Unravelling bipolar disorder: insights from the biggest genetic study to date
Bipolar disorder is a leading contributor to the global burden of disease with significant impact on quality of life, functioning and suicide risk (Carvalho et al. 2020; Aghababaie-Babaki et al. 2023). Current treatments involve mood stabilisers, antipsychotics, and antidepressants, combined with psychosocial interventions. However, about one-third of patients relapse within a year (Geddes et al. 2013). In addition, the majority of the underlying genetic determinants of bipolar disorder remain unknown (Lichtenstein et al. 2009).
A genome-wide association study (GWAS) conducted by the Psychiatric Genomics Consortium previously identified 64 risk loci associated with bipolar disorder and highlighted calcium channel antagonists as potential drug targets (Mullins et al. 2021). Additionally, neuroimaging studies have found reduced cortical thickness, lower subcortical volume, and altered white matter integrity in bipolar disorder, along with brain changes linked to medication use (Ching et al. 2022). However, while recent genetic and neuroimaging research has advanced our understanding of pathophysiology, research has been almost exclusively conducted in European ancestry populations, restricting the generalisation of findings.
In a recent study published in Nature by Kevin S. O’Connell et al (2025), the authors have conducted the largest to date multi-ancestry GWAS meta-analysis of bipolar disorder. They examined the impact of bipolar subtypes, ancestry, and patient source on genetic architecture and provide new insights into its neurobiology and potential targets for precision medicine and novel treatments.
Bipolar disorder is a leading contributor to the global burden of disease with significant impact on quality of life, functioning and suicide risk.
Methods
Individuals from 79 cohorts were included in the GWAS meta-analysis for a total of 158,036 cases with bipolar disorder and 2,796,499 controls for comparison purposes. Bipolar disorder cases were pulled from 3 different sources: clinical including semi-structured clinical interviews, community including medical records, registries and questionnaires data, and self-reported surveys. Cases of four different ancestral groups (European, East Asian, African American, and Latino) and each bipolar subtype (BDI vs. BDII) were included.
First, meta-analyses were performed separately for each ascertainment source, bipolar subtype, and ancestral group. Then, a multi-ancestry analysis of all the dataset was conducted. Finally, polygenic risk score analyses and single-cell enrichment studies to explore affected neuronal populations were conducted. Polygenic risk score analyses were performed in 55 European ancestry cohorts including 40,992 cases and 80,215 controls, as well as in one cohort of African American ancestry (347 cases and 669 controls), and three cohorts of East Asian ancestry (4,473 cases and 65,923 controls). Polygenic risk score analysis is a method used to estimate an individual’s genetic predisposition to a particular disease or trait by calculating a weighted sum of their genetic variants across the genome. Credible genes—genes considered likely to play a causal role in the underlying pathophysiological mechanism of a disorder—and potential drug targets were investigated.
The study analysed a massive sample to reveal genetic insights into bipolar disorder.
Results
Differences in source of patients and BD subtypes led to differences in genetic architecture
Bipolar disorder ascertained from clinical sample was more heritable than bipolar disorder ascertained from community samples or self-reported. In addition, genetic correlation was strong between clinical and community samples, and between self-reported and community samples. Bipolar disorder from self-reported samples had the greatest polygenicity and bipolar disorder in clinical samples was the most discoverable. Almost all variants were shared between bipolar disorder from community and clinical samples. Additionally, bipolar disorder type 1 (BDI) was more heritable than bipolar 2 (BDII), although there was a high correlation between both subtypes. However, genetic correlation between BDI and self-reported bipolar disorder was significantly lower than between BDII and self-reported bipolar.
Ancestry-specific GWAS meta-analyses
In the European ancestry population, 261 independent genome-wide significant variants mapping to 221 loci associated with bipolar disorder were identified. In the East Asian ancestry meta-analysis, 2 loci were identified, one of which was novel. In the African American and Latino ancestry meta-analyses, no genome-wide significant loci were identified.
Multi-ancestry meta-analysis
The multi-ancestry meta-analysis identified 337 genome-wide significant variants mapping to 298 loci, including 267 novel loci. Of the 64 previously reported bipolar disorder-associated loci, 31 met genome-wide significance in the current study. The direction of association for all top variants from the previous GWAS was consistent with the current meta-analysis. When the effect of ancestry was considered, one locus was strongly associated in the East Asian ancestry meta-analysis. All other loci had stronger association in the European ancestry meta-analysis.
Genetic correlations with human diseases and traits
Major depressive disorder, post-traumatic stress disorder, attention deficit-hyperactivity disorder, borderline personality disorder and autism spectrum disorder were more strongly associated with the full meta-analysis, BDII subtype and bipolar disorder in the community and self-reported samples, than with BDI subtype and bipolar disorder in clinical cohorts. On the opposite, schizophrenia was more strongly genetically correlated with the bipolar disorder meta-analysis excluding self-reported sample and with BDI subtype and bipolar in clinical samples.
Polygenic association with bipolar disorder
In the European ancestry cohorts, the variance explained by the multi-ancestry GWAS when excluding the self-reported samples was greater than the multi-ancestry GWAS with self-reported samples and the European ancestry GWAS without self-reported samples. Individuals with polygenic risk score (derived from the multi-ancestry GWAS without self-reported data) in the top 20% had a 7.06 increased likelihood of being affected with bipolar disorder compared to individuals in the middle quintile.
Pathway, tissue and cell-type enrichment
Six significantly enriched gene sets relating to synapse and transcription factor activity were identified. Single-cell enrichment analyses indicated involvement of neuronal populations from several brain regions including hippocampal pyramidal neurons and interneurons of the prefrontal cortex and hippocampus. Enrichment of specific dopamine-related and calcium-related biological processes and molecular functions, as well as GABAergic interneuron development, were also highlighted. Interestingly, single-cell enrichment analysis of non-brain mouse tissues identified a significant enrichment in the enteroendocrine cells of the large intestine and delta cells of the pancreas.
Credible bipolar disorder-associated genes and drug target
Using several approaches to map loci to genes, a set of 36 credible genes were identified. Genes SP4, TTC12, and MED24 were strong candidates. Specifically, regulation of SP4 in astrocytes and GABAergic neurons was linked to a specific genetic variant identified in the multi-ancestry GWAS. Eight of the 36 identified credible genes were genes known to be involved in synaptic function. Additionally, sixteen credible genes showed evidence of tractability by small molecule, suggesting their potential as drug targets. Notably, two genes were linked to lithium target.
Researchers uncovered 289 loci, with 267 newly identified, in this large multi-ancestry analysis.
Conclusions
This study provides novel insights into the genetic basis of BD by conducting the largest GWAS of bipolar disorder, including populations of European, East Asian, African American and Latino ancestry, and identifying 289 significant BD-associated loci. Differences in sample source and bipolar subtypes led to differences in genetic architecture. Synapse, interneurons of the prefrontal cortex and hippocampus, and hippocampal pyramidal neurons emerged as particularly relevant. Finally, dopamine-related and calcium-related biological processes were identified as important in exploratory analyses.
The study highlighted key brain cells and regions involved in bipolar disorder.
Strengths and limitations
One of the major strengths of this study is its unprecedented scale, analysing data from over 158,000 individuals with bipolar disorder and nearly 2.8 million controls. This large sample size significantly enhances statistical power, leading to the identification of more than four times the number of loci found in previous studies. Another strength is the study’s multi-ancestry approach, which includes individuals of European, East Asian, African American, and Latino descent. This diversity improves the generalisability of findings and allows for the discovery of ancestry-specific associations. Similarly, including bipolar disorder cases ascertained from different sources facilitates the investigation of source-specific differences in the genetic architecture of bipolar.
Despite its strengths, the study has a few limitations. A European linkage disequilibrium reference panel was used to analyse the multi-ancestry meta-analyses, thus, failing to fully capture the heterogeneity of within ancestry groups. Additionally, the genetic findings are strongest in European ancestry populations, limiting their applicability to underrepresented groups like African and Latino populations. Another limitation is the inclusion of samples with minimal phenotype, especially in non-European samples, which can result in association signals with low specificity. Finally, it is worth commenting on the reliance on self-reported bipolar disorder cases, which show lower heritability estimates and may introduce diagnostic inaccuracies. While the study attempts to address this by stratifying data by ascertainment method, misclassification remains a concern.
Cells in the large intestine and pancreas are potentially implicated in bipolar disorder
Implications for practice
Findings from this study have several implications for practice. First of all, this study’s findings provide a foundation for personalised medicine approaches in bipolar disorder. Polygenic risk scores derived from the study also have potential applications in risk prediction. Although polygenic risk scores are not yet clinically useful as standalone diagnostic tools, they could be integrated with other tools (see, for example, my last blog: Is brain imaging the future for bipolar disorder diagnosis in adolescents?) or risk factors, such as family history and environmental influences, to identify individuals at high risk for bipolar disorder earlier. This early identification may facilitate preventive interventions, such as lifestyle modifications, psychoeducation, or close psychiatric monitoring, and reduce the duration of untreated illness (Buoli et al. 2020) by assisting in earlier diagnosis.
Additionally, authors highlight genetic pathways linked to bipolar disorder, including synaptic function, calcium signalling, and dopamine regulation. These findings reinforce the role of calcium channel blockers as potential therapeutic agents (Dubovsky et al. 2018) and may contribute to drug repurposing strategies. Future research could explore new drugs targeting genes like SP4, which has been implicated in both bipolar disorder and schizophrenia.
The study also demonstrates genetic differences between bipolar subtypes (BDI vs. BDII) and between cases identified through clinical settings versus self-reported data. This highlights the need for more precise diagnostic criteria that account for genetic heterogeneity. Improved classification may help clinicians refine treatment plans, ensuring that bipolar subtypes receive the most appropriate interventions.
Finally, a novel finding of this study is the genetic enrichment observed in enteroendocrine cells of the large intestine, suggesting a potential link between bipolar and gut health. This supports emerging research on the gut-brain axis, raising the possibility that interventions (Campbell et al., 2025) or microbiome-targeted treatments could complement psychiatric care for bipolar disorder.
Genetic differences between bipolar subtypes underline the need for precise classification.
Statement of interests
Emiliana directly reports to one of the authors on this paper and collaborates on a different project with another author on this paper. She was not involved in the production of this study and declares no conflict of interest in relation to this study.
Links
Primary paper
O’Connell, K. S., Koromina, M., van der Veen, T., Boltz, T., David, F. S., Yang, J. M. K., Lin, K. H., Wang, X., Coleman, J. R. I., Mitchell, B. L., McGrouther, C. C., Rangan, A. V., Lind, P. A., Koch, E., Harder, A., Parker, N., Bendl, J., Adorjan, K., Agerbo, E., … Andreassen, O. A. Genomics yields biological and phenotypic insights into bipolar disorder. Nature (2025), doi: https://doi.org/10.1038/s41586-024-08468-9
Other references
Aghababaie-Babaki, P., Malekpour, M. R., Mohammadi, E., Saeedi Moghaddam, S., Rashidi, M. M., Ghanbari, A., Heidari-Foroozan, M., Esfahani, Z., Mohammadi Fateh, S., Hajebi, A., Haghshenas, R., Foroutan Mehr, E., Rezaei, N., Larijani, B., & Farzadfar, F. Global, regional, and national burden and quality of care index (QCI) of bipolar disorder: A systematic analysis of the Global Burden of Disease Study 1990 to 2019. The International journal of social psychiatry (2023), 69(8):1958–1970, doi: https://doi.org/10.1177/00207640231182358
Buoli M, Cesana BM, Fagiolini A, et al. Which factors delay treatment in bipolar disorder? A nationwide study focussed on duration of untreated illness. Early Intervention in Psychiatry (2021), 15:1136–1145. https://doi.org/10.1111/eip.13051
Campbell, I. H., Needham, N., Grossi, H., Kamenska, I., Luz, S., Sheehan, S., … Smith, D. J. A pilot study of a ketogenic diet in bipolar disorder: clinical, metabolic and magnetic resonance spectroscopy findings. BJPsych Open (2025), 11(2), e34, doi: https://doi.org/10.1192/bjo.2024.841
Carvalho, A. F., Flirth, J. & Vieta, E. Bipolar Disorder. The New England Journal of Medicine (2020) 383:58-66, doi: Â https://doi.org/10.1056/NEJMra1906193
Ching, C. R. K., Hibar, D. P., Gurholt, T. P., Nunes, A., Thomopoulos, S. I., Abé, C., Agartz, I., Brouwer, R. M., Cannon, D. M., de Zwarte, S. M. C., Eyler, L. T., Favre, P., Hajek, T., Haukvik, U. K., Houenou, J., Landén, M., Lett, T. A., McDonald, C., Nabulsi, L., Patel, Y., … ENIGMA Bipolar Disorder Working Group. What we learn about bipolar disorder from large-scale neuroimaging: Findings and future directions from the ENIGMA Bipolar Disorder Working Group. Human brain mapping (2022), 43(1):56–82, doi: https://doi.org/10.1002/hbm.25098
Dubovsky, S. L. Applications of calcium channel blockers in psychiatry: pharmacokinetic and pharmacodynamic aspects of treatment of bipolar disorder. Expert Opinion on Drug Metabolism & Toxicology (2018), 15(1):35–47. doi: https://doi.org/10.1080/17425255.2019.1558206
Geddes, J. R. & Miklowitz D. J. Treatment of bipolar disorder. The Lancet (2013) 381 (9878):1672-1682, doi: https://doi.org/10.1016/S0140-6736(13)60857-0
Lichtenstein, P., Yip, B. H., Björk, C., Pawitan, Y., Cannon, T. D., Sullivan, P. F., & Hultman, C. M. Common genetic determinants of schizophrenia and bipolar disorder in Swedish families: a population-based study. The Lancet (2009), 373(9659):234–239, doi: https://doi.org/10.1016/S0140-6736(09)60072-6
Mullins, N., Forstner, A.J., O’Connell, K.S., … & Andreassen, O. A. Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nature Genetics (2021) 53:817–829, doi: https://doi.org/10.1038/s41588-021-00857-4
