The central challenge in psychiatric drug development is causality. Epidemiological studies frequently reveal associations between a protein, a brain-imaging measure, or a lifestyle factor and the risk of a disorder, but associations alone cannot determine whether the factor drives the illness or is simply a downstream consequence of it. This ambiguity has contributed to an unusually high failure rate in late-stage psychiatric trials, where candidate drugs built on correlational evidence often fail to demonstrate benefit. Mendelian randomization addresses the problem by exploiting the random allocation of genetic variants at conception, a process analogous to the randomization in a clinical trial but conducted by nature across entire populations.

Mendelian randomization rests on a simple principle: if a genetic variant that robustly alters the level or activity of a biological molecule—such as a protein circulating in the blood or the expression of a gene in the brain—is also associated with altered risk of a disorder, this supports a causal role for that molecule in the disease process. Because genes are fixed at conception and generally not influenced by the disease itself or by environmental confounders in the same way as measured exposures, the approach can approximate the effect of a lifelong intervention. It is, in effect, a way of asking what would happen if a drug permanently modified a target, before any drug has been developed.

Yet Mendelian randomization has a well-known vulnerability: linkage disequilibrium, the phenomenon whereby neighboring genetic variants are inherited together as blocks. A variant that influences disease risk may sit near, but not within, the region controlling a candidate protein, producing a spurious causal signal. Colocalization analysis was developed to resolve exactly this problem. By examining the fine-grained pattern of genetic association in a genomic region, colocalization tests whether the same variant is driving both the protein-level signal and the disease signal. When the evidence indicates a shared causal variant, confidence that the protein genuinely participates in the disorder rises substantially, and the combination of the two methods has become a gold-standard screen for drug target prioritization.

The study brought together genome-wide association data from very large international consortia, including hundreds of thousands of participants for depression and tens of thousands for schizophrenia and bipolar disorder, alongside proteomic and transcriptomic datasets that map the abundance of thousands of proteins and genes across tissues. By systematically testing genetically predicted levels of each candidate molecule against genetic liability to each disorder, and then confirming overlaps through colocalization, the researchers were able to narrow an enormous search space of possible targets down to a short list backed by converging lines of evidence. Several of the prioritized targets encode proteins with known drugability profiles, meaning that compounds directed against them either already exist for other indications or fall within chemical classes amenable to pharmaceutical development.

Among the most striking implications of the work is the degree of shared biology it reveals across diagnostic boundaries. Depression, anxiety disorders, schizophrenia, and bipolar disorder are clinically distinct categories, but genetic studies have repeatedly shown that they overlap substantially at the level of inherited risk. The analysis reflected this reality, identifying targets whose causal signals appeared in more than one disorder. For drug developers, such pleiotropic targets carry both promise and caution: a single molecule acting on a shared pathway could potentially benefit multiple patient groups, while safety considerations become correspondingly broader, since modulating the target may influence several facets of brain function at once.

The study also underscored the importance of tissue context. Genetic variants that influence protein levels in the blood do not always do so in the brain, and psychiatric symptoms arise from neural circuitry rather than peripheral biochemistry. Where the underlying data permitted, the researchers examined whether the causal signals were consistent with expression in brain regions implicated in mood regulation, cognition, and reward processing. This layer of analysis matters for translation, because a target that appears compelling in plasma proteomics may prove irrelevant to central nervous system function, whereas one whose genetic regulation is demonstrably active in neural tissue represents a far stronger candidate for psychiatric intervention.

For patients and clinicians, the timeframe for impact should be understood realistically. Genetic target prioritization does not produce a treatment; it produces a hypothesis with unusually strong evidential support. The targets identified now require the full pipeline of experimental validation—cellular models, animal studies, medicinal chemistry, and ultimately clinical trials—before any new therapy reaches the clinic. Nevertheless, the value of the approach lies in its ability to redirect investment. History from other therapeutic areas, notably cardiology, shows that drugs developed against genetically validated targets are substantially more likely to succeed in trials than those based on other forms of evidence. Bringing the same discipline to psychiatry could measurably improve one of the least productive areas of modern pharmaceutical research.

The work also illustrates how the scale of open genetic datasets is reshaping biomedical science. The conclusions rest on the cumulative contributions of hundreds of thousands of research participants whose DNA and clinical information were aggregated across consortia worldwide, together with publicly funded resources cataloguing protein and gene regulation. No single laboratory could have assembled statistical power of this magnitude. As these datasets continue to grow, and as proteomic measurements become more comprehensive and more finely mapped, the resolution of target-screening studies of this kind will only improve, potentially extending the framework to rarer conditions, to treatment-response phenotypes, and to the prediction of side effects before trials begin.

Caution remains warranted. Mendelian randomization estimates the effect of lifelong genetic perturbation, whereas drugs act acutely and often on specific tissue compartments; developmental compensation can blunt the relevance of genetic findings; and any individual target signal requires replication in independent datasets before it can be considered settled. The colocalization framework, while powerful, depends on the density of genetic fine-mapping in each region and can be inconclusive where multiple variants contribute. Even so, the study represents a concrete advance in a field that urgently needs one: a short list of molecular targets for common mental disorders, each supported by human genetic evidence of causality, each assessed for shared mechanisms across conditions, and each grounded in data from the populations the resulting medicines would ultimately serve.

Subject of Research: Identification of therapeutic targets for common mental disorders using Mendelian randomization and colocalization

Article Title: Potential therapeutic targets for common mental disorders identified through Mendelian randomization and colocalization

Article References: Xiong, Z., Li, Z., Ji, X., Chen, H., Li, J., Peng, T., Huang, Z., Yang, L., Dong, X., Zhou, W., & Zhang, H. (2026). Potential therapeutic targets for common mental disorders identified through Mendelian randomization and colocalization. Translational Psychiatry. https://doi.org/10.1038/s41398-026-04416-5

Image Credits: AI Generated

DOI: 10.1038/s41398-026-04416-5

Keywords: Mendelian randomization, colocalization, psychiatric genetics, drug targets, depression, schizophrenia, bipolar disorder, anxiety, Translational Psychiatry, drug development, proteomics, genome-wide association studies

Copy number variants, or CNVs, are deletions or duplications of stretches of DNA that can span anywhere from a few hundred bases to millions of bases and can remove, add, or disrupt entire genes. Unlike single-nucleotide variants, which change a single DNA letter, CNVs reshape the genome’s structural landscape, and when they occur in genes critical to brain development they can have outsized effects on neurodevelopmental and psychiatric outcomes. Several recurrent CNVs, such as deletions at the 22q11.2 locus, have been known for years to dramatically elevate schizophrenia risk, but nearly all of that knowledge was built on cohorts of predominantly European descent. Whether the same structural variants, or entirely different ones, contribute to schizophrenia in East Asian populations, which make up a substantial fraction of the world’s population and carry distinct patterns of genomic variation, remained an open and important question.

The new study addressed that question by assembling and analyzing genome-wide data from individuals of East Asian ancestry, comparing the burden of rare copy number variants in people diagnosed with schizophrenia against unaffected controls. The analytic strategy relied on high-quality genotyping arrays and sequencing-based calls that allow researchers to detect deletions and duplications across the genome, followed by careful filtering to remove likely artifacts and annotation of each variant against gene content, known disease loci, and measures of a gene’s intolerance to loss-of-function variation. Burden tests, which ask whether cases collectively carry more large, rare, gene-disrupting CNVs than controls, form the statistical backbone of this kind of work, and the study applied them with the sample sizes needed to detect effects that individual variants alone would be too rare to reveal.

The results confirmed that the fundamental burden signal holds across ancestries. People with schizophrenia in East Asian cohorts carried a significant excess of rare CNVs, particularly those that are large, that remove or duplicate many genes, and that overlap genes previously implicated in neurodevelopmental disorders. This is precisely the pattern observed in European studies, and its replication in an East Asian setting carries real weight: it suggests that the structural-variant contribution to schizophrenia is not an artifact of any one population’s genomic history or ascertainment, but a genuine and broadly shared feature of the disorder’s genetic architecture. For clinicians and researchers in East Asia, it also validates the use of CNV screening in psychiatric care and research contexts far beyond the populations in which those tools were originally developed.

Beyond confirming the overall burden, the analyses pinpointed specific rare copy number variants associated with schizophrenia in East Asian populations, contributing new population-specific resolution to a catalog of risk loci that has been heavily Eurocentric. Some of these signals overlap with CNV loci already known from European studies, reinforcing their status as reproducible schizophrenia risk factors, while the East Asian data add power and detail to their characterization. Because the frequencies of specific structural variants differ across populations, owing to drift, demographic history, and selection, mapping them in East Asian genomes is essential for building risk models and genetic counseling frameworks that actually fit the populations being served.

The most ambitious component of the work, however, was its meta-analysis. By combining East Asian results with those from large European ancestry studies, the investigators boosted statistical power well beyond what either cohort could achieve alone and searched for CNV loci associated with schizophrenia across ancestries. This trans-ancestry approach identified additional risk loci that no single population had the numbers to confirm on its own. The logic is straightforward: if a rare variant’s effect is genuine, pooling evidence across populations with different linkage disequilibrium patterns and different variant spectrums reduces confounding and sharpens the signal. Structural variants, which are often individually very rare and recently arisen, benefit especially from this strategy because their pathogenicity is less dependent on population-specific genetic background than that of common variants.

A striking unifying theme emerged from the annotation of these loci. The genes disrupted by the associated CNVs were significantly enriched for those that are intolerant to loss-of-function variants, meaning that in population sequencing databases, damaging mutations in these genes appear far less often than expected by chance. Genes under strong purifying selection in this way are typically those in which gene dosage matters: losing one copy, or gaining an extra one, perturbs biological systems enough to be selected against. In the brain, dosage-sensitive genes cluster in pathways governing synaptic function, neuronal development, and signaling. The finding that schizophrenia-associated CNVs converge on loss-of-function intolerant genes ties the disorder’s structural-variant risk to the same dosage-sensitive neurodevelopmental biology implicated by de novo mutations in autism, developmental delay, and congenital anomalies, reinforcing a picture of overlapping genetic mechanisms across neuropsychiatric conditions.

The implications run in several directions at once. Scientifically, the study helps close a long-standing gap in psychiatric genetics, demonstrating that rare structural variation is a universal contributor to schizophrenia risk and supplying East Asian-specific loci that will refine global catalogs of disease genes. Methodologically, it shows the value of building genomic resources in understudied populations and then integrating them through meta-analysis rather than extrapolating from European data. Clinically, dosage-sensitive CNV loci identified across ancestries could inform the emerging practice of returning secondary findings to psychiatric patients, since carriers of known pathogenic CNVs may benefit from surveillance for associated medical comorbidities. And for drug discovery, each new risk locus is a pointer toward biology, with dosage-sensitive genes offering mechanistic hypotheses about synaptic and developmental processes that go awry in psychosis.

The study also arrives amid a broader recalibration of how the field thinks about the genetics of schizophrenia. Genome-wide association studies have catalogued hundreds of common variant loci that collectively explain a large share of heritability but individually contribute tiny effects, while rare, high-impact variants such as large CNVs explain a smaller but more mechanistically tractable slice of risk. Rare structural variants, particularly those spanning multiple loss-of-function intolerant genes, are among the strongest known genetic risk factors for the disorder, and the demonstration that this risk architecture replicates in East Asian populations strengthens confidence that findings from these variants will translate broadly. The remaining challenges are considerable: sample sizes for rare variant discovery in non-European populations still lag far behind those in Europe, detection and comparison of CNVs across platforms and ancestries remains technically demanding, and translating locus discovery into biological understanding requires functional work well beyond association statistics.

Still, the trajectory is clear. Schizophrenia genetics has moved from single candidate genes to genome-wide surveys, from one continent to many, and from catalogs of associations to mechanistic principles such as dosage sensitivity and loss-of-function intolerance that bind risk loci into coherent biological stories. By showing that East Asian populations carry the same excess of rare, gene-disrupting copy number variants, and by using trans-ancestry pooling to surface additional risk loci enriched in genes that evolution refuses to let break, this work takes a significant step toward a genetic account of schizophrenia that genuinely fits the global population it affects. It is a reminder that the path to understanding a universal human illness must, by necessity, run through all of humanity’s genomes.

Subject of Research: Rare copy number variants associated with schizophrenia in East Asian populations

Article Title: Contribution of copy number variants to schizophrenia in East Asian populations

Article References: Chen, Y., Feng, Q., Lam, M., Yu, M., Sun, Y., Huai, C., Jana, B., Fu, J., Liao, C., Ye, R., Kim, S., Tubbs, J. D., Shanta, O., Thiruvahindrapuram, B., Jen, Y., Zhao, G., Wang, J., Stanley Global Asia Initiatives, Schwab, S. G., … Huang, H. (2026). Contribution of copy number variants to schizophrenia in East Asian populations. Nature Genetics. https://doi.org/10.1038/s41588-026-02732-6

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02732-6

Keywords: schizophrenia, copy number variants, East Asian populations, genomics, rare variants, meta-analysis, genetic risk loci, loss-of-function intolerance, Nature Genetics, psychiatric genetics, population genetics, neurodevelopmental genes

The question is important because comorbidity changes both the experience of illness and the way healthcare systems respond to it. Depression may alter sleep, appetite, physical activity, stress physiology and adherence to treatment, while chronic physical disease can produce pain, disability, social isolation and uncertainty—factors that can increase the risk of depressive symptoms. These relationships can run in both directions, making simple cause-and-effect explanations unreliable. A person who has depression and diabetes, for example, may share inherited biological risk for both conditions, develop one illness partly as a consequence of the other, receive medications that affect the second condition, or be exposed to social and environmental pressures that influence both. A genetic study of pleiotropy attempts to separate these overlapping explanations rather than treating every association as evidence that one disease directly causes another.

In genetics, pleiotropy describes the influence of one genetic variant, gene or biological pathway on more than one observable characteristic. The concept is not inherently harmful: many genes participate in several physiological systems, and the same molecular machinery may contribute to brain function, immune regulation, energy metabolism or cardiovascular biology. When researchers compare genetic associations for major depressive disorder with those for physical diseases, they can search for shared signals. Such overlap might point to common mechanisms, including inflammation, hormonal regulation, neuronal signalling, mitochondrial activity or the processing of stress. It might also reveal that apparently separate clinical diagnoses are partly different expressions of a shared underlying vulnerability. However, genetic overlap is not the same as a direct causal pathway. A shared variant can affect two diseases independently, or it can influence an intermediate trait that connects them.

Major depressive disorder is particularly challenging to study because it is clinically heterogeneous. Diagnostic criteria group together people with different combinations of symptoms, including persistent sadness, loss of interest or pleasure, changes in sleep and appetite, impaired concentration, fatigue and feelings of worthlessness or guilt. The disorder also varies in age of onset, duration, recurrence, severity and response to treatment. This diversity means that a genetic association detected across thousands of participants may represent only one component of a much broader biological landscape. Physical comorbidities are similarly diverse. Cardiovascular disease, obesity, type 2 diabetes, autoimmune illness and chronic pain have distinct causes, yet each may intersect with depression through overlapping biological and social routes. A study examining pleiotropy across multiple conditions therefore confronts the problem of identifying reproducible shared biology amid substantial variation.

Modern psychiatric genetics generally addresses these questions through large-scale association data. Genome-wide association studies scan millions of DNA positions across the genomes of many participants and compare variant frequencies between people with and without a trait or diagnosis. The resulting associations are usually small in effect, but collectively they can be used to estimate polygenic liability: the combined contribution of many variants to an individual’s statistical risk. Researchers can then compare polygenic patterns for major depressive disorder with those associated with physical diseases. Statistical measures such as genetic correlation quantify whether the same genetic variants tend to influence two traits in the same direction or in opposite directions. More detailed analyses can test whether overlap is concentrated in particular genomic regions, genes, tissues or biological pathways.

Those approaches can generate powerful clues, but they require careful interpretation. Genetic correlation may arise from genuine shared biology, from correlations between study samples, or from differences in ancestry, diagnosis, age, sex, socioeconomic conditions and other sources of bias. Depression-related genetic studies have historically relied heavily on participants of European ancestry, limiting how confidently their results can be generalized to other populations. Diagnostic definitions can also differ between clinical records, self-reported questionnaires and structured research assessments. Physical disease studies may use similarly varied criteria. These issues matter because a statistical signal can appear stronger or weaker depending on how traits are measured and who is included. Genetic findings also describe population-level probabilities; they do not determine whether a particular person will develop depression or any associated physical illness.

The word “dissecting” in the article title suggests an effort to move beyond a single measure of genetic correlation and examine the components of shared risk in greater detail. In principle, such work can distinguish broad pleiotropy—where many variants each contribute modestly to several traits—from more localized overlap involving particular genomic regions. It can also investigate whether the genetic architecture connecting depression with a physical disease is concentrated in biological pathways or tissues relevant to the brain, immune system, endocrine organs or metabolism. Other analyses may ask whether the apparent relationship is driven by a subset of symptoms, by disease severity, or by factors such as smoking, body mass index and sleep. The supplied source does not state which of these analyses the authors performed, so these possibilities should not be presented as reported results of the study.

The clinical significance of this research lies in the possibility of improving how depression and physical illness are recognized and treated together. If robust shared mechanisms can be identified, they might help researchers find therapeutic targets that benefit both mental and physical health. Genetic evidence could also support better risk stratification, although it is not currently a substitute for clinical assessment. A biological pathway associated with both depression and cardiovascular disease, for instance, might motivate studies of anti-inflammatory or metabolic interventions, but only controlled clinical trials can establish whether changing that pathway improves outcomes. Likewise, evidence of shared inherited risk could encourage integrated care, in which mental-health screening is routine for patients with chronic disease and physical-health monitoring is routine for people receiving psychiatric treatment. The goal would be more coordinated care, not the reduction of complex illnesses to a DNA score.

The publication of “Dissecting pleiotropy between major depressive disorder and physical disease comorbidities” places the study within a rapidly expanding effort to understand mental illness as part of whole-body biology. Its subject reflects a shift away from the old division between psychiatric and physical disease, while also highlighting the danger of replacing one oversimplification with another. Depression is not merely a chemical imbalance, and genetic overlap does not erase the roles of life experience, healthcare access, infection, trauma, medication, behaviour or chance. Because the available source contains only the bibliographic record, the article’s specific discoveries, datasets and implications cannot be independently summarized here. What can be said is that the research addresses a central problem in contemporary genetics: determining how shared biological influences contribute to the striking and clinically consequential coexistence of major depressive disorder with physical disease.