The genetics here is real, weak and spread across hundreds of positions. The pharmacogenomics is the useful part, and it answers a narrower question than the tests sold on it claim: how much of a drug you are exposed to, never which antidepressant will work for you.
Depression is common, treatable and heritable in the way most common conditions are: many positions each contributing a very small amount, none of them decisive. The largest studies have found more than a hundred independent variants, and together they explain a small fraction of who becomes depressed. There is no depression gene, and the honest reading of the genetics is that it describes populations rather than people.
What genetics does have to say to an individual is about medicines, and it says less than most people have been told.
It can tell you about dose. Two enzymes, CYP2D6 and CYP2C19, clear most antidepressants. People at the extremes of either — clearing very slowly or very fast — reach concentrations well outside what a standard dose intends. Too much means side effects that make people stop; too little means taking a drug that was never going to reach a working level. CPIC publishes dosing guidance on exactly this.
It cannot tell you which drug will work. This is the distinction that consumer "find the right antidepressant for you" tests blur, and it is worth being blunt about. CPIC's 2023 guideline looked at SLC6A4 and HTR2A — the two genes those tests lean on hardest — and recommended against using them to guide therapy. Whether a particular antidepressant will lift a particular person's depression is not currently predictable from a genotype.
So the useful sentence is narrow and true: genotype can change the dose and the odds of tolerating a drug. Choosing the drug remains a clinical judgement.
Depression is diagnosed by interview rather than measured, and the same diagnosis covers people whose illnesses look quite different. That makes the phenotype noisy, and a noisy phenotype needs enormous samples to find small effects. The studies here ran to hundreds of thousands of people, and the variants they found still shift risk by amounts too small to matter to any individual.
The two positions on this page are narrower still: each comes from a study of depression under a specific circumstance — one in people without trauma exposure, one interacting with stressful life events. Read them as research findings about how the question was asked, not as anything about a reader.
A related but separate page on this site, Depression (Self-Reported Symptoms), covers a different kind of study — a self-report symptom scale across the general population rather than a clinical diagnosis. The two are kept apart because they do not always find the same signal.
Positions joined since this page was written
What this is The text above discusses the variants this page was written around. Since then the catalogue has joined 92 more positions to it, by shared trait or shared paper. They are listed here by the paper each came from; the text does not describe them, and each variant page carries that study's own record.
Meng X et al. 2024, Nature genetics rs12136984 (near LINC01360), rs12374612 (near ASCC3), rs324901 (MIR9-2HG), rs534415 (near SPPL3), rs1526193 (RSRC1), rs7415357 (LINC02796), rs72786643 (ARB2A), rs3091404 (IGSF6), rs9351533 (near ADH5P4), rs12886138 (MTHFD1), rs199929 (CACNA1E), rs2008551 (RAB27B) and 42 more — PMID:38177345
Yao X et al. 2021, Translational psychiatry rs349588 (NIHCOLE), rs1367635 (DCC), rs2029591 (BSN), rs35466861 (AGBL4), rs1848750 (near HIVEP2), rs59283172 (near RN7SKP120), rs12589834 (ESR2), rs2276138 (SPTBN2), rs1414622 (HS6ST3), rs6479487 (FAM120A), rs8035777 (near MRPL46), rs72948506 (ERBB4) — PMID:33479212
Coleman JRI et al. 2020, Biological psychiatry rs113872034 (TUBGCP6), rs61533748 (LINC01830), rs58235352 (near ACADS), rs16851203 (LINC01645), rs10759881 (ASTN2), rs11137850 (near MTND2P8), rs1518367 (PLCL1), rs1828385 (LINC02853), rs34866621 (PCLO), rs55993664 (near PSMC1P5), rs769087 (CACNA1C) — PMID:31926635
Li Z et al. 2023, Frontiers in psychiatry rs853676 (ZSCAN31), rs17689471 (LINC02210-CRHR1), rs199457 (NSF), rs1396862 (MAPT-AS1), rs2303473 (RPL24) — PMID:37426090
GENDEP Investigators 2013, The American journal of psychiatry rs17651119 (MYO10), rs12054895 (near PURPL) — PMID:23377640
Hyde CL et al. 2016, Nature genetics rs4543289 (LINC03000) — PMID:27479909
Verma A et al. 2024, Science (New York, N.Y.) rs17781288 (LRFN5-DT) — PMID:39024449
Chen D et al. 2023, Frontiers in psychiatry rs76025409 (near NIHCOLE) — PMID:38250258
Li X et al. 2018, Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology rs1457257 (near Metazoa_SRP) — PMID:29728651
Grove J et al. 2019, Nature genetics rs16854048 (BEND4) — PMID:30804558
Huider F et al. 2026, Molecular psychiatry rs3818852 (PALMD) — PMID:42230966
Mullins N et al. 2019, The American journal of psychiatry rs45593736 (ARL5B) — PMID:31164008
Coleman JRI et al. 2020, Molecular psychiatry rs11515172 (near LINC03131) — PMID:31969693
2026-05-20 · Complex correlations between mitochondrial DNA variants and gut microbiome in major depressive disorder: a genome-wide association analysis. BMC Psychiatry. 2026. DOI:10.1186/s12888-026-08132-8
Both gut microbiome disturbances and impaired mitochondrial function are independently linked to major depressive disorder (MDD), but how they interact hadn't been studied together. This small pilot study used shotgun metagenomic sequencing on 63 MDD patients and 30 healthy controls, finding 101 differentially enriched gut microbial species between groups, then ran GWAS specifically on mitochondrial DNA variants against those microbiome features. It identified 68 mitochondrial SNPs (mtSNPs) shared between the two groups at Bonferroni-corrected significance, suggesting real -- if not yet mechanistically explained -- genetic crosstalk between mitochondrial variation and gut microbial composition in the context of MDD. Given the small sample size (63 cases), this should be read as an early, hypothesis-generating pilot rather than a definitive finding, but it's a genuinely novel angle connecting two separate MDD research threads (gut-brain axis and mitochondrial dysfunction) that are usually studied apart. This site's major depressive disorder page carries 13 variants; the specific mtSNPs from this study are not currently among them, and would need a larger replication study before being added as established findings.
2026-01-19 · Unraveling the genetic interplay and therapeutic potentials between major depressive disorder and metabolic syndrome: multi-ancestry and multi-trait genome-wide association analyses. BMC Psychiatry. 2026. DOI:10.1186/s12888-026-07816-5
Major depressive disorder (MDD) and metabolic syndrome (MetS) frequently co-occur, and this study used GWAS datasets spanning European and East Asian populations to test the genetic relationship in both directions rather than assuming depression simply follows from metabolic illness or vice versa. It found bidirectional genetic correlations and shared pleiotropic genes between the two conditions, and cross-trait analysis strengthened the statistical robustness of new MDD-associated markers beyond what single-trait analysis alone would support. Using BLISS technology (a method for identifying disease-linked proteins from GWAS data), the study pinpointed proteins significantly linked to both conditions. Mendelian randomization further explored how differing dietary patterns across populations influence MDD and MetS risk. Most practically, computational drug-repurposing analysis flagged two existing drug classes as candidates worth investigating for both conditions: Cytochrome P450 inhibitors and HDAC (histone deacetylase) inhibitors -- neither currently used for depression or metabolic syndrome specifically, making this a genuinely novel repurposing angle rather than confirmation of an existing treatment relationship. The study's own summary did not specify individual gene or protein names in detail beyond the drug-class findings, so this site's major depressive disorder and metabolic syndrome pages record this as shared-biology and drug-repurposing context rather than new named variants.
2026-07-28 · Bright U et al. The genetics of fibromyalgia and its relationships to psychiatric and medical traits. Nature Communications 2026, PMID:42521668
Fibromyalgia shares much of its genetics with depression and PTSD
Fibromyalgia, a chronic pain syndrome, has long been hard to explain. A genome-wide study combining 85,139 cases and 1,642,433 controls across European, African and Latin American ancestry found 10 loci in European-ancestry participants, 12 across ancestries and 45 in a multi-trait analysis with pain, and strong genetic correlations with chronic pain, post-traumatic stress disorder and depression, alongside a moderate negative one with physical activity. Its analyses point to neuronal mechanisms. 16 of its positions already have pages here, four of them on the fibromyalgia page. PMID:42521668.
The enzymes, and what they do to which drugs. CYP2C19 converts tertiary-amine tricyclics into secondary amines; CYP2D6 clears both forms and most SSRIs to some degree. Fluvoxamine, paroxetine and fluoxetine also inhibit these enzymes themselves, so one antidepressant can change how another drug is handled — an interaction that has nothing to do with genotype and is more often the explanation.
What SLC6A4 and HTR2A are, and why they keep appearing. They encode the serotonin transporter and a serotonin receptor, which makes them the obvious candidates and the ones most heavily marketed. The evidence for using them to select a drug has not held up, and the current CPIC guideline recommends against it.
Depression is not one thing. Genetic studies find partial overlap with bipolar disorder, schizophrenia and anxiety — several of the positions in this catalogue are labelled with more than one of those at once. That overlap is a real finding about the categories, not an error in the data.
Nothing here is diagnostic or urgent. If you are unwell, the useful step is a clinician. If you are already on an antidepressant, do not change it on the strength of a page — stopping abruptly is unsafe and the risk is immediate, where any genetic consideration is not.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Major Depressive Disorder comes down to these specific, well-studied positions — not a diagnosis. 105 positions are linked to this page; the ones this page's own text discusses are shown first.
THRB · rs138272922
See detailed info →BMP2 · rs10485715
See detailed info →NEGR1 · rs2422320
See detailed info →MUC21 · rs9368649
See detailed info →PLEK2 · rs9323497
See detailed info →ZSCAN31 · rs853676
See detailed info →near NIHCOLE · rs76025409
See detailed info →LINC02210-CRHR1 · rs17689471
See detailed info →The studies behind these variants recruited participants from different ancestries — a result found in one population doesn't always transfer to another. Based on 16 of 105 linked studies with a resolved discovery ancestry.
Research-derived gene–drug associations only — not a prescription, dosing guide, or medical advice. Always follow your prescriber's guidance.
| Gene | Drug | What the research shows |
|---|---|---|
| CYP2D6 | Tricyclic antidepressants (amitriptyline, nortriptyline, clomipramine, desipramine, doxepin, imipramine, trimipramine) | Tricyclics are handled by two enzymes working in sequence: CYP2C19 turns a tertiary amine such as amitriptyline, clomipramine, doxepin, imipramine or trimipramine into its secondary amine — which is itself an active drug, and in two cases a drug prescribed in its own right, nortriptyline and desipramine — and CYP2D6 clears both. A CYP2D6 poor metaboliser accumulates more drug than intended and is more likely to stop because of dry mouth, sedation or a racing heart; an ultrarapid metaboliser may never reach a useful level. A CYP2C19 poor metaboliser shifts the balance towards the parent drug. CPIC recommends avoiding a tricyclic or halving the starting dose at these extremes, with therapeutic drug monitoring if one is used anyway. What this does not do is say which antidepressant will work. It is about dose and side effects; efficacy is a clinical judgement no genotype currently predicts. Nothing here should be used to start, stop or change a medicine — stopping an antidepressant abruptly is unsafe. (CPIC Guideline for CYP2D6 and CYP2C19 Genotypes and Dosing of Tricyclic Antidepressants: 2016 update (Clin Pharmacol Ther 2017, PMID 27997040); original guideline PMID 23486447.) |
Free to quote and reuse under CC BY 4.0. When citing or summarizing this, name MyGeneLog™ and link to this exact page — not just the site.
Major Depressive Disorder. MyGeneLog™. https://www.mygenelog.com/conditions/major-depressive-disorder
No, and the tests sold on that promise lean hardest on two genes — SLC6A4 and HTR2A — that CPIC’s 2023 guideline recommends against using to guide therapy. What genotype can inform is dose and the likelihood of side effects, through CYP2D6 and CYP2C19. Which drug helps remains a clinical judgement.
They tell you essentially nothing on their own. Depression is spread across hundreds of positions of very small effect, and these two come from studies of depression under particular circumstances rather than of depression in general.
Partly. Twin studies usually put heritability around a third — real, and much lower than schizophrenia or bipolar disorder. Most people with a depressed parent or sibling never develop it.
Not on the strength of a page. Stopping an antidepressant abruptly is unsafe, and that risk is immediate where a genetic consideration is not. If a result seems relevant, take it to the person prescribing.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog™. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.