Skip to content

Genetics

TL;DR — MDD is moderately heritable and extremely polygenic: liability reflects thousands of variants of tiny effect plus environments and their correlation with genetic liability. A GWAS meta-analysis of 807,553 people identified 102 independent variants and implicated synaptic and prefrontal biology (Howard 2019, PMID 30718901), but no variant is clinically diagnostic. Genetic overlap with bipolar disorder, schizophrenia, anxiety, neuroticism, pain, and sleep helps explain comorbidity while complicating diagnostic specificity. Polygenic scores are useful research variables but currently add little individual-level predictive value beyond clinical history. Candidate-gene and simple gene-by-stress claims should be treated cautiously unless replicated at genome-wide scale.

Architecture at a glance

Feature Evidence Clinical meaning
Family aggregation Depression clusters in families Genes and shared environment both contribute
Twin heritability Moderate, lower than bipolar disorder/schizophrenia Most individual variance is not genetic determination
Common variants Thousands of small effects Large samples required; no single “depression gene”
Rare variants Less well mapped Not a routine explanatory pathway for typical MDD
Cross-disorder sharing Extensive Diagnostic boundaries are not genetic boundaries
Polygenic scores Group-level risk gradients Insufficient discrimination for diagnosis or prescribing

Reviews synthesize twin, family, molecular, and GWAS evidence as a polygenic liability interacting with development and environment (Kendall 2021, PMID 33682643; Flint 2023, PMID 36702864).

A broad disease primer places twin heritability at approximately 35%, while emphasizing recurrent-course and severity heterogeneity (Otte 2016, PMID 27629598). That population estimate is an orientation point, not an individual causal fraction.

GWAS trajectory

Psychiatric GWAS initially struggled because phenotype heterogeneity and small effects demand enormous samples. Howard and colleagues combined cohorts totaling 807,553 participants and identified 102 independent variants, with enrichment in genes expressed in prefrontal regions and in synaptic pathways (Howard 2019, PMID 30718901). The advance is biological mapping, not a clinical test.

Scale problem Consequence
Tiny per-allele effects Odds ratios do not identify individuals
Broad phenotypes More power but less diagnostic precision
European-ancestry dominance Reduced portability across ancestries
Sample overlap Can inflate genetic correlation/prediction
Winner's curse Discovery effects shrink in replication

Shared liability

A cross-disorder GWAS found shared loci across MDD, bipolar disorder, schizophrenia, autism spectrum disorder, and ADHD (Cross-Disorder Group 2013, PMID 23453885). Genetic correlation does not mean identical biology; it indicates overlapping inherited liability. For depression, overlap with neuroticism, sleep traits, pain, smoking, and cardiometabolic traits can represent pleiotropy, mediation, or phenotype contamination.

This sharing explains why a genetic test cannot cleanly distinguish unipolar from bipolar depression. Clinical longitudinal history remains more informative than current polygenic scores.

Heritability is not inevitability

Heritability is population- and environment-specific. It does not state what fraction of one person's episode was “genetic,” nor whether prevention is possible. Environmental variance includes measurement error, idiosyncratic experience, and biological processes not tagged by inherited DNA.

Misinterpretation Correction
“40% heritable” means 40% of a person's depression is genetic Heritability partitions variance in a specified population
A family history proves inherited causation Families share genes, exposures, and care patterns
High polygenic risk predicts inevitable MDD It shifts probability; most outcomes remain uncertain
Genetic association identifies drug target Association-to-mechanism mapping requires functional validation

Gene–environment interplay

Three mechanisms matter: interaction (effects differ by exposure), correlation (genetic liability influences exposure probability), and mediation (one phenotype influences another). Small candidate-gene interaction studies were vulnerable to underpowering and selective reporting. Robust work requires preregistered hypotheses, adequate samples, precise exposure timing, ancestry-aware models, and replication.

Depression genetics therefore does not replace social explanation. Polygenic liability can shape sensitivity to adversity, while poverty, violence, isolation, illness, and care access remain modifiable exposures.

Epigenetic mechanisms offer one route by which developmental exposures may produce persistent transcriptional regulation, but human peripheral-tissue signals are cell-composition-sensitive and do not establish that an exposure caused a brain change (Yuan 2023, PMID 37644009). Epigenetic association should therefore be held to the same replication and causal-inference standards as other biomarkers.

Pharmacogenomics

Commercial panels mostly use pharmacokinetic genes to flag drug–gene interactions; they do not read the polygenic biology of depression. In the PRIME Care randomized trial, pharmacogenomic testing changed prescribing toward fewer predicted drug–gene interactions, but the remission advantage was small and not persistent at 24 weeks (Oslin 2022, PMID 35819423). This is a useful distinction: improving medication compatibility is not the same as predicting which mechanism will work.

Use case Current readiness
Avoiding selected drug–gene interactions Limited, actionable for some drugs/alleles
Diagnosing MDD Not ready
Distinguishing bipolar from unipolar depression Not ready
Selecting SSRI vs psychotherapy vs TMS Not ready
Population etiologic research Ready and productive

From loci to mechanisms

The route from association to treatment is long: fine-map credible variants, identify target genes and cell types, establish developmental timing, test causal perturbations, and validate human relevance. Depression loci often sit in regulatory regions and may affect multiple genes. Multi-omic integration can prioritize hypotheses but also amplifies analytic flexibility.

Equity and portability

GWAS discovery has disproportionately represented European ancestries. Because linkage disequilibrium and allele frequency differ, polygenic-score accuracy falls in underrepresented populations. Using poorly portable scores clinically could widen disparities. Diverse recruitment and locally validated models are scientific requirements, not optional representation exercises.

Genetic evidence deepening

The scale of discovery has changed: depression genetics now spans millions of participants and hundreds of loci, yet individual prediction remains weak. Sample size increases locus count; phenotype depth and ancestry diversity determine what those loci mean.

Question Quantitative anchor Interpretation
How familial is MDD? Family-study OR 2.84 (95% CI 2.31–3.49); twin heritability 37% (31–42%) Moderate inherited liability coexists with large person-specific environmental variance
What did scale add? Million Veteran Program meta-analysis exceeded 1.2 million participants and fine-mapped 178 loci Loci nominate pathways, not deterministic alleles
Does diversity matter? Multi-ancestry work found European loci transferred less often than expected European-only scores are not a neutral default
How predictive are scores? A 2025 trans-ancestry study reported at most 5.8% liability variance explained in Europeans Useful for etiologic stratification; insufficient for diagnosis
Is TRD genetically distinct? EHR-derived TRD GWAS found SNP heritability around 2–4.2% and metabolic overlap Surrogate phenotype choice may define the signal

The central controversy is power versus phenotype precision. Broad self-report definitions yield discovery power; deeply characterized recurrent, melancholic, atypical, or treatment-resistant samples may offer specificity but are much smaller.

Additional live-search evidence ledger

The records below were added after full PubMed E-utilities retrieval on 2026-08-30. The ledger states the evidentiary role of each record and preserves the design limitation that should travel with its citation.

  • Grotzinger AD 2026 — Mapping the genetic landscape across 14 psychiatric disorders. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Grotzinger AD 2026, PMID 41372416)

  • Lyu N 2025 — Genome-wide association and DNA methylation analyses of SSRI treatment response in major depressive disorder. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Lyu N 2025, PMID 41152819)

  • Kang J 2024 — Genome-Wide Association Study of Treatment-Resistant Depression: Shared Biology With Metabolic Traits. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Kang J 2024, PMID 38745458)

  • Mitchell BL 2022 — The Australian Genetics of Depression Study: New Risk Loci and Dissecting Heterogeneity Between Subtypes. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Mitchell BL 2022, PMID 34924174)

  • Levey DF 2021 — Bi-ancestral depression GWAS in the Million Veteran Program and meta-analysis in >1.2 million individuals highlight new therapeutic directions. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Levey DF 2021, PMID 34045744)

  • Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium. 2025 — Trans-ancestry genome-wide study of depression identifies 697 associations implicating cell types and pharmacotherapies. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium. 2025, PMID 39814019)

  • Meng X 2024 — Multi-ancestry genome-wide association study of major depression aids locus discovery, fine mapping, gene prioritization and causal inference. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Meng X 2024, PMID 38177345)

  • Nievergelt CM 2024 — Genome-wide association analyses identify 95 risk loci and provide insights into the neurobiology of post-traumatic stress disorder. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Nievergelt CM 2024, PMID 38637617)

  • Docherty AR 2023 — GWAS Meta-Analysis of Suicide Attempt: Identification of 12 Genome-Wide Significant Loci and Implication of Genetic Risks for Specific Health Factors. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Docherty AR 2023, PMID 37777856)

  • Lorincz-Comi N 2026 — Combining xQTL and genome-wide association studies from diverse populations improves druggable gene discovery. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Lorincz-Comi N 2026, PMID 41690969)

  • Liu HY 2025 — Bidirectional causal relationship between depression and Type 2 diabetes: a multi-ancestry and sex stratified Mendelian Randomization analysis. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Liu HY 2025, PMID 40659072)

  • Zhou H 2023 — Multi-ancestry study of the genetics of problematic alcohol use in over 1 million individuals. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Zhou H 2023, PMID 38062264)

  • Sullivan PF 2000 — Genetic epidemiology of major depression: review and meta-analysis. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Sullivan PF 2000, PMID 11007705)

  • Suktas A 2024 — Genetic polymorphism involved in major depressive disorder: a systemic review and meta-analysis. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Suktas A 2024, PMID 39438912)

  • Coleman JRI 2022 — Editorial: Genome-wide Association Studies of Internalizing Symptoms: A Big Step on a Long Road. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Coleman JRI 2022, PMID 35487336)

  • Perlis RH 2010 — Genome-wide association study of suicide attempts in mood disorder patients. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Perlis RH 2010, PMID 21041247)

  • Hamet P 2005 — Genetics and genomics of depression. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Hamet P 2005, PMID 15877306)

  • Itokawa M 2007 — [Molecular biology of depressive disorders]. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Itokawa M 2007, PMID 17876981)

Open questions

  • Can symptom- or course-defined phenotypes yield genetic signals more predictive than broad MDD while retaining adequate power (Howard 2019, PMID 30718901)?
  • Which associated loci converge on perturbable cell types and developmental windows (Flint 2023, PMID 36702864)?
  • Can multi-ancestry scores achieve equitable calibration and meaningful incremental value over clinical history?
  • Which pharmacogenomic decisions improve durable patient outcomes rather than only medication matching (Oslin 2022, PMID 35819423)?

References

  1. Kendall KM, et al. The genetic basis of major depression. Psychological Medicine. 2021. PMID 33682643
  2. Flint J, Kendler KS. The genetic basis of major depressive disorder. Molecular Psychiatry. 2023. PMID 36702864
  3. Howard DM, et al. Genome-wide meta-analysis of depression identifies 102 independent variants. Nature Neuroscience. 2019. PMID 30718901
  4. Cross-Disorder Group of the Psychiatric Genomics Consortium. Identification of risk loci with shared effects on five major psychiatric disorders. Lancet. 2013. PMID 23453885
  5. Oslin DW, et al. Effect of Pharmacogenomic Testing for Drug-Gene Interactions on Medication Selection and Remission in MDD. JAMA. 2022. PMID 35819423
  6. Otte C, et al. Major depressive disorder. Nature Reviews Disease Primers. 2016. PMID 27629598
  7. Yuan M, et al. Epigenetic regulation in major depression and other stress-related disorders. Signal Transduction and Targeted Therapy. 2023. PMID 37644009
  8. Grotzinger AD, et al. Mapping the genetic landscape across 14 psychiatric disorders. Nature. 2026;649:406-415. PMID 41372416
  9. Lyu N, et al. Genome-wide association and DNA methylation analyses of SSRI treatment response in major depressive disorder. BMC psychiatry. 2025;25:1030. PMID 41152819
  10. Kang J, et al. Genome-Wide Association Study of Treatment-Resistant Depression: Shared Biology With Metabolic Traits. The American journal of psychiatry. 2024;181:608-619. PMID 38745458
  11. Mitchell BL, et al. The Australian Genetics of Depression Study: New Risk Loci and Dissecting Heterogeneity Between Subtypes. Biological psychiatry. 2022;92:227-235. PMID 34924174
  12. Levey DF, et al. Bi-ancestral depression GWAS in the Million Veteran Program and meta-analysis in >1.2 million individuals highlight new therapeutic directions. Nature neuroscience. 2021;24:954-963. PMID 34045744
  13. Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium., et al. Trans-ancestry genome-wide study of depression identifies 697 associations implicating cell types and pharmacotherapies. Cell. 2025;188:640-652.e9. PMID 39814019
  14. Meng X, et al. Multi-ancestry genome-wide association study of major depression aids locus discovery, fine mapping, gene prioritization and causal inference. Nature genetics. 2024;56:222-233. PMID 38177345
  15. Nievergelt CM, et al. Genome-wide association analyses identify 95 risk loci and provide insights into the neurobiology of post-traumatic stress disorder. Nature genetics. 2024;56:792-808. PMID 38637617
  16. Docherty AR, et al. GWAS Meta-Analysis of Suicide Attempt: Identification of 12 Genome-Wide Significant Loci and Implication of Genetic Risks for Specific Health Factors. The American journal of psychiatry. 2023;180:723-738. PMID 37777856
  17. Lorincz-Comi N, et al. Combining xQTL and genome-wide association studies from diverse populations improves druggable gene discovery. Nature communications. 2026;17:2801. PMID 41690969
  18. Liu HY, et al. Bidirectional causal relationship between depression and Type 2 diabetes: a multi-ancestry and sex stratified Mendelian Randomization analysis. Journal of affective disorders. 2025;390:119879. PMID 40659072
  19. Zhou H, et al. Multi-ancestry study of the genetics of problematic alcohol use in over 1 million individuals. Nature medicine. 2023;29:3184-3192. PMID 38062264
  20. Sullivan PF, et al. Genetic epidemiology of major depression: review and meta-analysis. The American journal of psychiatry. 2000;157:1552-62. PMID 11007705
  21. Suktas A, et al. Genetic polymorphism involved in major depressive disorder: a systemic review and meta-analysis. BMC psychiatry. 2024;24:716. PMID 39438912
  22. Coleman JRI. Editorial: Genome-wide Association Studies of Internalizing Symptoms: A Big Step on a Long Road. Journal of the American Academy of Child and Adolescent Psychiatry. 2022;61:864-865. PMID 35487336
  23. Perlis RH, et al. Genome-wide association study of suicide attempts in mood disorder patients. The American journal of psychiatry. 2010;167:1499-507. PMID 21041247
  24. Hamet P, et al. Genetics and genomics of depression. Metabolism: clinical and experimental. 2005;54:10-5. PMID 15877306
  25. Itokawa M, et al. [Molecular biology of depressive disorders]. Nihon rinsho. Japanese journal of clinical medicine. 2007;65:1599-606. PMID 17876981