_D calculation (e.g., threshold used), as well as the type and timing of environmental exposures investigated, which hampers an integrative conclusion or meta-analytical approach. For example, cross-sectional studies relying on self-reported exposures are susceptible to mood-congruent recall bias, as depressive states may inflate the reporting of environmental adversity [96, 97]. In addition, depression outcomes were defined in markedly different ways: Lifetime history of depression or MDD status, usually assessed in case–control designs, differs fundamentally from dimensional measures of acute depressive symptoms that reflect short-term severity and may include subclinical cases. Such variation in conceptualization and measurement further hinders direct comparison of results across studies. Moreover, several of the included studies used partially overlapping samples and data sources (e.g., UK Biobank, PGC-derived datasets), with genetic and phenotypic information from the same individuals appearing in more than one study. While these studies often investigated different environmental exposures, such overlap reduces the independence of findings and complicates the interpretation of apparent replication.

While evidence for robust interactions of polygenic liability and environmental exposures on depressive phenotypes remains inconclusive, it is reasonable to assume that PRS_D are only beginning to reach their full potential in terms of predictive accuracy. The estimated SNP heritability and phenotypic variance explained by PRS_D is steadily rising proportional to the sample size of the GWAS they are built upon [15], thereby increasing predictive power to stratify individuals at risk upon specific environmental exposures. This upward trend is expected to continue by integrating genetic risk variants that are not yet captured by traditional PRS_D once whole-genome sequencing data from MDD patients become increasingly available [98]. Since sufficiently predictive PRS_D are a key prerequisite for well-powered GxE analyses, further improvements in score construction are expected to enhance the feasibility and precision of such studies. Besides their potential in aiding risk stratification, pathway-specific PRS based on neuroimaging, proteomic or other multi-omic data might further advance our etiological understanding of underlying biological mechanisms implicated in the interplay of GxE in MDD. In the past years, substantial progress has been made in constructing polygenic scores based on gene transcription profiles targeting specific biological systems implicated in MDD, including inflammation [99], cellular stress responses [100] and neurotransmitter signaling pathways [101, 102] that could be easily combined with measures of environmental exposures in future GxE studies (e.g., [103]). To conclude, recent advances in PRSs construction have the potential to improve the investigation of GxE effects in the pathogenesis of MDD, which may in turn inform more targeted prevention strategies for individuals at elevated risk.

Supplementary information

Supplement^{(391.6KB, docx)}

Acknowledgements

The current study was supported by grants from the German Research Foundation to Nina Alexander (AL 1484/6-1 and Project-ID 521379614 – SFB/TRR 393). The publication was funded by the Open Access budget of MSH Medical School Hamburg.

Author contributions

SI and NA conceived and designed the study. SI and JE conducted the systematic literature search and study selection. SI performed quality assessment under the supervision of NA. Data extraction procedures were developed by all authors. SI carried out data extraction, with NA reviewing all extracted data. SI drafted the manuscript, and NA revised and finalized it. JE and SV contributed to the interpretation of the findings and critically reviewed the manuscript. SI created the visualizations. NA supervised the project. All authors have read and approved the final version of the manuscript.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Sabrina Illius, Email: sabrina.illius@medicalschool-hamburg.de.

Nina Alexander, Email: nina.alexander@uni-marburg.de.

Supplementary information

The online version contains supplementary material available at 10.1038/s41398-025-03793-7.

References

1.Sullivan PF, Neale MC, Kendler KS. Genetic epidemiology of major depression: review and meta-analysis. Am J Psychiatry. 2000;157:1552–62. [DOI] [PubMed] [Google Scholar]
2.Flint J, Kendler KS. The genetics of major depression. Neuron. 2014;81:1214. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Monroe SM, Simons AD. Diathesis-stress theories in the context of life stress research: Implications for the depressive disorders. Psychol Bull. 1991;110:406–25. [DOI] [PubMed] [Google Scholar]
4.Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT Gene. Science. 2003;301:386–9. [DOI] [PubMed] [Google Scholar]
5.Culverhouse RC, Saccone NL, Horton AC, Ma Y, Anstey KJ, Banaschewski T, et al. Collaborative meta-analysis finds no evidence of a strong interaction between stress and 5-HTTLPR genotype contributing to the development of depression. Mol Psychiatry. 2018;23:133–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Hyde CL, Nagle MW, Tian C, Chen X, Paciga SA, Wendland JR, et al. Identification of 15 genetic loci associated with risk of major depression in individuals of European descent. Nat Genet. 2016;48:1031–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Wray NR, Ripke S, Mattheisen M, Trzaskowski M, Byrne EM, Abdellaoui A, et al. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet. 2018;50:668–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Howard DM, Adams MJ, Shirali M, Clarke TK, Marioni RE, Davies G, et al. Genome-wide association study of depression phenotypes in UK Biobank identifies variants in excitatory synaptic pathways. Nat Commun. 2018;9:1470. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Howard DM, Adams MJ, Clarke TK, Hafferty JD, Gibson J, Shirali M, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci. 2019;22:343–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Levey DF, Stein MB, Wendt FR, Pathak GA, Zhou H, Aslan M, et al. Bi-ancestral depression GWAS in the Million Veteran Program and meta-analysis in >1.2 million individuals highlight new therapeutic directions. Nat Neurosci. 2021;24:954–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Arnau-Soler A, Macdonald-Dunlop E, Adams MJ, Clarke T-K, MacIntyre DJ, Milburn K, et al. Genome-wide by environment interaction studies of depressive symptoms and psychosocial stress in UK Biobank and generation scotland. Transl Psychiatry. 2019;9:14. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Dunn EC, Wiste A, Radmanesh F, Almli LM, Gogarten SM, Sofer T, et al. Genome-Wide association study (GWAS) and genome-wide environment interaction study (GWEIS) of depressive symptoms in African American and hispanic/latina women. Depress Anxiety. 2016;33:265–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Ikeda M, Shimasaki A, Takahashi A, Kondo K, Saito T, Kawase K, et al. Genome-Wide environment interaction between depressive state and stressful life events. J Clin Psychiatry. 2016;77:e29–30. [DOI] [PubMed] [Google Scholar]
14.Choi SW, Mak TSH, O’Reilly PF. Tutorial: a guide to performing polygenic risk score analyses. Nat Protoc. 2020;15:2759–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Mitchell BL, Thorp JG, Wu Y, Campos AI, Nyholt DR, Gordon SD, et al. Polygenic risk scores derived from varying definitions of depression and risk of depression. JAMA Psychiatry. 2021;78:1152–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Ripke S, Wray NR, Lewis CM, Hamilton SP, Weissman MM, Breen G, et al. A mega-analysis of genome-wide association studies for major depressive disorder. Mol Psychiatry. 2013;18:497–511. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Chen Y, Pan C, Cheng S, Li C, Zhang H, Zhang Z, et al. The interaction of early life factors and depression-associated loci affecting the age at onset of the depression. Transl Psychiatry. 2022;12:294. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Goltermann J, Redlich R, Grotegerd D, Dohm K, Leehr EJ, Böhnlein J, et al. Childhood maltreatment and cognitive functioning: the role of depression, parental education, and polygenic predisposition. Neuropsychopharmacology. 2021;46:891–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Halldorsdottir T, Piechaczek C, Soares de Matos AP, Czamara D, Pehl V, Wagenbuechler P, et al. Polygenic risk: predicting depression outcomes in clinical and epidemiological cohorts of youths. Am J Psychiatry. 2019;176:615–25. [DOI] [PubMed] [Google Scholar]
20.Iob E, Ajnakina O, Steptoe A. The interactive association of adverse childhood experiences and polygenic susceptibility with depressive symptoms and chronic inflammation in older adults: a prospective cohort study. Psychol Med. 2023;53:1426–36. [DOI] [PubMed] [Google Scholar]
21.Joo YY, Moon SY, Wang HH, Kim H, Lee EJ, Kim JH, et al. Association of genome-wide polygenic scores for multiple psychiatric and common traits in preadolescent youths at risk of suicide. JAMA Netw Open. 2022;5:e2148585. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Kochunov P, Ma Y, Hatch KS, Gao S, Jahanshad N, Thompson PM, et al. Brain-wide versus genome-wide vulnerability biomarkers for severe mental illnesses. Hum Brain Mapp. 2022;43:4970–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Mullins N, Power RA, Fisher HL, Hanscombe KB, Euesden J, Iniesta R, et al. Polygenic interactions with environmental adversity in the aetiology of major depressive disorder. Psychol Med. 2016;46:759–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Musci RJ, Masyn KE, Benke K, Maher B, Uhl G, Ialongo NS. The effects of the interplay of genetics and early environmental risk on the course of internalizing symptoms from late childhood through adolescence. Dev Psychopathol. 2016;28:225–37. [DOI] [PubMed] [Google Scholar]
25.Pitharouli MC, Hagenaars SP, Glanville KP, Coleman JRI, Hotopf M, Lewis CM, et al. Elevated C-reactive protein in patients with depression, independent of genetic, health, and psychosocial factors: results from the UK Biobank. Am J Psychiatry. 2021;178:522–9. [DOI] [PubMed] [Google Scholar]
26.Peyrot WJ, Milaneschi Y, Abdellaoui A, Sullivan PF, Hottenga JJ, Boomsma DI, et al. Effect of polygenic risk scores on depression in childhood trauma. Br J Psychiatry J Ment Sci. 2014;205:113–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Peyrot WJ, Van der Auwera S, Milaneschi Y, Dolan CV, Madden PAF, Sullivan PF, et al. Does childhood trauma moderate polygenic risk for depression? a meta-analysis of 5765 subjects from the psychiatric genomics consortium. Biol Psychiatry. 2018;84:138–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Schür RR, Schijven D, Boks MP, Rutten BPF, Stein MB, Veldink JH, et al. The effect of genetic vulnerability and military deployment on the development of post-traumatic stress disorder and depressive symptoms. Eur Neuropsychopharmacol. 2019;29:405–15. [DOI] [PubMed] [Google Scholar]
29.Smeeth D, May AK, Karam EG, Rieder MJ, Elzagallaai AA, Van Uum S, et al. Risk and resilience in Syrian refugee children: a multisystem analysis. Dev Psychopathol. 2023;35:2275–87. [DOI] [PubMed] [Google Scholar]
30.Taylor KS, Steptoe A, Iob E. The relationship of adverse childhood experiences, hair cortisol, C-reactive protein, and polygenic susceptibility with older adults’ psychological distress during the COVID-19 pandemic. Mol Psychiatry. 2022;27:5038–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Wang R, Lifelines Cohort, Study, Hartman CA, Snieder H. Stress-related exposures amplify the effects of genetic susceptibility on depression and anxiety. Transl Psychiatry. 2023;13:27. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Agerbo E, Trabjerg BB, Børglum AD, Schork AJ, Vilhjálmsson BJ, Pedersen CB, et al. Risk of early-onset depression associated with polygenic liability, parental psychiatric history, and socioeconomic status. JAMA Psychiatry. 2021;78:387–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Armitage JM, Wang RAH, Davis OSP, Haworth CMA. A polygenic approach to understanding resilience to peer victimisation. Behav Genet. 2022;52:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Kosciuszko M, Steptoe A, Ajnakina O. Genetic propensity, socioeconomic status, and trajectories of depression over a course of 14 years in older adults. Transl Psychiatry. 2023;13:68. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Lehto K, Hägg S, Lu D, Karlsson R, Pedersen NL, Mosing MA. Childhood adoption and mental health in adulthood: the role of gene-environment correlations and interactions in the UK Biobank. Biol Psychiatry. 2020;87:708–16. [DOI] [PubMed] [Google Scholar]
36.Nelemans SA, Boks M, Lin B, Oldehinkel T, van Lier P, Branje S, et al. Polygenic risk for major depression interacts with parental criticism in predicting adolescent depressive symptom development. J Youth Adolesc. 2021;50:159–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Perret LC, Boivin M, Morneau-Vaillancourt G, Andlauer TFM, Paquin S, Langevin S, et al. Polygenic risk score and peer victimisation independently predict depressive symptoms in adolescence: results from the Quebec Longitudinal Study of Children Development. J Child Psychol Psychiatry. 2023;64:388–96. [DOI] [PubMed] [Google Scholar]
38.Zhang X, Yan H, Yu H, Zhao X, Shah S, Dong Z, et al. Childhood urbanicity interacts with polygenic risk for depression to affect stress-related medial prefrontal function. Transl Psychiatry. 2021;11:522. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Acosta H, Kantojarvi K, Tuulari JJ, Lewis JD, Hashempour N, Scheinin NM, et al. Sex-specific association between infant caudate volumes and a polygenic risk score for major depressive disorder. J Neurosci Res. 2020;98:2529–40. [DOI] [PubMed] [Google Scholar]
40.Acosta H, Kantojärvi K, Hashempour N, Pelto J, Scheinin NM, Lehtola SJ, et al. Partial support for an interaction between a polygenic risk score for major depressive disorder and prenatal maternal depressive symptoms on infant right amygdalar volumes. Cereb Cortex. 2020;30:6121–34. [DOI] [PubMed] [Google Scholar]
41.Qiu A, Shen M, Buss C, Chong YS, Kwek K, Saw SM, et al. Effects of antenatal maternal depressive symptoms and socio-economic status on neonatal brain development are modulated by genetic risk. Cereb Cortex. 2017;27:3080–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Wang MJ, Dunn EC, Okereke OI, Kraft P, Zhu Y, Smoller JW. Maternal vitamin D status during pregnancy and offspring risk of childhood/adolescent depression: results from the avon longitudinal study of parents and children (ALSPAC). J Affect Disord. 2020;265:255–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Chen LM, Pokhvisneva I, Lahti-Pulkkinen M, Kvist T, Baldwin JR, Parent C, et al. Independent prediction of child psychiatric symptoms by maternal mental health and child polygenic risk scores. J Am Acad Child Adolesc Psychiatry. 2024;63:640–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Arnau-Soler A, Adams MJ, Clarke TK, MacIntyre DJ, Milburn K, Navrady L, et al. A validation of the diathesis-stress model for depression in Generation Scotland. Transl Psychiatry. 2019;9:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Choi KW, Chen CY, Ursano RJ, Sun X, Jain S, Kessler RC, et al. Prospective study of polygenic risk, protective factors, and incident depression following combat deployment in US Army soldiers. Psychol Med. 2020;50:737–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Domingue BW, Liu H, Okbay A, Belsky DW. Genetic heterogeneity in depressive symptoms following the death of a spouse: polygenic score analysis of the U.S. health and retirement study. Am J Psychiatry. 2017;174:963–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Fang Y, Scott L, Song P, Burmeister M, Sen S. Genomic prediction of depression risk and resilience under stress. Nat Hum Behav. 2020;4:111–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Hayes JP, Logue MW, Sadeh N, Spielberg JM, Verfaellie M, Hayes SM, et al. Mild traumatic brain injury is associated with reduced cortical thickness in those at risk for alzheimer’s disease. Brain J Neurol. 2017;140:813–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Musliner KL, Seifuddin F, Judy JA, Pirooznia M, Goes FS, Zandi PP. Polygenic risk, stressful life events and depressive symptoms in older adults: a polygenic score analysis. Psychol Med. 2015;45:1709–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Turner CA, Khalil H, Murphy-Weinberg V, Hagenauer MH, Gates L, Tang Y, et al. The impact of COVID-19 on a college freshman sample reveals genetic and nongenetic forms of susceptibility and resilience to stress. Proc Natl Acad Sci. 2023;120:e2305779120. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Cleary JL, Fang Y, Zahodne LB, Bohnert ASB, Burmeister M, Sen S. Polygenic risk and social support in predicting depression under stress. Am J Psychiatry. 2023;180:139–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Peter HL, Giglberger M, Streit F, Frank J, Kreuzpointner L, Rietschel M, et al. Association of polygenic scores for depression and neuroticism with perceived stress in daily life during a long-lasting stress period. Genes Brain Behav. 2023;22:e12872. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Cacioppo JT, Hughes ME, Waite LJ, Hawkley LC, Thisted RA. Loneliness as a specific risk factor for depressive symptoms: cross-sectional and longitudinal analyses. Psychol Aging. 2006;21:140–51. [DOI] [PubMed] [Google Scholar]
54.Colodro-Conde L, Couvy-Duchesne B, Zhu G, Coventry WL, Byrne EM, Gordon S, et al. A direct test of the diathesis–stress model for depression. Mol Psychiatry. 2018;23:1590–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Giannelis A, Palmos A, Hagenaars SP, Breen G, Lewis CM, Mutz J. Examining the association between family status and depression in the UK Biobank. J Affect Disord. 2021;279:585–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Steen OD, Ori APS, Wardenaar KJ, van Loo HM. Loneliness associates strongly with anxiety and depression during the COVID pandemic, especially in men and younger adults. Sci Rep. 2022;12:9517. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Stringa N, Milaneschi Y, van Schoor NM, Suanet B, van der Lee S, Holstege H, et al. Genetic liability for depression, social factors and their interaction effect in depressive symptoms and depression over time in older adults. Am J Geriatr Psychiatry. 2020;28:844–55. [DOI] [PubMed] [Google Scholar]
58.Cao Z, Yang H, Ye Y, Zhang Y, Li S, Zhao H, et al. Polygenic risk score, healthy lifestyles, and risk of incident depression. Transl Psychiatry. 2021;11:189. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Choi KW, Zheutlin AB, Karlson RA, Wang MJ, Dunn EC, Stein MB, et al. Physical activity offsets genetic risk for incident depression assessed via electronic health records in a biobank cohort study. Depress Anxiety. 2020;37:106–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Chen Y, Yang H, Zhang Y, Zhou L, Lin J, Wang Y. Night shift work, genetic risk, and the risk of depression: a prospective cohort study. J Affect Disord. 2024;354:735–42. [DOI] [PubMed] [Google Scholar]
61.Lin J, Yang H, Zhang Y, Cao Z, Li D, Sun L, et al. Association of time spent in outdoor light and genetic risk with the incidence of depression. Transl Psychiatry. 2023;13:40. [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Fu Z, Liu Q, Liang J, Weng Z, Li W, Xu J, et al. Air pollution, genetic factors and the risk of depression. Sci Total Environ. 2022;850:158001. [DOI] [PubMed] [Google Scholar]
63.Wesseldijk LW, Lu Y, Karlsson R, Ullén F, Mosing MA. A comprehensive investigation into the genetic relationship between music engagement and mental health. Transl Psychiatry. 2023;13:15. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Coleman JRI, Peyrot WJ, Purves KL, Davis KAS, Rayner C, Choi SW, et al. Genome-wide gene-environment analyses of major depressive disorder and reported lifetime traumatic experiences in UK Biobank. Mol Psychiatry. 2020;25:1430–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.García-González J, Ramírez J, Howard DM, Brennan CH, Munroe PB, Keers R. The effects of polygenic risk for psychiatric disorders and smoking behaviour on psychotic experiences in UK Biobank. Transl Psychiatry. 2020;10:330. [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Peterson RE, Cai N, Dahl AW, Bigdeli TB, Edwards AC, Webb BT, et al. Molecular genetic analysis subdivided by adversity exposure suggests etiologic heterogeneity in major depression. Am J Psychiatry. 2018;175:545–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Polimanti R, Kaufman J, Zhao H, Kranzler HR, Ursano RJ, Kessler RC, et al. Trauma exposure interacts with the genetic risk of bipolar disorder in alcohol misuse of US soldiers. Acta Psychiatr Scand. 2018;137:148–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Musliner KL, Andersen KK, Agerbo E, Albiñana C, Vilhjalmsson BJ, Rajagopal VM, et al. Polygenic liability, stressful life events and risk for secondary-treated depression in early life: a nationwide register-based case-cohort study. Psychol Med. 2021;53:217–26. [DOI] [PubMed] [Google Scholar]
69.Thorp JG, Gerring ZF, Colodro-Conde L, Byrne EM, Medland SE, Middeldorp CM, et al. The association between trauma exposure, polygenic risk and individual depression symptoms. Psychiatry Res. 2023;321:115101. [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Lipsky RK, Garrett ME, Dennis MF, Hauser MA, Beckham JC, Ashley-Koch AE, et al. Impact of traumatic life events and polygenic risk scores for major depression and posttraumatic stress disorder on Iraq/Afghanistan Veterans. J Psychiatr Res. 2023;158:15–9. [DOI] [PubMed] [Google Scholar]
71.Plomin R, DeFries JC, Loehlin JC. Genotype-environment interaction and correlation in the analysis of human behavior. Psychol Bull. 1977;84:309–22. [PubMed] [Google Scholar]
72.Kendler KS, Baker JH. Genetic influences on measures of the environment: a systematic review. Psychol Med. 2007;37:615–26. [DOI] [PubMed] [Google Scholar]
73.Purcell S. Variance components models for gene-environment interaction in twin analysis. Twin Res. 2002;5:554–71. [DOI] [PubMed] [Google Scholar]
74.Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097. [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Sharew NT, Clark SR, Schubert KO, Amare AT. Pharmacogenomic scores in psychiatry: systematic review of current evidence. Transl Psychiatry. 2024;14:322. [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Okbay A, Baselmans BML, De Neve JE, Turley P, Nivard MG, Fontana MA, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nat Genet. 2016;48:624–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Cai N, Revez JA, Adams MJ, Andlauer TFM, Breen G, Byrne EM, et al. Minimal phenotyping yields genome-wide association signals of low specificity for major depression. Nat Genet. 2020;52:437–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Privé F, Aschard H, Carmi S, Folkersen L, Hoggart C, O’Reilly PF, et al. Portability of 245 polygenic scores when derived from the UK Biobank and applied to 9 ancestry groups from the same cohort. Am J Hum Genet. 2022;109:12–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Turley P, Walters RK, Maghzian O, Okbay A, Lee JJ, Fontana MA, et al. Multi-trait analysis of genome-wide association summary statistics using MTAG. Nat Genet. 2018;50:229–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Becker J, Burik CAP, Goldman G, Wang N, Jayashankar H, Bennett M, et al. Resource profile and user guide of the polygenic index repository. Nat Hum Behav. 2021;5:1744–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Zhu Z, Zheng Z, Zhang F, Wu Y, Trzaskowski M, Maier R, et al. Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat Commun. 2018;9:224. [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Misztal MC, Tio ES, Mohan A, Felsky D. Interactions between genetic risk for 21 neurodevelopmental and psychiatric disorders and sport activity on youth mental health. Psychiatry Res. 2023;330:115550. [DOI] [PubMed] [Google Scholar]
83.Van den Bergh BRH, van den Heuvel MI, Lahti M, Braeken M, de Rooij SR, Entringer S, et al. Prenatal developmental origins of behavior and mental health: The influence of maternal stress in pregnancy. Prenat Stress Brain Disord Later Life. 2020;117:26–64. [DOI] [PubMed] [Google Scholar]
84.Pet MA, Brouwer-Brolsma EM. The impact of maternal vitamin d status on offspring brain development and function: a systematic review. Adv Nutr. 2016;7:665–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Kendler KS, Karkowski LM, Prescott CA. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry. 1999;156:837–41. [DOI] [PubMed] [Google Scholar]
86.Belsky J, Bakermans-Kranenburg MJ, van IJzendoorn MH. For better and for worse: differential susceptibility to environmental influences. Curr Dir Psychol Sci. 2007;16:300–4. [Google Scholar]
87.Solmi F, Dykxhoorn J, Kirkbride JB. Urban-rural differences in major mental health conditions. In: Mental health and illness in the city. New York, NY, US: Springer Science + Business Media; 2017; 27–132.
88.Sund ER, Van Lenthe FJ, Avendano M, Raina P, Krokstad S. Does urbanicity modify the relationship between a polygenic risk score for depression and mental health symptoms? Cross-sectional evidence from the observational HUNT Study in Norway. J Epidemiol Community Health. 2021;75:420–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
89.Kendler KS, Kessler RC, Walters EE, MacLean C, Neale MC, Heath AC, et al. Stressful life events, genetic liability, and onset of an episode of major depression in women. Am J Psychiatry. 1995;152:833–42. [DOI] [PubMed] [Google Scholar]
90.Werme J, Van Der Sluis S, Posthuma D, De Leeuw CA. Genome-wide gene-environment interactions in neuroticism: an exploratory study across 25 environments. Transl Psychiatry. 2021;11:180. [DOI] [PMC free article] [PubMed] [Google Scholar]
91.Machlitt-Northen S, Keers R, Munroe PB, Howard DM, Trubetskoy V, Pluess M. Polygenic scores for schizophrenia and major depression are associated with psychosocial risk factors in children: evidence of gene–environment correlation. J Child Psychol Psychiatry. 2022;63:1140–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
92.Feurer C, McGeary JE, Brick LA, Knopik VS, Carper MM, Palmer RHC, et al. Associations between depression-relevant genetic risk and youth stress exposure: evidence of gene–environment correlations. J Psychopathol Clin Sci. 2022;131:457–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
93.Dalvie S, Maihofer AX, Coleman JRI, Bradley B, Breen G, Brick LA, et al. Genomic influences on self-reported childhood maltreatment. Transl Psychiatry. 2020;10:38. [DOI] [PMC free article] [PubMed] [Google Scholar]
94.Warrier V, Kwong ASF, Luo M, Dalvie S, Croft J, Sallis HM, et al. Gene–environment correlations and causal effects of childhood maltreatment on physical and mental health: a genetically informed approach. Lancet Psychiatry. 2021;8:373–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
95.Kendler KS, Karkowski LM, Prescott CA. The assessment of dependence in the study of stressful life events: validation using a twin design. Psychol Med. 1999;29:1455–60. [DOI] [PubMed] [Google Scholar]
96.Goltermann J, Meinert S, Hülsmann C, Dohm K, Grotegerd D, Redlich R, et al. Temporal stability and state-dependence of retrospective self-reports of childhood maltreatment in healthy and depressed adults. Psychol Assess. 2023;35:12–22. [DOI] [PubMed] [Google Scholar]
97.Colombo D, Suso-Ribera C, Fernández-Álvarez J, Cipresso P, Garcia-Palacios A, Riva G, et al. Affect recall bias: being resilient by distorting reality. Cogn Ther Res. 2020;44:906–18. [Google Scholar]
98.Sanders SJ, Neale BM, Huang H, Werling DM, An JY, Dong S, et al. Whole genome sequencing in psychiatric disorders: the WGSPD consortium. Nat Neurosci. 2017;20:1661–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
99.Kappelmann N, Czamara D, Rost N, Moser S, Schmoll V, Trastulla L, et al. Polygenic risk for immuno-metabolic markers and specific depressive symptoms: a multi-sample network analysis study. Brain Behav Immun. 2021;95:256–68. [DOI] [PubMed] [Google Scholar]
100.Penner-Goeke S, Bothe M, Rek N, Kreitmaier P, Pöhlchen D, Kühnel A, et al. High-throughput screening of glucocorticoid-induced enhancer activity reveals mechanisms of stress-related psychiatric disorders. Proc Natl Acad Sci. 2023;120:e2305773120. [DOI] [PMC free article] [PubMed] [Google Scholar]
101.Miguel PM, Pereira LO, Barth B, de Mendonça Filho EJ, Pokhvisneva I, Nguyen TTT, et al. Prefrontal cortex dopamine transporter gene network moderates the effect of perinatal hypoxic-ischemic conditions on cognitive flexibility and brain gray matter density in children. atten.-deficithyperactivity. Disord Genes Symptoms. 2019;86:621–30. [DOI] [PubMed] [Google Scholar]
102.Restrepo-Lozano JM, Pokhvisneva I, Wang Z, Patel S, Meaney MJ, Silveira PP, et al. Corticolimbic DCC gene co-expression networks as predictors of impulsivity in children. Mol Psychiatry. 2022;27:2742–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
103.Acosta H, Kantojärvi K, Tuulari JJ, Lewis JD, Hashempour N, Scheinin NM, et al. Association of cumulative prenatal adversity with infant subcortical structure volumes and child problem behavior and its moderation by a coexpression polygenic risk score of the serotonin system. Dev Psychopathol 2023;1–16. [DOI] [PubMed]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement^{(391.6KB, docx)}

[CR1] 1.Sullivan PF, Neale MC, Kendler KS. Genetic epidemiology of major depression: review and meta-analysis. Am J Psychiatry. 2000;157:1552–62. [DOI] [PubMed] [Google Scholar]

[CR2] 2.Flint J, Kendler KS. The genetics of major depression. Neuron. 2014;81:1214. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Monroe SM, Simons AD. Diathesis-stress theories in the context of life stress research: Implications for the depressive disorders. Psychol Bull. 1991;110:406–25. [DOI] [PubMed] [Google Scholar]

[CR4] 4.Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT Gene. Science. 2003;301:386–9. [DOI] [PubMed] [Google Scholar]

[CR5] 5.Culverhouse RC, Saccone NL, Horton AC, Ma Y, Anstey KJ, Banaschewski T, et al. Collaborative meta-analysis finds no evidence of a strong interaction between stress and 5-HTTLPR genotype contributing to the development of depression. Mol Psychiatry. 2018;23:133–42. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Hyde CL, Nagle MW, Tian C, Chen X, Paciga SA, Wendland JR, et al. Identification of 15 genetic loci associated with risk of major depression in individuals of European descent. Nat Genet. 2016;48:1031–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Wray NR, Ripke S, Mattheisen M, Trzaskowski M, Byrne EM, Abdellaoui A, et al. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet. 2018;50:668–81. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Howard DM, Adams MJ, Shirali M, Clarke TK, Marioni RE, Davies G, et al. Genome-wide association study of depression phenotypes in UK Biobank identifies variants in excitatory synaptic pathways. Nat Commun. 2018;9:1470. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Howard DM, Adams MJ, Clarke TK, Hafferty JD, Gibson J, Shirali M, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci. 2019;22:343–52. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Levey DF, Stein MB, Wendt FR, Pathak GA, Zhou H, Aslan M, et al. Bi-ancestral depression GWAS in the Million Veteran Program and meta-analysis in >1.2 million individuals highlight new therapeutic directions. Nat Neurosci. 2021;24:954–63. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Arnau-Soler A, Macdonald-Dunlop E, Adams MJ, Clarke T-K, MacIntyre DJ, Milburn K, et al. Genome-wide by environment interaction studies of depressive symptoms and psychosocial stress in UK Biobank and generation scotland. Transl Psychiatry. 2019;9:14. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Dunn EC, Wiste A, Radmanesh F, Almli LM, Gogarten SM, Sofer T, et al. Genome-Wide association study (GWAS) and genome-wide environment interaction study (GWEIS) of depressive symptoms in African American and hispanic/latina women. Depress Anxiety. 2016;33:265–80. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Ikeda M, Shimasaki A, Takahashi A, Kondo K, Saito T, Kawase K, et al. Genome-Wide environment interaction between depressive state and stressful life events. J Clin Psychiatry. 2016;77:e29–30. [DOI] [PubMed] [Google Scholar]

[CR14] 14.Choi SW, Mak TSH, O’Reilly PF. Tutorial: a guide to performing polygenic risk score analyses. Nat Protoc. 2020;15:2759–72. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Mitchell BL, Thorp JG, Wu Y, Campos AI, Nyholt DR, Gordon SD, et al. Polygenic risk scores derived from varying definitions of depression and risk of depression. JAMA Psychiatry. 2021;78:1152–60. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Ripke S, Wray NR, Lewis CM, Hamilton SP, Weissman MM, Breen G, et al. A mega-analysis of genome-wide association studies for major depressive disorder. Mol Psychiatry. 2013;18:497–511. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Chen Y, Pan C, Cheng S, Li C, Zhang H, Zhang Z, et al. The interaction of early life factors and depression-associated loci affecting the age at onset of the depression. Transl Psychiatry. 2022;12:294. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Goltermann J, Redlich R, Grotegerd D, Dohm K, Leehr EJ, Böhnlein J, et al. Childhood maltreatment and cognitive functioning: the role of depression, parental education, and polygenic predisposition. Neuropsychopharmacology. 2021;46:891–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Halldorsdottir T, Piechaczek C, Soares de Matos AP, Czamara D, Pehl V, Wagenbuechler P, et al. Polygenic risk: predicting depression outcomes in clinical and epidemiological cohorts of youths. Am J Psychiatry. 2019;176:615–25. [DOI] [PubMed] [Google Scholar]

[CR20] 20.Iob E, Ajnakina O, Steptoe A. The interactive association of adverse childhood experiences and polygenic susceptibility with depressive symptoms and chronic inflammation in older adults: a prospective cohort study. Psychol Med. 2023;53:1426–36. [DOI] [PubMed] [Google Scholar]

[CR21] 21.Joo YY, Moon SY, Wang HH, Kim H, Lee EJ, Kim JH, et al. Association of genome-wide polygenic scores for multiple psychiatric and common traits in preadolescent youths at risk of suicide. JAMA Netw Open. 2022;5:e2148585. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Kochunov P, Ma Y, Hatch KS, Gao S, Jahanshad N, Thompson PM, et al. Brain-wide versus genome-wide vulnerability biomarkers for severe mental illnesses. Hum Brain Mapp. 2022;43:4970–83. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Mullins N, Power RA, Fisher HL, Hanscombe KB, Euesden J, Iniesta R, et al. Polygenic interactions with environmental adversity in the aetiology of major depressive disorder. Psychol Med. 2016;46:759–70. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Musci RJ, Masyn KE, Benke K, Maher B, Uhl G, Ialongo NS. The effects of the interplay of genetics and early environmental risk on the course of internalizing symptoms from late childhood through adolescence. Dev Psychopathol. 2016;28:225–37. [DOI] [PubMed] [Google Scholar]

[CR25] 25.Pitharouli MC, Hagenaars SP, Glanville KP, Coleman JRI, Hotopf M, Lewis CM, et al. Elevated C-reactive protein in patients with depression, independent of genetic, health, and psychosocial factors: results from the UK Biobank. Am J Psychiatry. 2021;178:522–9. [DOI] [PubMed] [Google Scholar]

[CR26] 26.Peyrot WJ, Milaneschi Y, Abdellaoui A, Sullivan PF, Hottenga JJ, Boomsma DI, et al. Effect of polygenic risk scores on depression in childhood trauma. Br J Psychiatry J Ment Sci. 2014;205:113–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Peyrot WJ, Van der Auwera S, Milaneschi Y, Dolan CV, Madden PAF, Sullivan PF, et al. Does childhood trauma moderate polygenic risk for depression? a meta-analysis of 5765 subjects from the psychiatric genomics consortium. Biol Psychiatry. 2018;84:138–47. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Schür RR, Schijven D, Boks MP, Rutten BPF, Stein MB, Veldink JH, et al. The effect of genetic vulnerability and military deployment on the development of post-traumatic stress disorder and depressive symptoms. Eur Neuropsychopharmacol. 2019;29:405–15. [DOI] [PubMed] [Google Scholar]

[CR29] 29.Smeeth D, May AK, Karam EG, Rieder MJ, Elzagallaai AA, Van Uum S, et al. Risk and resilience in Syrian refugee children: a multisystem analysis. Dev Psychopathol. 2023;35:2275–87. [DOI] [PubMed] [Google Scholar]

[CR30] 30.Taylor KS, Steptoe A, Iob E. The relationship of adverse childhood experiences, hair cortisol, C-reactive protein, and polygenic susceptibility with older adults’ psychological distress during the COVID-19 pandemic. Mol Psychiatry. 2022;27:5038–48. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Wang R, Lifelines Cohort, Study, Hartman CA, Snieder H. Stress-related exposures amplify the effects of genetic susceptibility on depression and anxiety. Transl Psychiatry. 2023;13:27. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Agerbo E, Trabjerg BB, Børglum AD, Schork AJ, Vilhjálmsson BJ, Pedersen CB, et al. Risk of early-onset depression associated with polygenic liability, parental psychiatric history, and socioeconomic status. JAMA Psychiatry. 2021;78:387–97. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Armitage JM, Wang RAH, Davis OSP, Haworth CMA. A polygenic approach to understanding resilience to peer victimisation. Behav Genet. 2022;52:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Kosciuszko M, Steptoe A, Ajnakina O. Genetic propensity, socioeconomic status, and trajectories of depression over a course of 14 years in older adults. Transl Psychiatry. 2023;13:68. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Lehto K, Hägg S, Lu D, Karlsson R, Pedersen NL, Mosing MA. Childhood adoption and mental health in adulthood: the role of gene-environment correlations and interactions in the UK Biobank. Biol Psychiatry. 2020;87:708–16. [DOI] [PubMed] [Google Scholar]

[CR36] 36.Nelemans SA, Boks M, Lin B, Oldehinkel T, van Lier P, Branje S, et al. Polygenic risk for major depression interacts with parental criticism in predicting adolescent depressive symptom development. J Youth Adolesc. 2021;50:159–76. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Perret LC, Boivin M, Morneau-Vaillancourt G, Andlauer TFM, Paquin S, Langevin S, et al. Polygenic risk score and peer victimisation independently predict depressive symptoms in adolescence: results from the Quebec Longitudinal Study of Children Development. J Child Psychol Psychiatry. 2023;64:388–96. [DOI] [PubMed] [Google Scholar]

[CR38] 38.Zhang X, Yan H, Yu H, Zhao X, Shah S, Dong Z, et al. Childhood urbanicity interacts with polygenic risk for depression to affect stress-related medial prefrontal function. Transl Psychiatry. 2021;11:522. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Acosta H, Kantojarvi K, Tuulari JJ, Lewis JD, Hashempour N, Scheinin NM, et al. Sex-specific association between infant caudate volumes and a polygenic risk score for major depressive disorder. J Neurosci Res. 2020;98:2529–40. [DOI] [PubMed] [Google Scholar]

[CR40] 40.Acosta H, Kantojärvi K, Hashempour N, Pelto J, Scheinin NM, Lehtola SJ, et al. Partial support for an interaction between a polygenic risk score for major depressive disorder and prenatal maternal depressive symptoms on infant right amygdalar volumes. Cereb Cortex. 2020;30:6121–34. [DOI] [PubMed] [Google Scholar]

[CR41] 41.Qiu A, Shen M, Buss C, Chong YS, Kwek K, Saw SM, et al. Effects of antenatal maternal depressive symptoms and socio-economic status on neonatal brain development are modulated by genetic risk. Cereb Cortex. 2017;27:3080–92. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Wang MJ, Dunn EC, Okereke OI, Kraft P, Zhu Y, Smoller JW. Maternal vitamin D status during pregnancy and offspring risk of childhood/adolescent depression: results from the avon longitudinal study of parents and children (ALSPAC). J Affect Disord. 2020;265:255–62. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Chen LM, Pokhvisneva I, Lahti-Pulkkinen M, Kvist T, Baldwin JR, Parent C, et al. Independent prediction of child psychiatric symptoms by maternal mental health and child polygenic risk scores. J Am Acad Child Adolesc Psychiatry. 2024;63:640–51. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Arnau-Soler A, Adams MJ, Clarke TK, MacIntyre DJ, Milburn K, Navrady L, et al. A validation of the diathesis-stress model for depression in Generation Scotland. Transl Psychiatry. 2019;9:25. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Choi KW, Chen CY, Ursano RJ, Sun X, Jain S, Kessler RC, et al. Prospective study of polygenic risk, protective factors, and incident depression following combat deployment in US Army soldiers. Psychol Med. 2020;50:737–45. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Domingue BW, Liu H, Okbay A, Belsky DW. Genetic heterogeneity in depressive symptoms following the death of a spouse: polygenic score analysis of the U.S. health and retirement study. Am J Psychiatry. 2017;174:963–70. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Fang Y, Scott L, Song P, Burmeister M, Sen S. Genomic prediction of depression risk and resilience under stress. Nat Hum Behav. 2020;4:111–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Hayes JP, Logue MW, Sadeh N, Spielberg JM, Verfaellie M, Hayes SM, et al. Mild traumatic brain injury is associated with reduced cortical thickness in those at risk for alzheimer’s disease. Brain J Neurol. 2017;140:813–25. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Musliner KL, Seifuddin F, Judy JA, Pirooznia M, Goes FS, Zandi PP. Polygenic risk, stressful life events and depressive symptoms in older adults: a polygenic score analysis. Psychol Med. 2015;45:1709–20. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Turner CA, Khalil H, Murphy-Weinberg V, Hagenauer MH, Gates L, Tang Y, et al. The impact of COVID-19 on a college freshman sample reveals genetic and nongenetic forms of susceptibility and resilience to stress. Proc Natl Acad Sci. 2023;120:e2305779120. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Cleary JL, Fang Y, Zahodne LB, Bohnert ASB, Burmeister M, Sen S. Polygenic risk and social support in predicting depression under stress. Am J Psychiatry. 2023;180:139–45. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Peter HL, Giglberger M, Streit F, Frank J, Kreuzpointner L, Rietschel M, et al. Association of polygenic scores for depression and neuroticism with perceived stress in daily life during a long-lasting stress period. Genes Brain Behav. 2023;22:e12872. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Cacioppo JT, Hughes ME, Waite LJ, Hawkley LC, Thisted RA. Loneliness as a specific risk factor for depressive symptoms: cross-sectional and longitudinal analyses. Psychol Aging. 2006;21:140–51. [DOI] [PubMed] [Google Scholar]

[CR54] 54.Colodro-Conde L, Couvy-Duchesne B, Zhu G, Coventry WL, Byrne EM, Gordon S, et al. A direct test of the diathesis–stress model for depression. Mol Psychiatry. 2018;23:1590–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Giannelis A, Palmos A, Hagenaars SP, Breen G, Lewis CM, Mutz J. Examining the association between family status and depression in the UK Biobank. J Affect Disord. 2021;279:585–98. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Steen OD, Ori APS, Wardenaar KJ, van Loo HM. Loneliness associates strongly with anxiety and depression during the COVID pandemic, especially in men and younger adults. Sci Rep. 2022;12:9517. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Stringa N, Milaneschi Y, van Schoor NM, Suanet B, van der Lee S, Holstege H, et al. Genetic liability for depression, social factors and their interaction effect in depressive symptoms and depression over time in older adults. Am J Geriatr Psychiatry. 2020;28:844–55. [DOI] [PubMed] [Google Scholar]

[CR58] 58.Cao Z, Yang H, Ye Y, Zhang Y, Li S, Zhao H, et al. Polygenic risk score, healthy lifestyles, and risk of incident depression. Transl Psychiatry. 2021;11:189. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Choi KW, Zheutlin AB, Karlson RA, Wang MJ, Dunn EC, Stein MB, et al. Physical activity offsets genetic risk for incident depression assessed via electronic health records in a biobank cohort study. Depress Anxiety. 2020;37:106–14. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Chen Y, Yang H, Zhang Y, Zhou L, Lin J, Wang Y. Night shift work, genetic risk, and the risk of depression: a prospective cohort study. J Affect Disord. 2024;354:735–42. [DOI] [PubMed] [Google Scholar]

[CR61] 61.Lin J, Yang H, Zhang Y, Cao Z, Li D, Sun L, et al. Association of time spent in outdoor light and genetic risk with the incidence of depression. Transl Psychiatry. 2023;13:40. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Fu Z, Liu Q, Liang J, Weng Z, Li W, Xu J, et al. Air pollution, genetic factors and the risk of depression. Sci Total Environ. 2022;850:158001. [DOI] [PubMed] [Google Scholar]

[CR63] 63.Wesseldijk LW, Lu Y, Karlsson R, Ullén F, Mosing MA. A comprehensive investigation into the genetic relationship between music engagement and mental health. Transl Psychiatry. 2023;13:15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR64] 64.Coleman JRI, Peyrot WJ, Purves KL, Davis KAS, Rayner C, Choi SW, et al. Genome-wide gene-environment analyses of major depressive disorder and reported lifetime traumatic experiences in UK Biobank. Mol Psychiatry. 2020;25:1430–46. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.García-González J, Ramírez J, Howard DM, Brennan CH, Munroe PB, Keers R. The effects of polygenic risk for psychiatric disorders and smoking behaviour on psychotic experiences in UK Biobank. Transl Psychiatry. 2020;10:330. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Peterson RE, Cai N, Dahl AW, Bigdeli TB, Edwards AC, Webb BT, et al. Molecular genetic analysis subdivided by adversity exposure suggests etiologic heterogeneity in major depression. Am J Psychiatry. 2018;175:545–54. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Polimanti R, Kaufman J, Zhao H, Kranzler HR, Ursano RJ, Kessler RC, et al. Trauma exposure interacts with the genetic risk of bipolar disorder in alcohol misuse of US soldiers. Acta Psychiatr Scand. 2018;137:148–56. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Musliner KL, Andersen KK, Agerbo E, Albiñana C, Vilhjalmsson BJ, Rajagopal VM, et al. Polygenic liability, stressful life events and risk for secondary-treated depression in early life: a nationwide register-based case-cohort study. Psychol Med. 2021;53:217–26. [DOI] [PubMed] [Google Scholar]

[CR69] 69.Thorp JG, Gerring ZF, Colodro-Conde L, Byrne EM, Medland SE, Middeldorp CM, et al. The association between trauma exposure, polygenic risk and individual depression symptoms. Psychiatry Res. 2023;321:115101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Lipsky RK, Garrett ME, Dennis MF, Hauser MA, Beckham JC, Ashley-Koch AE, et al. Impact of traumatic life events and polygenic risk scores for major depression and posttraumatic stress disorder on Iraq/Afghanistan Veterans. J Psychiatr Res. 2023;158:15–9. [DOI] [PubMed] [Google Scholar]

[CR71] 71.Plomin R, DeFries JC, Loehlin JC. Genotype-environment interaction and correlation in the analysis of human behavior. Psychol Bull. 1977;84:309–22. [PubMed] [Google Scholar]

[CR72] 72.Kendler KS, Baker JH. Genetic influences on measures of the environment: a systematic review. Psychol Med. 2007;37:615–26. [DOI] [PubMed] [Google Scholar]

[CR73] 73.Purcell S. Variance components models for gene-environment interaction in twin analysis. Twin Res. 2002;5:554–71. [DOI] [PubMed] [Google Scholar]

[CR74] 74.Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.Sharew NT, Clark SR, Schubert KO, Amare AT. Pharmacogenomic scores in psychiatry: systematic review of current evidence. Transl Psychiatry. 2024;14:322. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Okbay A, Baselmans BML, De Neve JE, Turley P, Nivard MG, Fontana MA, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nat Genet. 2016;48:624–33. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.Cai N, Revez JA, Adams MJ, Andlauer TFM, Breen G, Byrne EM, et al. Minimal phenotyping yields genome-wide association signals of low specificity for major depression. Nat Genet. 2020;52:437–47. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR78] 78.Privé F, Aschard H, Carmi S, Folkersen L, Hoggart C, O’Reilly PF, et al. Portability of 245 polygenic scores when derived from the UK Biobank and applied to 9 ancestry groups from the same cohort. Am J Hum Genet. 2022;109:12–23. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR79] 79.Turley P, Walters RK, Maghzian O, Okbay A, Lee JJ, Fontana MA, et al. Multi-trait analysis of genome-wide association summary statistics using MTAG. Nat Genet. 2018;50:229–37. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR80] 80.Becker J, Burik CAP, Goldman G, Wang N, Jayashankar H, Bennett M, et al. Resource profile and user guide of the polygenic index repository. Nat Hum Behav. 2021;5:1744–58. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR81] 81.Zhu Z, Zheng Z, Zhang F, Wu Y, Trzaskowski M, Maier R, et al. Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat Commun. 2018;9:224. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR82] 82.Misztal MC, Tio ES, Mohan A, Felsky D. Interactions between genetic risk for 21 neurodevelopmental and psychiatric disorders and sport activity on youth mental health. Psychiatry Res. 2023;330:115550. [DOI] [PubMed] [Google Scholar]

[CR83] 83.Van den Bergh BRH, van den Heuvel MI, Lahti M, Braeken M, de Rooij SR, Entringer S, et al. Prenatal developmental origins of behavior and mental health: The influence of maternal stress in pregnancy. Prenat Stress Brain Disord Later Life. 2020;117:26–64. [DOI] [PubMed] [Google Scholar]

[CR84] 84.Pet MA, Brouwer-Brolsma EM. The impact of maternal vitamin d status on offspring brain development and function: a systematic review. Adv Nutr. 2016;7:665–78. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR85] 85.Kendler KS, Karkowski LM, Prescott CA. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry. 1999;156:837–41. [DOI] [PubMed] [Google Scholar]

[CR86] 86.Belsky J, Bakermans-Kranenburg MJ, van IJzendoorn MH. For better and for worse: differential susceptibility to environmental influences. Curr Dir Psychol Sci. 2007;16:300–4. [Google Scholar]

[CR87] 87.Solmi F, Dykxhoorn J, Kirkbride JB. Urban-rural differences in major mental health conditions. In: Mental health and illness in the city. New York, NY, US: Springer Science + Business Media; 2017; 27–132.

[CR88] 88.Sund ER, Van Lenthe FJ, Avendano M, Raina P, Krokstad S. Does urbanicity modify the relationship between a polygenic risk score for depression and mental health symptoms? Cross-sectional evidence from the observational HUNT Study in Norway. J Epidemiol Community Health. 2021;75:420–5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR89] 89.Kendler KS, Kessler RC, Walters EE, MacLean C, Neale MC, Heath AC, et al. Stressful life events, genetic liability, and onset of an episode of major depression in women. Am J Psychiatry. 1995;152:833–42. [DOI] [PubMed] [Google Scholar]

[CR90] 90.Werme J, Van Der Sluis S, Posthuma D, De Leeuw CA. Genome-wide gene-environment interactions in neuroticism: an exploratory study across 25 environments. Transl Psychiatry. 2021;11:180. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR91] 91.Machlitt-Northen S, Keers R, Munroe PB, Howard DM, Trubetskoy V, Pluess M. Polygenic scores for schizophrenia and major depression are associated with psychosocial risk factors in children: evidence of gene–environment correlation. J Child Psychol Psychiatry. 2022;63:1140–52. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR92] 92.Feurer C, McGeary JE, Brick LA, Knopik VS, Carper MM, Palmer RHC, et al. Associations between depression-relevant genetic risk and youth stress exposure: evidence of gene–environment correlations. J Psychopathol Clin Sci. 2022;131:457–66. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR93] 93.Dalvie S, Maihofer AX, Coleman JRI, Bradley B, Breen G, Brick LA, et al. Genomic influences on self-reported childhood maltreatment. Transl Psychiatry. 2020;10:38. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR94] 94.Warrier V, Kwong ASF, Luo M, Dalvie S, Croft J, Sallis HM, et al. Gene–environment correlations and causal effects of childhood maltreatment on physical and mental health: a genetically informed approach. Lancet Psychiatry. 2021;8:373–86. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR95] 95.Kendler KS, Karkowski LM, Prescott CA. The assessment of dependence in the study of stressful life events: validation using a twin design. Psychol Med. 1999;29:1455–60. [DOI] [PubMed] [Google Scholar]

[CR96] 96.Goltermann J, Meinert S, Hülsmann C, Dohm K, Grotegerd D, Redlich R, et al. Temporal stability and state-dependence of retrospective self-reports of childhood maltreatment in healthy and depressed adults. Psychol Assess. 2023;35:12–22. [DOI] [PubMed] [Google Scholar]

[CR97] 97.Colombo D, Suso-Ribera C, Fernández-Álvarez J, Cipresso P, Garcia-Palacios A, Riva G, et al. Affect recall bias: being resilient by distorting reality. Cogn Ther Res. 2020;44:906–18. [Google Scholar]

[CR98] 98.Sanders SJ, Neale BM, Huang H, Werling DM, An JY, Dong S, et al. Whole genome sequencing in psychiatric disorders: the WGSPD consortium. Nat Neurosci. 2017;20:1661–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR99] 99.Kappelmann N, Czamara D, Rost N, Moser S, Schmoll V, Trastulla L, et al. Polygenic risk for immuno-metabolic markers and specific depressive symptoms: a multi-sample network analysis study. Brain Behav Immun. 2021;95:256–68. [DOI] [PubMed] [Google Scholar]

[CR100] 100.Penner-Goeke S, Bothe M, Rek N, Kreitmaier P, Pöhlchen D, Kühnel A, et al. High-throughput screening of glucocorticoid-induced enhancer activity reveals mechanisms of stress-related psychiatric disorders. Proc Natl Acad Sci. 2023;120:e2305773120. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR101] 101.Miguel PM, Pereira LO, Barth B, de Mendonça Filho EJ, Pokhvisneva I, Nguyen TTT, et al. Prefrontal cortex dopamine transporter gene network moderates the effect of perinatal hypoxic-ischemic conditions on cognitive flexibility and brain gray matter density in children. atten.-deficithyperactivity. Disord Genes Symptoms. 2019;86:621–30. [DOI] [PubMed] [Google Scholar]

[CR102] 102.Restrepo-Lozano JM, Pokhvisneva I, Wang Z, Patel S, Meaney MJ, Silveira PP, et al. Corticolimbic DCC gene co-expression networks as predictors of impulsivity in children. Mol Psychiatry. 2022;27:2742–50. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR103] 103.Acosta H, Kantojärvi K, Tuulari JJ, Lewis JD, Hashempour N, Scheinin NM, et al. Association of cumulative prenatal adversity with infant subcortical structure volumes and child problem behavior and its moderation by a coexpression polygenic risk score of the serotonin system. Dev Psychopathol 2023;1–16. [DOI] [PubMed]

PERMALINK

The predictive value of polygenic risk scores for depression in gene-environment interaction studies: a systematic review

Sabrina Illius

Julian Eder

Susanne Vogel

Nina Alexander

Abstract

Background

Methods

Results

Conclusion

Introduction

Material and methods

Search strategy

Inclusion and exclusion criteria

Quality assessment

Data extraction

Results

Study selection and characteristics

Fig. 1.

Quality assessment results

Studies investigating the interaction between PRSD and early environmental exposures on depression-related outcomes

Childhood trauma

Table 1.

Other forms of childhood adversity and protective factors

Prenatal adversity

Studies investigating the interaction between PRSD and adult or cumulative environmental exposures on depression-related outcomes

Trauma and stressful life events in adulthood

Table 2.

Adverse and protective social environments in adulthood

Other adverse and protective environmental exposures in adulthood

Cumulative lifetime adversity

Studies investigating the interaction between PRSD and environmental exposures on depression-related intermediate phenotypes

Discussion

Supplementary information

Acknowledgements

Author contributions

Funding

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Studies investigating the interaction between PRS_D and early environmental exposures on depression-related outcomes

Studies investigating the interaction between PRS_D and adult or cumulative environmental exposures on depression-related outcomes

Studies investigating the interaction between PRS_D and environmental exposures on depression-related intermediate phenotypes