henhaobio
Sign in

Nature GeneticsGeneticsPeer-reviewed journal article

Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression

eQTLGen, Naomi R. WrayiD, 23andMe, Stephan RipkeiD, Manuel MattheiseniD, Maciej TrzaskowskiiD, Enda M. ByrneiD, Abdel AbdellaouiiD, Mark J. AdamsiD, Esben AgerboiD, Tracy AiriD, Till F. M. AndlaueriD, Silviu‐Alin BacanuiD, Marie Bækvad‐HanseniD, Aartjan T.F. BeekmaniD, Tim B. BigdeliiD, Elisabeth B. BinderiD, Douglas BlackwoodiD, Julien BryoisiD, Henriette N. ButtenschøniD, Jonas Bybjerg‐GrauholmiD, Na CaiiD, Enrique CastelaoiD, Jane ChristenseniD, Toni‐Kim ClarkeiD, , Niamh MullinsiD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

Major depressive disorder (MDD) is a notably complex illness with a lifetime prevalence of 14%.1 It is often chronic or recurrent and is thus accompanied by considerable morbidity, excess mortality, substantial costs, and heightened risk of suicide.2-7 MDD is a major cause of disability worldwide.8 We conducted a genome-wide association (GWA) meta-analysis in 130,664 MDD cases and 330,470 controls, and identified 44 independent loci that met criteria for statistical significance. We present extensive analyses of these results which provide new insights into the nature of MDD. The genetic findings were associated with clinical features of MDD, and implicated prefrontal and anterior cingulate cortex in the pathophysiology of MDD (regions exhibiting anatomical differences between MDD cases and controls). Genes that are targets of antidepressant medications were strongly enriched for MDD association signals (P=8.5x10-10), suggesting the relevance of these findings for improved pharmacotherapy of MDD. Sets of genes involved in gene splicing and in creating isoforms were also enriched for smaller MDD GWA P-values, and these gene sets have also been implicated in schizophrenia and autism. Genetic risk for MDD was correlated with that for many adult and childhood onset psychiatric disorders. Our analyses suggested important relations of genetic risk for MDD with educational attainment, body mass, and schizophrenia: the genetic basis of lower educational attainment and higher body mass were putatively causal for MDD whereas MDD and schizophrenia reflected a partly shared biological etiology. All humans carry lesser or greater numbers of genetic risk factors for MDD, and a continuous measure of risk underlies the observed clinical phenotype. MDD is not a distinct entity that neatly demarcates normalcy from pathology but rather a useful clinical construct associated with a range of adverse outcomes and the end result of a complex process of intertwined genetic and environmental effects. These findings help refine and define the fundamental basis of MDD.

In-depth summary

Claude Opus 5.5 · xhigh effort · 5 Nov 2026 · from the full text of version 1 (24 Jul 2017) · CC BY-NC-ND 4.0

Genome-wide association meta-analysis of 130,664 major depression cases and 330,470 controls identifies 44 independent risk loci enriched in cortical neurons, conserved mammalian regions, splicing networks, and antidepressant targets.

Earlier genome-wide association studies of major depressive disorder (MDD) struggled to detect susceptibility loci because common risk variants have small individual effect sizes and the disorder is etiologically heterogeneous. Combining 29 clinically ascertained cohorts with six large expanded cohorts reveals 44 risk loci and links genetic liability to cortical neurons, alternative splicing networks, and antidepressant targets.

Key results

  1. ShownMeta-analysis of seven strongly correlated European-ancestry MDD cohorts identified 44 independent genome-wide significant loci, including 30 not previously reported, and estimated SNP heritability at 8.9%.

    Across 130,664 MDD cases and 330,470 controls (9.6 million imputed SNPs; weighted mean genetic correlation rg = 0.76, SE 0.028 between anchor and expanded cohorts; LDSC intercept 1.018, SE 0.009), 44 loci reached P < 5 x 10^-8 (led by OLFM4 and NEGR1, plus two independent signals in RBFOX1), with liability-scale SNP heritability estimated at 8.9% (SE 0.004, assuming 15% lifetime risk). · Abstract; Analysis of MDD anchor with six expanded cohorts…; Implications of the individual loci for the…

  2. ShownOut-of-sample genetic risk scores significantly predicted clinical MDD status and were elevated in patients with earlier onset, higher symptom count, and recurrent or chronic illness.

    Across the 29 anchor samples, genetic risk scores explained 1.9% of liability variance, ranked cases above controls with probability 0.57, and gave an OR of 2.4 for the 10th vs 1st decile, with higher scores in early vs late onset, severe (>=8 criteria) vs moderate (5-7) MDD, recurrent (N=5,574) vs single-episode (N=12,968) MDD in iPSYCH, and chronic Stage 4 (n=705) vs Stage 2 (n=388) MDD in NESDA. · Analysis of MDD anchor with six expanded cohorts…; Genetic risk score (GRS) analyses; section 33; Figure legends

  3. ShownMDD heritability was enriched in prefrontal and anterior cingulate cortex, cortical neurons, and mammalian-conserved regulatory regions, with Hi-C and eQTL data localizing target genes.

    Partitioned LDSC enriched only in GTEx brain tissues (led by prefrontal and anterior cingulate cortex), neurons, and regions conserved across 29 Eutherian mammals (20.9-fold, P = 1.4 x 10^-15) plus brain open chromatin and H3K4me1, not exons; SMR and prefrontal TWAS identified 13 and 17 eQTL genes, and cortical Hi-C (q < 10^-4) linked rs1432639 (P = 4.6 x 10^-15) and rs12129573 (P = 4.0 x 10^-12) to NEGR1. · Implications for the biology of MDD using…; Relation of MDD GWA findings to tissue and…; DNA looping using Hi-C; Figure legends

  4. ShownGene-set analyses implicated 19 biological pathways, including RNA splicing regulators and synaptic genes, and showed strong enrichment for known antidepressant drug targets.

    MAGMA identified 153 significant genes (P < 2.6 x 10^-6; 45 in the extended MHC) and 19 FDR < 0.05 pathways (including RBFOX1/2/3 and CELF4 regulatory networks, FMRP targets, synaptic/neuronal morphogenesis genes, and voltage-gated calcium channels), while 42 antidepressant target gene sets were enriched for smaller MDD association P-values (rank enrichment P = 8.5 x 10^-10). · Abstract; Implications of the individual loci for the…; Implications for the biology of MDD based on the…; Gene-wise and pathway analysis

  5. ShownMendelian randomization supported putative causal effects of higher body mass index and lower educational attainment on MDD and bidirectional shared etiology with schizophrenia.

    Across 221 traits, MDD correlated with psychiatric traits, depressive symptoms (rg = 0.80, SE 0.059 in anchor), sleep, neuroticism, BMI, education, and CAD (rg = 0.12); GSMR showed a 1.15-fold MDD increase per SD of BMI (PGSMR = 2.7 x 10^-7) and 0.89-fold decrease per SD of education years (PGSMR = 8.8 x 10^-7) without reverse causality (PGSMR = 0.81, 0.28), plus bidirectional effects with schizophrenia (rg = 0.34). · Implications for a deeper understanding of the…; Mendelian randomization (MR).121; section 33

Methods, data and reproducibility

Performs a fixed-effects GWA meta-analysis of 9.6 million 1000-Genomes-imputed SNPs across 29 clinically interviewed PGC anchor samples and six expanded European-ancestry cohorts using logistic regression in ricopili. Integrates out-of-sample genetic risk scoring, partitioned LDSC, SMR and TWAS/FUSION eQTL mapping, cortical easy Hi-C, MAGMA pathway tests, and bidirectional GSMR.

  • PGC MDD anchor cohort (29 clinically ascertained European-ancestry samples) and 6 expanded cohorts (deCODE, Generation Scotland, GERA, iPSYCH, UK Biobank, 23andMe) · human · n = 130,664 cases, 330,470 controls
  • Independent clinical inpatient MDD validation cohort and screened controls from Munster, Germany · human · n = 845 inpatient MDD cases and 834 controls
  • Postmortem adult dorsolateral prefrontal cortex (European ancestry) and fetal cerebrum (gestation weeks 17-19, African ancestry, NIH NeuroBiobank) for easy Hi-C · human · n = 3 adult and 3 fetal brain samples (~1 billion reads per sample)
  • Dorsolateral prefrontal cortex ATAC-seq open chromatin profiles · human · n = 288 samples (135 controls, 137 schizophrenia, 10 bipolar, 6 affective disorder)
  • eQTL and transcriptome reference cohorts (eQTLGen, BIOS, NESDA/NTR, GEUVADIS, Rosmap, GTEx v6p, and CommonMind Consortium prefrontal cortex) · human · n = eQTLGen N=14,115; BIOS N=2,116; NESDA N=4,896; CMC N=467; Rosmap N=494 · syn5650509
  • 1000 Genomes Phase 1/3 imputation and LD reference panels and Han Chinese CONVERGE female MDD cohort for trans-ancestry comparison · human · n = 2,186 phased haplotypes (1000G); 5,303 cases in CONVERGE
Methods & tools
ricopili, IMPUTE2, SHAPEIT, PLINK, LDSC, LD-Hub, SMR, TWAS/FUSION, MAGMA, GSMR, POPCORN, Bowtie
Software & environment
NCBI Build 37 / UCSC hg19; MAGMA v1.06 / v1.6; MSigDB v5.2; DGIdb v.2.0; R packages pROC, diffbind, and gstat; GTEx v6p [S15, S16, S18, S21].
Reproducibility gaps
Full genome-wide summary statistics for the 23andMe cohort require a separate data access agreement with 23andMe; public PGC summary statistics exclude 23andMe beyond the top 10,000 SNPs [S22].
Limitations (3)
  • All seven primary meta-analysis cohorts were of European ancestry, and neither of the two genome-wide significant loci from the Han Chinese CONVERGE study replicated at P < 1 x 10^-6 in the European dataset (trans-ancestry rg = 0.33, SE 0.03) [S4, S13, S20]. · stated by the authors
  • Forty-five of the 153 significant genes from MAGMA gene-wise analysis fell within the extended MHC region, where high linkage disequilibrium and gene density complicate causal gene localization [S4, S14]. · stated by the authors
  • Too few primate-conserved or human-accelerated genomic regions were available for confident partitioned LDSC evaluation, and MDD-to-coronary-artery-disease Mendelian randomization had high standard error requiring more MDD SNP instruments [S5, S7]. · stated by the authors

Read this if you study the genetic architecture of psychiatric disorders, integrate GWAS with brain functional genomics, or prioritize antidepressant drug targets.

Citations
2,997
FWCI
44.95
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISAls '23Jermy '21Cai '19Werme '23Howard '18Howard '19
85% semantic matchGenetic & Genomic Medicine
Thomas D. Als, Mitja Kurki, Jakob Grove+40 more·Nature Medicine·Jul 1, 2023· 326 citations