You might have heard people talk about ‘the depression gene’ or ‘the schizophrenia gene’ – these comments are misleading. While there is substantial evidence for a genetic component to mental health disorders, we now understand that there are complex influences at play, which go far beyond the activity of any single gene.
Psychiatric disorders overlap to a degree, both in terms of sharing symptoms and sharing associations with particular genetic variants. Previous analyses of schizophrenia and bipolar disorder found more than 100 pleiotropic loci (Lee et al., 2019). This means that genetic variation in the same location in the genome was associated with risk for both conditions.
Currently, diagnoses are based on reported signs and symptoms rather than underlying pathophysiology (Rush & Hisham, 2018). This approach works pragmatically, but it does not necessarily reflect the underlying biology of these conditions. If we better understand the mechanisms shared across disorders, then perhaps we can identify more precise treatment targets to treat disorders more effectively. Mapping out the genetic similarities and differences between the most common psychiatric disorders is a good place to start.
In a third major study from the Psychiatric Genomics Consortium Cross-Disorder Working Group (CDG3), multiple statistical and functional analyses were used to investigate genetic associations (i.e. shared variants) between 14 different psychiatric disorders (Grotzinger et al., 2025). By identifying which genetic factors are shared and which are unique, the study set out to improve our understanding of the biological connections between mental health conditions.
Psychiatric disorders may look distinct, but their underlying biology often overlaps.
Methods
Eight disorders from the Consortium’s previous study were included (ADHD, anorexia nervosa, autism spectrum disorder, bipolar disorder, major depression, OCD, schizophrenia and Tourette’s syndrome), alongside six new ones (alcohol use disorder, anxiety disorders, PTSD, nicotine dependence, opioid use disorder and cannabis use disorder). Average cases increased by roughly 165% compared to the previous study.
Causal modelling was used to explore genetic associations between disorders. Findings were supported by Gene Ontology enrichment, which allowed researchers to link genetic variants to the biological functions of the genes they were found in. This helps to illustrate how differences might affect key processes in the body.
Sample sizes and statistical power varied across the 14 disorders. Data included participants from different ancestral backgrounds, but representation was uneven. Analyses were restricted to European ancestry to allow comparison with global reference panels that help fill in any gaps in the study data, thus improving the statistical power.
Results
The case sample size was 1,056,201. First, the authors used linkage disequilibrium (which measures combinations of genetic variants that are inherited together) to search for ‘genetic overlap’ between psychiatric disorders. The European ancestry group found overlap between disorders meaning that some genome-wide variants were common to multiple conditions.
Further analysis explored the direction of these shared effects. Most shared variants had concordant effects (i.e. the direction of risk was the same across disorders). Fewer variants had discordant effects, where the direction of risk differed despite sharing variants, or were specific to individual disorders and not shared with others.
Shared genetic patterns
Genomic structural equation modelling of the linkage disequilibrium data grouped the 14 disorders into five factors representing shared genetic patterns:
- Compulsive factor: Anorexia nervosa, OCD, Tourette’s syndrome, anxiety disorders
- SB factor: schizophrenia, bipolar disorder
- Neurodevelopmental factor: ASD, ADHD, Tourette’s syndrome
- Internalising factor: PTSD, major depression, anxiety disorders
- SUD factor: opioid use disorder, cannabis use disorder, alcohol use disorder, nicotine dependence, ADHD
The first four factors were similar to previous models, supporting the robustness of the approach despite increases in sample size and the number of disorders included. Moderate correlations between the five factors suggested that a higher-order factor (the p-factor) may explain common variance across all disorders. The Internalising factor loaded most strongly onto the p-factor (0.95), meaning that internalising disorders shared the most genetic risk with the higher-order, general psychopathology factor.
While measures of pleiotropy provide an average of shared genetic signals across the genome, this overlap is unlikely to be consistent across all regions. A specialised analysis was used to identify ‘hotspots’ where pairs of disorders were particularly likely to match up. The most pleiotropic ‘hotspot’ was found on chromosome 11, which contained 17 significant positive correlations across eight disorders. This region is known to contain a gene cluster which previous studies had already linked to symptoms of psychiatric disorders.
Finally, the authors identified SNPs (single nucleotide polymorphisms, otherwise known as common genetic variants, which may or may not be implicated in dysfunction) associated with the five factors or p-factor. The SB and Internalising factors produced the most ‘hits’, meaning people with these disorders shared the most SNPs. Functional follow-up showed that p-factor genes were more active in processes relating to gene expression. Genes associated with the five lower-order factors were found to be active in both foetal and adult brains. For example, SB factor genes were active in foetal interneurons and in adult deep-layer excitatory neurons, suggesting important roles of these genes in developmental neural activity.
Many psychiatric disorders share genetic variants, which cluster into five main factors and a general p-factor, highlighting overlapping biological risk across conditions.
Conclusions
This study surveyed shared genetic influences across 14 psychiatric disorders at the level of the whole genome, smaller genomic regions and individual loci. Initial analyses suggested significant cross-disorder genetic overlap, which was further organised into a five-factor model that grouped correlated disorders into categories.
101 genomic regions were identified where variants were correlated with multiple disorders. The strongest ‘hotspot’ was on chromosome 11, which was linked to eight of the 14 disorders and contained genes previously associated with psychiatric traits (Mota et al., 2015) (Bidwell et al., 2015).
The higher-order p-factor was most strongly related to Internalising disorders. Genes identified by the p-factor model were active in key biological processes, while the five lower-order factors captured more specific genetic signals. The authors suggest that this reflects a transdiagnostic genetic vulnerability: a general risk for psychopathology most strongly expressed in internalising disorders, alongside more targeted influences that shape the symptoms seen in distinct disorders.
These study findings could be used to better understand how common genetic variants can contribute to presentations of psychiatric disorders.
Strengths and limitations
Strengths
This study combined a range of statistical and functional analyses that help us to map the complex genetic landscape of psychiatric disorders. Functional annotation was used to verify findings within the same study, showing a commitment to robust results.
The authors were transparent about the use of self-reported diagnoses. They performed sensitivity analysis excluding these cases and found that the five-factor model still fit the data well.
To place the data into a wider ‘clinical context’, the authors estimated genetic correlations between the five factors, the higher-order p-factor and 31 other traits linked to psychiatric disorders; including memory skills, adult BMI and sleep duration. The Internalising and SUD factors were most closely linked to cognitive outcomes. The inclusion of this measure suggests the authors’ interest in addressing how statistical analyses of genetic data can be contextualised using additional measures.
Limitations
The full range of analyses were only performed using the European dataset, as the African and East Asian datasets were substantially smaller. While the authors found high levels of genetic overlap between disorders in the European population, correlations found in the African population were not significant, likely due to a lack of statistical power. This embodies a general problem across studies that use genetic databases: the data is derived overwhelmingly from people of European ancestry, thereby limiting the generalisability of the findings. As such, researchers and policymakers are thinking about ways to prevent the further perpetuation of health inequalities due to uneven representation in genetic data (NHS, 2024).
However, the authors of this study did take steps to address the issue. They used Popcorn, a Python package used in previous cross-ethnicity genetic studies, to try to generalise their findings across different ancestral groups. The package correlates effect sizes of common genetic variants that are specific to populations of non-European ancestry. Despite this effort, the authors stated that their results were still underpowered for many comparisons. Future studies would benefit from a greater quantity and quality of cross-ancestry genetic data.
The study combined robust statistical and functional analyses to map shared genetic risks, but results were largely based on European ancestry, highlighting the need for more diverse datasets.
Implications for practice
This study provides a detailed update on the genetics of psychiatric disorders across diagnostically distinct conditions. It contributes to ongoing debates around how we should classify psychiatric disorders: if certain disorders are both symptomatically and genetically similar, how do we know that we are drawing accurate boundaries between one disorder and another? Understanding shared genetic signals may encourage thinking about psychiatric conditions more dimensionally, rather than as strictly separate categories.
One big question mark about the study’s utility for policy and practice is the partial exclusion of people of non-European ancestry. Although the authors took certain measures to include genetic data from populations of African and East Asian descent, these were smaller and underpowered analyses. Variants differ in frequency across populations, meaning findings based predominantly on European data may not generalise. Relying on these results to guide clinical decision-making risks marginalising service users from underrepresented ancestries and perpetuating health inequalities (NHS, 2024). Addressing these gaps will require larger, more diverse genetic datasets and continued efforts to remove barriers to participation for minoritised communities, such as language differences and unequal access to research institutions.
On another note, this work could contribute to improving pharmacological interventions for psychiatric disorders. Selective serotonin re-uptake inhibitors (SSRIs) are prescribed across multiple disorders (Murphy et al., 2021). By clarifying which genetic signals are shared across disorders, then perhaps we could routinely stratify conditions according to their genetic background. This way, new or improved therapeutics could be developed that target the identified shared signal. For example, internalising disorders showed the strongest link to the higher order p-factor. In future, interventions could be developed to target shared genetic pathways within such groups, potentially improving efficacy across multiple related conditions. However, the feasibility of stratification in clinical practice will need first to be explored.
The work of the Psychiatric Genomics Consortium’s Cross-Disorder Working Group, at its essence, is an intricate portrayal of the wide-ranging genetic influences on psychiatric disorders and a reminder that to ascribe a psychiatric disorder to a single gene would be inaccurate, even if ‘the depression gene’ makes a snappier news headline than ‘depression’s rich and complex genomic architecture’.
Taken together, this study highlights the potential for psychiatric genetics to influence how we understand, classify, and eventually treat mental health conditions. It encourages a balance between appreciating broad shared mechanisms and recognising disorder-specific signals. At the same time, it underscores the need for inclusivity in research to ensure that insights benefit all populations, and not just those of European ancestry.
The study shows shared genetic risk across psychiatric disorders, highlights the need for more diverse research, and points to future opportunities for stratified, more precise interventions.
Statement of interests
Sophie Webb has no conflicts of interests to declare.
Editor
Edited by Éimear Foley. AI tools assisted with language refinement and formatting during the editorial phase.
Links
Primary paper
Andrew Grotzinger, Josefin Werme, Wouter Peyrot, Oleksandr Frei, Christiaan de Leeuw… Phil Lee, Kenneth Kendler, Jordan Smoller et al (2025) Mapping the genetic landscape across 14 psychiatric disorders. Nature 2025 649 406-415. doi: https://doi.org/10.1038/s41586-025-09820-3
Other references
Lee PH, Anttila V, Won H et al (2019) Genomic relationships, novel loci, and pleiotropic mechanisms across eight psychiatric disorders. Cell 2019 179(7) 1469-1482. doi: 10.1016/j.cell.2019.11.020 PMID: 31835028 PMCID: PMC7077032
Rush JA & Hisham I (2018) Speculations on the future of psychiatric diagnosis. The Journal of Nervous and Mental Disease 2018 206(6) 481-487. doi: 10.1097/NMD.0000000000000821
Mota NR, Rovaris DL, Kappel DB et al (2015) NCAM1-TTC12-ANKK1-DRD2 gene cluster and the clinical and genetic heterogeneity of adults with ADHD. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics 2015 168(6) 433-444. doi: 10.1002/ajmg.b.32317 PMID: 25989041
Bidwell LC, McGeary JE, Gray JC et al (2015) NCAM1-TTC12-ANKK1-DRD2 variants and smoking motives as intermediate phenotypes for nicotine dependence. Psychopharmacology 2015 232(7) 1177-1186. doi:Â 10.1007/s00213-014-3748-2 PMID: 25273375 PCMID: PMC4361268
NHS Race and Health Observatory (2024) Ethnic inequities in genomics and precision medicine. https://www.nhsrho.org/wp-content/uploads/2024/06/RHO-Genomics-Report-June-2024.pdf
Murphy SE, Capitão LP, Giles SLC et al 2021 The knowns and unknowns of SSRI treatment in young people with depression and anxiety: Efficacy, predictors, and mechanisms of action. The Lancet Psychiatry 8(9) 824-835. doi: 10.1016/S2215-0366(21)00154-1Â
