The review, led by researchers synthesizing evidence from PubMed, Scopus and EBSCO up to July 2026, screened the literature according to PRISMA 2020 guidelines and ultimately included 31 original human studies: 18 on metformin and 13 on SGLT2 inhibitors. The authors deliberately chose an integrative framework, allowing them to combine pharmacogenetic association studies, pharmacokinetic investigations and clinical response data into a single narrative synthesis. Because the included studies varied so widely in design, populations and outcome measures, the team did not perform a formal meta-analysis or a uniform risk-of-bias assessment, instead weighing evidence by consistency, biological plausibility and sample size.

The scale of the clinical problem is enormous. According to the latest International Diabetes Federation Atlas figures cited in the review, roughly 589 million adults worldwide are living with type 2 diabetes, about one in nine adults, and that number is projected to reach 853 million by 2050. More than 90 percent of all adults with diabetes have the type 2 form, and approximately 252 million people, some 43 percent of the diabetic population, remain undiagnosed. With lifestyle change alone insufficient for most patients, pharmacotherapy is the backbone of treatment, which makes predictable drug response a global public health priority.

Metformin, the biguanide that has anchored first-line diabetes therapy for more than six decades, lowers glucose primarily by suppressing hepatic gluconeogenesis. It downregulates key gluconeogenic enzymes such as phosphoenolpyruvate carboxykinase and glucose-6-phosphatase, dampens glucagon-driven cAMP and protein kinase A signalling, and blocks mitochondrial glycerophosphate dehydrogenase, limiting the conversion of lactate and glycerol into glucose. It also activates AMP-activated protein kinase by inhibiting mitochondrial complex I, which reduces ATP and raises AMP, thereby curbing glucose production and lipogenesis while improving insulin sensitivity and promoting GLUT4-mediated glucose uptake in muscle and fat. Beyond glycemia, the drug has attracted interest for potential roles in cancer, polycystic ovary syndrome, non-alcoholic fatty liver disease and aging, and recent work suggests it may modulate the gut microbiome and stimulate the appetite-related metabolite N-lactoyl-phenylalanine.

Because metformin is not meaningfully metabolized, its fate in the body depends almost entirely on membrane transporters. Organic cation transporters encoded by SLC22A1 (OCT1), SLC22A2 (OCT2) and SLC22A3 (OCT3) mediate hepatic uptake, renal uptake and tissue distribution, while the multidrug and toxin extrusion proteins MATE1 and MATE2, encoded by SLC47A1 and SLC47A2, handle efflux into bile and urine. Polymorphisms in these genes can alter intracellular drug concentrations, systemic exposure and clearance, and thereby shift both glycemic efficacy and tolerability. The review identifies SLC22A1 as the most consistent genetic marker of metformin response: reduced-function OCT1 alleles have been linked across Han Chinese, Danish, Jordanian, Mexican-Mestizo and South Indian cohorts to attenuated HbA1c reductions, lower plasma metformin concentrations and higher rates of gastrointestinal intolerance. Notably, Danish data showed that carriers of the rs72552763 deletion experienced roughly 1.1 percent less HbA1c lowering at six months than non-carriers.

The evidence for the other transporter genes is more modest and less reproducible. SLC22A2 variants, particularly rs316019, appear to influence renal clearance of metformin more than glucose-lowering effect, with some studies showing reduced clearance in carriers and others finding no significant interaction at all. SLC22A3 and SLC47A1 show moderate, population-dependent associations with glycemic response, while SLC47A2 findings are inconsistent across South Indian and Iraqi cohorts. Intriguingly, the ATM gene, which encodes a kinase functionally linked to AMPK signalling rather than drug transport, has been associated with variability in metformin efficacy in both Chinese and large European cohorts, demonstrating that pharmacodynamic pathways, not just pharmacokinetic ones, shape who responds. The review’s authors stress that none of these individual associations yet supports routine genotype-guided metformin prescribing.

SGLT2 inhibitors, by contrast, represent a younger and genetically less charted pharmacogenomic landscape. This drug class, including canagliflozin, dapagliflozin, empagliflozin, ertugliflozin and bexagliflozin, descends conceptually from phlorizin, a compound isolated from apple tree bark whose glucose-blocking properties in the kidney were recognized as early as the 1930s. Modern gliflozins selectively inhibit SGLT2, the transporter responsible for reabsorbing roughly 90 to 97 percent of filtered glucose in the proximal tubule, forcing the excess glucose out in urine through an entirely insulin-independent mechanism. In persistent hyperglycemia, renal SGLT2 expression is upregulated, so blocking it counteracts a maladaptive amplification loop. Although their glucose-lowering effect is modest, large trials have established robust cardiovascular and renoprotective benefits, making them a cornerstone of contemporary care.

Pharmacogenetically, SGLT2 inhibitors are metabolized mainly by glucuronidation through UGT1A9 and UGT2B4. Reduced-function variants such as UGT1A9*3 (p.Met33Thr) increase systemic exposure to canagliflozin and ertugliflozin by about 1.26 to 1.45-fold in pharmacokinetic studies, but these changes have not translated into clinically meaningful differences in efficacy or dose requirements, and pooled analyses of more than 1,600 participants concluded that no dose adjustment is warranted. Variants in the drug target gene SLC5A2, which encodes SGLT2 itself, show associations with glucose homeostasis, HbA1c and diabetes risk in some populations, and one Slovenian study linked the rs9934336 A allele to higher fasting glucose and increased risk of diabetic retinopathy. Emerging but preliminary data suggest SLC5A2 variants may even modulate cardiorenal outcomes, possibly reflecting a form of natural SGLT2 inhibition that limits additional drug benefit. Meanwhile, the PNPLA3 I148M variant, a major determinant of fatty liver disease, appears to modify how much liver fat patients lose on dapagliflozin, with risk-allele carriers responding less to monotherapy but more when the drug is combined with omega-3 carboxylic acids.

The review’s comparative analysis highlights how mechanistically distinct these two drug classes are, and why their pharmacogenetic profiles differ so sharply. Metformin acts on the liver and depends on a well-characterized transporter network, giving it the most mature pharmacogenetic evidence base of any oral glucose-lowering agent. SGLT2 inhibitors act on the kidney, are cleared by hepatic glucuronidation, and carry strong cardiorenal benefits with very low hypoglycemia risk, but their genetic determinants remain emerging and modest in effect. The authors grade SLC22A1 as the most promising biomarker for clinical use, while most other metformin genes and all SGLT2-related markers remain investigational or exploratory. A recurring theme is that drug response is inherently polygenic: no single variant explains enough variance to guide treatment alone, and genetic effects operate within a web of clinical and environmental modifiers including renal function, diet, concomitant medications and the gut microbiome, which is estimated to contribute alongside genetics to roughly a third of the individual variation in metformin response.

Looking forward, the review outlines several priorities for translating pharmacogenomics into practice. Polygenic risk scores tailored to metformin and SGLT2 inhibitors must be developed, but current discovery cohorts are heavily Eurocentric, limiting generalizability to South Asian, Latino and African populations where genetic architecture and drug response may differ. Multi-omics integration, combining genomics with proteomics, metabolomics, metagenomics and epigenomics, is proposed as the path to more accurate prediction, potentially accelerated by artificial intelligence. Implementation science matters too: clinicians have expressed frustration with commercial pharmacogenetic panels that lack actionable diabetes-specific information and are not embedded in electronic medical records. Ultimately, the authors conclude that large prospective, multicenter, ethnically diverse interventional studies, ideally piggybacking on ongoing cardiovascular and renal trials of SGLT2 inhibitors, are essential before gene-guided glycemic therapy can move from promising biology to routine clinical decision-making.

Subject of Research: Pharmacogenetics of metformin and SGLT2 inhibitor response in type 2 diabetes

Article Title: Gene‐Guided Therapy in Type 2 Diabetes Mellitus: An Integrative Review of the Pharmacogenetics of Metformin and Sodium‐Glucose Cotransporter‐2 Inhibitors

Article References: Hossain, I., Haidar, M., Ratul, M. A., Sumona, J. F., Shabnaz, S., Nokib, A. F., & Fatema‐Tuz‐Zohora (2026). Gene‐Guided Therapy in Type 2 Diabetes Mellitus: An Integrative Review of the Pharmacogenetics of Metformin and Sodium‐Glucose Cotransporter‐2 Inhibitors. Endocrinology, Diabetes & Metabolism, 9(6), Article e70343. https://doi.org/10.1002/edm2.70343

Image Credits: AI Generated

DOI: 10.1002/edm2.70343

Keywords: type 2 diabetes, metformin, SGLT2 inhibitors, pharmacogenetics, OCT1, MATE1, UGT1A9, SLC5A2, precision medicine, polygenic risk scores, drug transporters, Health Science Reports

The scale of the evidence base underpinning PTSD genetics has grown rapidly in recent years. Earlier genome-wide association studies of the disorder were often underpowered and produced few reproducible findings, a common problem in psychiatric genomics. The field’s trajectory changed as consortia aggregated samples numbering in the tens and then hundreds of thousands, allowing robust identification of genetic loci. The recent discovery of dozens of loci and a set of genes with strong evidence of causal involvement in PTSD places the disorder on firmer genomic footing than it enjoyed even a decade ago. Against that backdrop, the current study’s approach of testing whether genetic effects differ by trauma type represents a natural next step: once broad genetic signals are established, the question becomes how those signals are expressed across the diverse range of experiences that people actually encounter.

One reason childhood experiences might show stronger gene–trauma interaction is the sheer duration and developmental timing of early adversity. The developing brain undergoes prolonged periods of plasticity, during which neural circuits governing fear learning, emotion regulation, and stress responsiveness are being calibrated by the environment. Chronic adversity during these windows can shape the set points of the hypothalamic–pituitary–adrenal axis, the body’s central stress-response system, and can influence epigenetic marks that regulate gene expression. In this sense, genetic variants that affect how the developing brain responds to an impoverished or hostile environment may have more opportunity to exert influence than variants acting on responses to a discrete adult event, such as a serious accident, which unfolds over a limited time frame.

The distinction the researchers observed between childhood abuse and childhood neglect is also noteworthy. Abuse involves active harm, whereas neglect involves the absence of needed care, and these two forms of adversity may operate through partly different mechanisms. Neglect can be harder to measure and recall, and its effects may be more diffuse, unfolding gradually through the accumulation of unmet needs rather than through discrete frightening events. The finding that neglect showed a more moderate association with PTSD symptoms but a greater likelihood of genetic moderation raises the possibility that individual differences in biological sensitivity to the absence of care are particularly consequential. It also underscores a methodological point: the apparent strength of gene–environment interaction can depend heavily on how reliably an exposure is measured, and retrospective self-reports of early experiences carry well-known limitations.

The specific genes highlighted in the study offer tantalizing, if preliminary, biological clues. FAM120A, which showed its strongest interaction with childhood emotional neglect, has been implicated in processes related to fear and stress processing in the brain. Genes involved in protein clearance, synaptic communication between neurons, and the regulation of gene expression fit broadly with contemporary models of PTSD pathophysiology, which emphasize altered stress-hormone signaling, changes in fear circuitry involving the amygdala and prefrontal cortex, and disruptions in the cellular machinery that maintains neuronal health. The appearance of SGCD, a gene better known for its role in muscle function, among the strong interactors is a reminder that genomic findings often point in unexpected directions, and that the biological roles of many risk genes remain to be worked out. The researchers’ own observation that no single biological process explained the pattern of trauma-specific interactions is an honest acknowledgment of how much remains unknown.

Methodologically, the study’s use of polygenic risk scores alongside gene-level and variant-level analyses reflects a layered strategy that has become standard in the field. Polygenic risk scores aggregate the small effects of many variants across the genome into a single index of inherited risk, and testing whether such scores moderate the link between trauma and symptoms is a direct way of probing gene–environment interaction at scale. Examining individual genes and variants can then help localize where in the genome these differential effects arise. Each level of analysis has strengths and weaknesses: polygenic scores capture broad inherited liability but are opaque to biological interpretation, while single-gene findings are more interpretable but more vulnerable to sampling noise. Convergence across levels, where it occurs, lends confidence to the results.

The samples drawn upon—UK Biobank alongside the GLAD, EDGI UK, and COPING cohorts—illustrate both the power and the constraints of large-scale volunteer-based genetic research in the United Kingdom. These cohorts collectively encompass well over a hundred thousand participants with genetic data and mental-health questionnaires, enabling analyses that would have been impossible in smaller clinical samples. At the same time, volunteer cohorts tend to over-represent participants of European ancestry and individuals who are relatively engaged with health research, which can limit the generalizability of findings. PTSD itself is also unevenly distributed across populations, with elevated rates documented among veterans, refugees, and communities exposed to sustained violence, groups that are often underrepresented in genomic biobanks. Extending this line of research to more diverse populations will be essential for understanding whether gene–trauma interactions hold across different ancestral backgrounds and cultural contexts.

The reliance on retrospective self-report of trauma is another consideration. Memories of childhood adversity are shaped by current mental state, and individuals with PTSD may recall or interpret past events differently from those without the disorder. Prospective longitudinal studies, in which adversity is documented as it occurs, offer a complementary approach and have sometimes yielded different estimates of the relationship between early adversity and later psychopathology. Triangulating across retrospective and prospective designs, and across self-report and official records, will help establish which of the observed interactions are robust and which may be artifacts of measurement.

From a clinical standpoint, the vision articulated by the researchers—using genetic information to identify trauma-exposed individuals who would benefit most from early support—raises both promise and caution. On the promise side, early intervention after trauma is known to improve outcomes, and targeted screening could direct limited mental-health resources toward those at highest risk. On the caution side, polygenic risk scores for psychiatric conditions currently explain only a modest fraction of the overall variation in risk, far too little to support individual-level prediction on their own. Any future screening application would need to combine genetic information with other risk factors, and would require careful attention to privacy, consent, and the potential for genetic determinism or discrimination. The history of psychiatric genetics offers repeated cautionary tales about overpromising clinical translation before the underlying science is mature.

There is also a broader conceptual implication for how trauma-related research is conducted. If different traumas interact differently with genetic risk, then diagnostic categories and research protocols that treat trauma as a unitary exposure may be missing important heterogeneity. This resonates with a wider movement in psychiatry toward more granular characterization of environmental exposures, recognizing that

Subject of Research: Gene–trauma interactions in posttraumatic stress disorder risk

Article Title: Genetic risk for PTSD may vary depending on the type of trauma experienced, data from more than 140,000 individuals suggests

Article References: Genetic risk for PTSD may vary depending on the type of trauma experienced, data from more than 140,000 individuals suggests. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: PTSD, genetics, childhood trauma, gene-environment interaction, polygenic risk scores, psychiatric genomics, UK Biobank, trauma, mental health, European Journal of Psychotraumatology, Genetic, risk

The numbers behind the crisis are sobering. Before the COVID-19 pandemic, more than 116 million Africans, roughly 11 percent of the continent’s population, were already living with depression, anxiety, post-traumatic stress disorder or substance-use disorders. Women bear a disproportionate share of this burden, with 66 million meeting criteria for clinical depression. Substance-use disorders are projected to rise by 40 percent by 2030, and Africa’s suicide rate already exceeds the global average, standing at 11 deaths per 100,000 people compared with a global mean of 9. These figures sit atop a health system stretched to breaking point by socio-economic disparity, chronic under-investment and the historically low priority given to mental health in many national budgets.

Service provision tells an even more alarming story. Estimates based on the World Health Organization’s Mental Health Atlas suggest that specialist treatment coverage in sub-Saharan Africa reaches only 10 to 12.8 percent for psychosis, 1.9 to 2.6 percent for bipolar disorder and 2.7 to 3.4 percent for moderate-to-severe depression, implying treatment gaps of roughly 87 to 98 percent. African governments spend on average just US$0.50 per person per year on mental health, a quarter of the minimum recommended for low- and middle-income countries. The workforce deficit is equally severe: only 1.4 mental health professionals serve every 100,000 inhabitants, against a global average of 9. Stigma and pluralistic explanatory models that attribute symptoms to spiritual or supernatural causes can further delay formal help-seeking, particularly where traditional or faith-based providers are the first point of contact.

The economic consequences compound the human toll. Between 2000 and 2015, disability-adjusted life years attributable to mental disorders grew by 52 percent, outpacing the continent’s 49 percent population growth. In 2015 alone, lost productivity from mental disorders amounted to 17.9 million disability years. Rising life expectancy adds another layer of complexity, as non-communicable diseases such as diabetes, cardiovascular disease and cancer increase in parallel, and the relationships run in both directions. Sustained antipsychotic use drives weight gain and metabolic risk; depression undermines adherence to diabetes medication; cardiovascular risk factors predict cognitive decline; and HIV doubles the prevalence of anxiety. This tangled web of mental illness, infectious disease and chronic conditions multiplies healthcare costs and demands integrated rather than siloed policy responses.

Against this backdrop, the review argues that genetics offers an essential and largely untapped dimension. Under the liability-threshold model, the formal framework used to conceptualise how multiple risk factors combine to cause disease, an individual’s underlying genetic liability interacts with environmental stressors such as conflict, malnutrition or toxic stress, pushing some people beyond the threshold of clinical susceptibility while others demonstrate resilience. Understanding this interplay could allow African health systems to stratify risk and target prevention within a precision public health framework, aligning genetic research with the United Nations Sustainable Development Goals and continental policy agendas such as Agenda 2063 and the Africa Health Strategy.

The scientific case rests on the peculiar architecture of psychiatric genetic risk. Over the past two decades, the Psychiatric Genomics Consortium, a network of more than 1,600 investigators in 66 countries that has produced 755 peer-reviewed papers since 2008, has shown that genetic risk for psychiatric disorders is extraordinarily polygenic. For most disorders, risk is distributed across thousands of genomic loci, with each common variant contributing only a tiny increment. Model-based estimates suggest that common variant liability involves roughly 5,600 variants for ADHD, 9,600 for schizophrenia and 14,500 for depression. These diffuse effects coexist with rare copy-number and sequence variants that affect fewer individuals but can raise carrier odds by tens to more than a hundred-fold, a pattern consistent with purifying selection against variants that impair neurodevelopment and reproductive fitness.

Despite these molecular differences, studies repeatedly point to shared biological themes across disorders: synaptic organisation and plasticity, regulation of gene expression during brain development, chromatin remodelling, ion-channel signalling and, particularly in schizophrenia, complement-related synaptic pruning and glutamatergic signalling. Disorder-enriched patterns are emerging too. Autism studies emphasise rare variants in genes such as CHD8, ARID1B, SCN2A and the SHANK family; schizophrenia shows convergence of common and rare variation in synapse biology and the function of complement C4A; bipolar disorder implicates GABAergic interneurons and ion channels such as CACNA1C and CACNB2; and a recent trans-ancestry genome-wide study of depression points to synaptic and receptor-clustering biology. Yet current discoveries still explain only a fraction of the heritability inferred from twin and family studies, with much of the remaining signal likely hidden in non-coding and repetitive DNA, complex structural variation and very rare alleles.

This is precisely where Africa becomes indispensable. In the 1000 Genomes Project, African ancestry genomes showed the highest median number of autosomal single-nucleotide variants, 4.3 million per genome compared with 3.5 to 3.6 million in non-African groupings, reflecting deeper population histories and less severe founder bottlenecks. Shorter haplotype blocks and lower linkage disequilibrium allow association signals to be resolved more precisely. The African Genome Variation Project, sequencing just 320 sub-Saharan Africans, identified 9.5 million novel variants absent from existing global reference catalogues. By contrast, the field’s heavy reliance on European ancestry cohorts has practical consequences: polygenic risk scores lose accuracy when applied to populations unlike the discovery cohort, with the greatest loss occurring in populations of African descent. Treating African ancestry as a homogenous category risks excluding clinically important variation, since genetic structure in eastern and southern Africa correlates with ethnolinguistic diversity.

Momentum is now building on the continent itself. NeuroGAP-Psychosis has completed genomic and phenotypic collections from 42,953 participants in Ethiopia, Uganda, Kenya and South Africa, creating Africa’s single largest collection of neuropsychiatric research data. The South African Xhosa schizophrenia study of 1,826 individuals identified an enrichment of private damaging variants in mutation-intolerant genes, with effect sizes generally larger than those observed in comparable subsamples of a much larger Swedish cohort. NeuroDev has enrolled nearly 5,000 children with autism, ADHD or intellectual disability, identifying pathogenic or likely pathogenic variants in 17 percent of South African and 38 percent of Kenyan probands in its first exome analysis. The NIMH’s Ancestral Populations Network targets 200,000 participants across more than 25 global sites, while Wellcome-funded DepGen Africa is sequencing 8,000 depression cases and 4,000 controls in Nigeria, Ethiopia and Malawi. In North Africa and Sudan, large families and consanguinity offer routes to rare homozygous variants that population-based studies may miss.

The authors argue that turning these sparks into a durable discovery engine requires investment far beyond sample collection. Africa, home to 18.5 percent of the world’s population, accounts for barely 1.3 percent of global research and development expenditure, 2 percent of scientific publications and less than 1 percent of patents. Only South Africa, Kenya and Egypt approach the African Union’s 1-percent-of-GDP research target, and even these trail the spending of BRICS nations, the European Union and the United States. Sequencing capacity remains uneven, though Africa CDC reports that 34 of 55 African Union countries had next-generation sequencing capability in public health laboratories by 2023, up from 7 in 2019. Cost-effective approaches such as blended genome-exome sequencing, which combines low-pass whole-genome with high-depth exome sequencing, are being implemented across multiple African initiatives, while high-throughput whole-genome sequencing protocols have been developed in Nigeria. The Broadband Commission estimates that up to $109 billion will be needed to achieve affordable broadband coverage by 2030, underscoring how genomic medicine depends on connectivity spanning academic, hospital and laboratory settings.

Ultimately, the review frames the next decade as a test of whether African neuropsychiatric genetics can deliver systems of value rather than merely larger sample sizes. The benchmarks include African-initiated multi-country investigations whose results replicate elsewhere, shared phenotype dictionaries and data-access procedures, sequencing and analytic capacity that endures beyond individual grants, and governance frameworks that command public trust. With more than 60 percent of Africa’s population under 25, the demographic dividend is enormous, and programmes such as H3Africa, H3ABioNet and the GINGER training initiative have already seeded a new generation of investigators. Bibliometric trends show African-led psychiatric genetics research attracting increasing citations and appearing in higher-impact journals. If sustained investment in cohorts, informatics, governance and workforce development follows, the continent could move from being a source of genomic insights to a producer and evaluator of genomically-informed medicine, defining globally relevant models of inclusive mental health research and equitable precision public health.

Subject of Research: The state and future of neuropsychiatric genetics research in Africa

Article Title: Neuropsychiatry genetics in Africa

Article References: Iyegbe, C. O., Chaouch, M., Isewon, I., Majara, L., Ayinde, O., A O’Hare, M., Mufford, M., Ubah, O. A., Larnaout, A., Grobler, A.-L., Ramaboli, M., Nto, N. J., Alagbe, E., Toikumo, S., Elmugadam, F., van der Merwe, C., Gureje, O., Bensaid, M., Abdulaziz, M., … Dahdouh, A. (2026). Neuropsychiatry genetics in Africa. Nature Communications, 17(1), Article 10534. https://doi.org/10.1038/s41467-026-78057-z

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78057-z

Keywords: neuropsychiatric genetics, Africa, genomic diversity, mental health, polygenic risk scores, Psychiatric Genomics Consortium, genome-wide association studies, precision public health, NeuroGAP-Psychosis, health policy, sequencing infrastructure, genomic governance

The research, led by Alexandros Rammos and Evie Stergiakouli at the University of Bristol together with colleagues at Cardiff University and the University of Exeter, drew on two of the United Kingdom’s most valuable population resources. The team analysed genetic and phenotypic data from 2,313 children with cleft lip and/or palate recruited through the Cleft Collective, the world’s largest cleft-specific cohort study, and compared them with 7,913 unaffected children from the Millennium Cohort Study, a nationally representative birth cohort. This scale matters, because questions about subtle genetic correlations demand thousands of participants before any signal can be distinguished from statistical noise.

The investigators deployed a battery of complementary genomic techniques, each probing a different aspect of the relationship between cleft biology and neurodevelopment. First, they used linkage disequilibrium score regression, a method that estimates the overall genetic correlation between two traits by examining whether the same stretches of the genome show association signals in both. If cleft liability and psychiatric conditions shared a common genetic architecture, this technique would detect it across hundreds of thousands of genetic variants simultaneously. Instead, the analysis revealed little evidence of genetic correlations between cleft lip and/or palate and any of the eight traits examined, which spanned cognitive, neurodevelopmental, and psychiatric domains.

Second, the team constructed polygenic risk scores for each child. These scores summarise the cumulative effect of thousands of common variants, each contributing a tiny amount, into a single number that estimates an individual’s inherited liability to a given trait. The researchers calculated scores for ADHD, autism, depression, anxiety, schizophrenia, bipolar disorder, educational attainment, and intelligence, and then asked two separate questions. Did children with clefts carry higher polygenic risk than controls? And within the cleft population, did polygenic scores predict actual developmental and behavioural outcomes measured through validated instruments such as the Strengths and Difficulties Questionnaire, the Ages and Stages Questionnaire, and the Moods and Feelings Questionnaire?

The results were revealing precisely because they diverged. Children with cleft lip and/or palate did not show elevated polygenic risk scores for any of the psychiatric or neurodevelopmental traits compared with population controls, nor lower scores for educational attainment or intelligence. Yet within the cleft group itself, the polygenic scores behaved exactly as they do in the general population: children with higher inherited liability to ADHD, for example, showed more ADHD-type behaviours, and children with higher educational attainment scores performed better developmentally. In other words, the genetic factors that shape neurodevelopment in children with clefts appear to be the same ones that shape it in everyone else, operating independently of cleft status rather than through a shared cleft-neurodevelopment pathway.

Third, the researchers turned to Mendelian randomization, a technique that exploits the random allocation of genetic variants at conception to test whether an exposure plausibly causes an outcome. Because genetic variants are inherited independently of the environmental and social confounders that plague observational research, they can serve as natural proxies. The team used genetic variants associated with cleft liability to ask whether genetic predisposition to clefts causally influences neurodevelopmental outcomes. The analysis provided no robust evidence that it does, further undermining the hypothesis that the cleft itself, through shared genetic roots, drives the increased rates of neurodevelopmental difficulty.

The study also examined the relationship between common and rare genetic variation, an area where prior findings had been much more positive. Previous work has established that rare neurodevelopmental copy number variants, large deletions or duplications of DNA segments that remove or multiply dozens of genes at once, are enriched among children with clefts and are associated with poorer developmental outcomes. If common and rare variants were acting through convergent biological pathways, one might expect children carrying these rare variants to also carry a heavier burden of common risk alleles. The team found no such difference: neurodevelopmental copy number variant carriers did not differ from non-carriers in their polygenic burden, suggesting the two classes of variation contribute independently.

The implications for clinical practice are significant. Because polygenic risk scores predict behavioural and developmental outcomes within the cleft population just as they do in the general population, these scores may retain clinical relevance for children with clefts, potentially helping to identify those who would benefit from early developmental monitoring and support. At the same time, the findings make clear that screening for clefts alone does not identify a population with elevated common variant liability to psychiatric illness. Clinicians and families should instead understand the increased neurodevelopmental risk in this group as arising from other sources, which may include the rare variants already identified, environmental exposures, perinatal complications, the social and surgical burdens associated with cleft care, or factors yet to be discovered.

The negative result also carries a broader lesson for psychiatric genetics. Comorbidities between apparently distinct conditions are often assumed to reflect shared genetic architecture, and in many cases that assumption holds. But this study demonstrates that a well-documented clinical association, between a major congenital malformation and neurodevelopmental difficulty, can exist without any detectable common genetic overlap. The mechanisms linking clefts to neurodevelopmental outcomes may be heterogeneous, involving rare mutations in some children, environmental and medical factors in others, and possibly the cumulative experience of living with a visible craniofacial difference and repeated surgical interventions. Disentangling these threads will require longitudinal data of the kind the Cleft Collective was designed to provide.

The research was funded by the Medical Research Council, with additional support from The Scar Free Foundation, The Underwood Trust, and the Vocational Training Charitable Trust, and it relied on the participation of thousands of families and the United Kingdom’s NHS cleft teams. By rigorously ruling out one of the most plausible genetic explanations for neurodevelopmental comorbidity in cleft lip and/or palate, the study sharpens the field’s focus on the mechanisms that remain. For the roughly one in seven hundred children born with an orofacial cleft, the message is cautiously reassuring: their inherited common genetic risk for psychiatric and cognitive difficulties looks much like anyone else’s, and the developmental challenges some of them face are not written into the same variants that shaped their cleft. Finding what does explain that risk is now the field’s central task.

Subject of Research: Common and rare genetic variant contributions to neurodevelopmental risk in children with orofacial clefts

Article Title: Common variant contributions to neurodevelopmental risk in orofacial clefts

Article References: Rammos, A., Lewis, S. J., Davies, A., Wren, Y., Humphries, K., Sandy, J., Sharp, G. C., Owen, M. J., van den Bree, M. B. M., & Stergiakouli, E. (2026). Common variant contributions to neurodevelopmental risk in orofacial clefts. BMC Medicine. https://doi.org/10.1186/s12916-026-05171-6

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05171-6

Keywords: orofacial clefts, cleft lip and palate, neurodevelopmental disorders, polygenic risk scores, Mendelian randomization, copy number variants, genetic correlation, ADHD, autism spectrum disorder, Cleft Collective, psychiatric genetics, BMC Medicine

The team approached the problem with two complementary technologies drawn from statistical genetics. The first was local genetic correlation analysis, carried out with a method called SUPERGNOVA, which estimates how strongly the genetic influences on two traits align or diverge within specific segments of the genome rather than averaging across all roughly three billion base pairs. The researchers applied this tool to sixteen specific cancers plus a pan-cancer phenotype, computing pairwise local correlations for every combination. The second approach relied on polygenic risk scores, or PRSs, which condense the small effects of hundreds of thousands of common genetic variants into a single number estimating an individual’s inherited liability to a disease. By testing whether a PRS built for one cancer predicted risk of another, the researchers could ask whether shared susceptibility extends from broad genomic architecture down to the level of actual prediction.

The genome-wide results set the stage for everything that followed. When the authors computed standard genome-wide genetic correlations across all cancer pairs, they found twenty significantly correlated pairs, a number that confirms earlier reports of modest pan-cancer sharing but also underscores how much of the genome’s influence on each cancer remains disease-specific. Genetic correlation at this scale is a blunt instrument: it asks whether the constellation of variants affecting one cancer tends, on average, to affect another in the same direction. A weak overall correlation does not mean the two cancers share nothing; it may simply mean that shared effects concentrated in a handful of genomic neighborhoods are diluted by thousands of variants acting independently.

That is precisely what the local analysis revealed. Zooming into genomic regions rather than whole chromosomes, Hu and colleagues identified eighty-two distinct regions where genetic signals were shared across sixty-six different cancer pairs. In other words, the number of cancer pairs showing evidence of shared genetic influence more than tripled once the analysis moved from the genome-wide to the regional scale. These regions are not random stretches of DNA. The team annotated them using genetic correlations with non-cancer traits and known associations catalogued in the GWAS Catalog, building a functional picture of where and why cancer susceptibilities converge.

Among the eighty-two shared regions, five stood out as hubs of what the authors call mutually correlated cancer clusters. In these locations, the local genetic effects on multiple cancers were correlated with one another in a structured web, rather than in isolated pairs. When the researchers examined what these hub regions do biologically, enrichment appeared in functional domains connected to inflammatory processes. The finding fits a broader theme in cancer genetics: chronic inflammation is a recognized driver of tumor initiation and progression in many tissues, and inherited variation in immune and inflammatory regulation could plausibly tilt risk for several cancers at once. A single regulatory variant that dampens or amplifies inflammatory signaling, for example, might leave detectable fingerprints in the genetic architecture of lung, colorectal, and other inflammation-associated malignancies simultaneously.

The polygenic risk score analysis added an independent line of evidence. Testing PRSs across cancer pairs, the researchers identified five pairs of cancers with statistically significant cross-cancer associations, meaning that a genetic risk score optimized for one cancer carried predictive information for the other. This is a meaningful step beyond correlation of summary statistics, because PRS association tests whether shared genetic liability is strong enough to translate into measurable prediction in independent data. Still, five significant pairs out of the many tested reinforces the study’s headline conclusion: cross-cancer prediction from genetics exists, but it is the exception rather than the rule, and it demands careful dissection before it can be interpreted biologically.

That dissection came from the study’s most technically inventive component, the decomposition of genome-wide PRSs into functionally informed components. Rather than treating a PRS as an indivisible number, the authors split it into pleiotropy-informed parts, which emphasize variants already known to influence multiple traits, and pathway-specific parts, which weight variants according to membership in defined biological pathways. When the significant cross-cancer PRS associations were decomposed this way, a clear pattern emerged: the shared predictive signal was concentrated in specific pleiotropy groups and in immune-related pathways, not spread diffusely across the genome. The cross-cancer associations, in other words, are not a generic reflection of genetic overlap between complex traits; they trace back to identifiable biological machinery.

The immune pathway enrichment ties the two analytical strands of the study together elegantly. The regional analysis pointed to inflammatory functional domains in the hub regions of shared local genetic correlation, and the PRS decomposition pointed to immune-related pathways as carriers of cross-cancer predictive signal. Two independent methods, operating at different scales of genomic resolution, converged on the same biological story. This kind of methodological triangulation matters in a field where pleiotropy, the phenomenon of one variant influencing many traits, can arise from genuine shared biology or from statistical artifacts such as sample overlap and population stratification. Agreement between local correlation and PRS-based prediction makes a genuine biological basis for the shared risk considerably more convincing.

The implications reach into both research practice and, eventually, clinical risk assessment. For researchers, the study supplies a framework: rather than asking only whether two cancers are genetically correlated overall, investigators can map where in the genome the correlation lives, which functional categories those regions implicate, and whether the shared signal is strong enough to survive PRS-based prediction testing. For clinicians and genetic counselors, the work tempers expectations while pointing to future opportunities. Current polygenic risk scores are largely cancer-specific, and the modest number of significant cross-cancer pairs suggests that a single pan-cancer score is not on the immediate horizon. But for the specific pairs identified, particularly those anchored in immune pathways, multi-cancer risk models that borrow strength across related malignancies could eventually sharpen risk stratification, especially as PRS methods improve and datasets grow more diverse.

The study also carries a cautionary note about scale. Findings that vanish at the genome-wide level but appear at the regional level illustrate a general principle in complex-trait genetics: aggregation can obscure as much as it reveals. Eighty-two shared regions hiding inside twenty significant genome-wide correlations is a quantitative demonstration that the architecture of shared disease risk is heterogeneous, concentrated in pockets of the genome where pleiotropic variants cluster. As biobank-scale datasets expand and methods like SUPERGNOVA and PRS decomposition mature, the expectation is that more of these hidden threads connecting cancers will come into view, each one a potential clue to mechanisms that tumor types as different as melanoma and leukemia might, unexpectedly, share. For now, Hu and colleagues have shown that the genetic conversation between cancers is quieter than a pan-cancer model would hope, but far more organized than a genome-wide average would suggest.

Subject of Research: Shared polygenic genetic risk across multiple cancer types

Article Title: Identifying shared polygenic risk across cancers

Article References: Hu, J., Muheyati, M., Xu, L., DeWan, A., & Zhao, H. (2026). Identifying shared polygenic risk across cancers. PLOS Genetics, 22(9), e1012308. https://doi.org/10.1371/journal.pgen.1012308

Image Credits: AI Generated

DOI: 10.1371/journal.pgen.1012308

Keywords: polygenic risk scores, genetic correlation, cancer genetics, pleiotropy, SUPERGNOVA, immune pathways, inflammation, GWAS, cross-cancer susceptibility, genomic regions, PLOS Genetics, risk prediction

Polygenic risk scores, commonly abbreviated as PRS, summarize the combined influence of thousands or millions of genetic variants on a person’s likelihood of developing a particular condition. Constructing them is a formidable statistical challenge. Modern biobanks measure hundreds of thousands of individuals at millions of genetic positions, and the number of potential predictors vastly exceeds the number of study participants. In this high-dimensional regime, ordinary regression breaks down, and researchers must rely on penalized regression methods that shrink or eliminate weak signals. The most famous of these, the Lasso introduced by Tibshirani in 1996, adds a penalty proportional to the sum of absolute coefficient values, driving the coefficients of uninformative variants all the way to zero and thereby selecting a sparse subset of predictors.

Despite its elegance, the Lasso faces difficulties when the number of variables dwarfs the number of observations, as it does in genomic data. Feature selection can be unstable: small perturbations in the data may cause the method to pick different variants, and correlated variants can compete with one another in ways that obscure the true signal. UniLasso addresses these problems with a two-stage strategy. In the first stage, the method performs simple univariate regressions, examining each genetic variant one at a time for its individual association with the trait of interest. The signs and magnitudes of these univariate coefficients are then used to guide the second stage, a multivariate penalized regression fitted on the full set of variants simultaneously. Variants with strong univariate evidence receive favorable treatment in the penalty, while those with weak or contradictory evidence are discouraged from entering the model.

The intuition behind this design is that univariate screening provides a stable, data-driven prior about which variants matter. By anchoring the multivariate fit to this prior, uniLasso stabilizes feature selection across datasets and reduces the variance of the resulting coefficient estimates. The authors emphasize that the univariate stage does not replace the multivariate fit; it merely informs it. The final model is still learned from individual-level data in the target population, which means the method can adapt to the specific ancestry, cohort composition, and phenotype definitions of the dataset at hand. This two-stage architecture also confers a computational advantage: the initial univariate screen is embarrassingly parallel and can be distributed across many processors, making the approach feasible for datasets with more than one million genetic variants measured on hundreds of thousands of people.

To test the method at true biobank scale, the team applied uniLasso to the UK Biobank, a population-based repository containing genetic, health, and lifestyle data from roughly half a million participants across the United Kingdom. The researchers adapted the algorithm to handle the sheer dimensionality of the resource, where more than a million variants must be evaluated simultaneously. They also introduced an extension called uniLasso ES, short for external scores, which incorporates summary statistics from pre-existing genome-wide association studies. These external signals, drawn from large meta-analyses that may involve millions of additional participants, guide the regression toward variants with prior evidence of association by informing penalty weights and imposing sign constraints that keep coefficient directions consistent with earlier findings.

The uniLasso ES framework is particularly significant because it bridges two traditionally separate paradigms in polygenic risk prediction. Summary-statistic methods, such as PRS-CS and lassosum2, rely exclusively on aggregated association results and linkage disequilibrium reference panels, avoiding the privacy and logistical hurdles of individual-level data but sacrificing some flexibility. Individual-level methods, by contrast, can model the full joint structure of the data but require access to raw genotypes. UniLasso ES occupies a middle ground: external summary statistics act as a soft guide, shaping the penalty landscape, while the definitive fitting happens on individual-level target data. The external evidence informs the model without dictating it, allowing the method to correct for differences between the discovery population and the target population.

The empirical results reported in the study are striking. Across a range of traits and disease outcomes in the UK Biobank, uniLasso attained predictive performance comparable to the standard Lasso while selecting substantially fewer variants. Sparser models offer tangible benefits beyond aesthetics. A risk score built from a few thousand variants rather than hundreds of thousands costs less to genotype in a clinical setting, reduces the storage and computational burden of scoring large patient cohorts, and, crucially, is easier to interrogate biologically. When a compact set of variants drives a prediction, researchers can more readily trace which genes and pathways contribute to the score, potentially generating new hypotheses about disease mechanisms.

Interpretability has been a persistent criticism of polygenic risk scores. Critics note that scores built from millions of tiny effects are statistical black boxes whose individual components rarely correspond to established biology. By producing models that are sparse by construction and whose selected variants carry coherent univariate evidence, uniLasso nudges the field toward scores that a geneticist can actually read. The method’s reliance on univariate coefficient signs also guards against a known pathology of penalized regression in the presence of correlated predictors, where a variant may enter the model with a counterintuitive sign simply because it is absorbing the effect of a nearby variant. Sign constraints informed by univariate evidence keep the fitted coefficients aligned with the marginal signal.

Benchmarking against competing approaches reinforced the method’s promise. The authors report that both uniLasso and uniLasso ES remained competitive with PRS-CS and lassosum2, two widely used PRS estimation methods that represent the current state of the art. Achieving parity with these established tools while producing markedly sparser models positions uniLasso as a practical alternative for research groups that need both accuracy and parsimony. The scalability of the implementation matters as much as its statistical properties; as biobanks grow toward cohorts of millions and sequencing expands the variant catalogue into the billions, methods that cannot be distributed efficiently will simply fall out of use.

The study arrives at a moment when polygenic risk scores are edging toward clinical application, with health systems beginning to explore their use in screening programs for conditions such as cardiovascular disease, diabetes, and several cancers. The promise of personalized prevention depends on scores that are accurate, portable across ancestries, and affordable to deploy. A method that delivers competitive accuracy with a fraction of the variants addresses all three concerns at once. The work also exemplifies a broader trend in statistics: the creative combination of classical ideas, in this case univariate screening and penalized regression, to meet the demands of modern data scale. As the authors and their collaborators continue to refine the framework, uniLasso may well become a standard tool in the geneticist’s arsenal, helping to convert the torrent of biobank data into models that clinicians and patients can understand and act upon.

Subject of Research: Scalable sparse regression methods for polygenic risk score computation in biobank-scale genomic data

Article Title: Univariate-guided sparse regression for Biobank-scale high-dimensional omics data

Article References: Richland, J., Kiiskinen, T., Wang, W., Lu, W. S., Narasimhan, B., Hastie, T., Rivas, M., & Tibshirani, R. (2026). Univariate-guided sparse regression for Biobank-scale high-dimensional omics data. PLOS Genetics, 22(9), e1012314. https://doi.org/10.1371/journal.pgen.1012314

Image Credits: AI Generated

DOI: 10.1371/journal.pgen.1012314

Keywords: polygenic risk scores, uniLasso, UK Biobank, sparse regression, Lasso, genome-wide association studies, high-dimensional statistics, summary statistics, predictive modeling, genomics, feature selection, PLOS Genetics

The research, led by Yang Liu and Dewen Dong of the Health Sciences Institute at China Medical University, together with colleagues including John R. Speakman and senior author Deliang Wen, drew on the UK Biobank, one of the largest biomedical databases in the world. The team focused on 13,931 participants who met the study’s definition of pre-clinical obesity: they had abnormal adiposity, established through body mass index and related anthropometric measures, but had not yet been diagnosed with any obesity-related cardiometabolic disease at baseline. This population is precisely where early prediction matters most, because it represents a window of opportunity in which intervention could still prevent disease rather than merely manage it.

The biological raw material came from the UK Biobank Pharma Proteomics Project, which used affinity-based assays to quantify 2,919 proteins circulating in each participant’s plasma. Proteins are the workhorses of physiology, and their concentrations in blood reflect the real-time state of tissues and organs in ways that genetic data cannot. A person’s genome is essentially fixed at conception, but the proteome is dynamic, shifting with inflammation, insulin resistance, liver fat, vascular stress, and countless other processes that precede overt disease. That dynamism is exactly what makes plasma proteins attractive candidates for early warning systems.

The first analytical pass was deliberately simple: the researchers tested each protein individually for associations with future diagnoses of type 2 diabetes, hypertension, and hyperlipidemia using Cox proportional hazards models, the standard statistical framework for time-to-event data in epidemiology. To guard against false positives across such a massive number of tests, they applied Bonferroni correction, one of the most stringent thresholds in statistics. Even so, the screen yielded 973 significant protein-disease associations involving 653 unique proteins. The sheer volume of signal confirmed what many smaller studies had hinted at: the blood proteome carries a deep reservoir of information about cardiometabolic futures.

Individual proteins, however, are noisy predictors on their own. The team’s real innovation was to combine them into composite proteomic risk scores, which they called ProteinScores, using elastic-net Cox regression. Elastic net is a machine learning technique that performs variable selection and regularization simultaneously, shrinking the coefficients of less informative proteins toward zero while retaining a parsimonious set of features. This approach tames the multicollinearity that plagues proteomic data, where hundreds of proteins travel together in correlated clusters reflecting shared biological pathways. The researchers built separate ProteinScores for three prediction horizons, five, ten, and fifteen years, recognizing that the molecular signatures of disease emerging within five years may differ from those portending trouble a decade or more out.

Evaluating the scores required rigorous benchmarking against increasingly demanding comparison models. The ProteinScores clearly improved discrimination beyond simple age- and sex-adjusted models and beyond baseline covariate models incorporating standard clinical variables across all prediction horizons. The tougher test came from extended clinical models that already included disease-specific indicators such as glycated hemoglobin, blood pressure, and lipid measurements. Against these, the incremental gains were more outcome- and horizon-dependent, with statistically significant improvements in the area under the receiver operating characteristic curve, or AUC, appearing mainly for selected longer-term predictions. That pattern makes physiological sense: clinical indicators capture the disease process once it is already underway, whereas proteins can reveal subclinical perturbations that precede measurable changes in glucose or blood pressure by years.

Discrimination alone does not make a risk score clinically useful; a model must also be well calibrated, meaning its predicted probabilities match observed event rates, and it must meaningfully reclassify patients into more accurate risk categories. The team assessed both. Calibration analyses and category-based net reclassification improvement, a metric quantifying how many people are moved into correct risk strata, showed that the ProteinScores provided additional risk-stratification information in several disease-horizon combinations, though the magnitude of benefit varied. The researchers also pitted their proteomic scores against two rival omics platforms available in the same cohort: polygenic risk scores from UK Biobank fields and metabolomic scores built from circulating metabolites. In this framework the ProteinScores showed favorable performance overall, but the relative ranking shifted by outcome and horizon, underscoring that no single molecular layer holds a monopoly on predictive power.

Beyond prediction, the study delivered biological insight. Pathway enrichment analyses of the proteins selected into the scores implicated lipid transport, lipoprotein remodeling, extracellular matrix organization, peptide hormone processing, and immune-inflammatory signaling. These are not random associations; they map onto the canonical axes of cardiometabolic disease biology. Lipid transport and lipoprotein remodeling connect directly to atherosclerosis and dyslipidemia, extracellular matrix organization to the fibrotic remodeling of adipose tissue that accompanies metabolic dysfunction, peptide hormone processing to the endocrine signaling disrupted in insulin resistance, and immune-inflammatory pathways to the chronic low-grade inflammation that characterizes unhealthy obesity. In effect, the risk scores distill the molecular conversation between fat tissue, liver, vasculature, and immune system into a single number.

The clinical implications are considerable. Current guidelines increasingly recognize that obesity is heterogeneous, distinguishing metabolically healthy from metabolically unhealthy obesity, but these categories are crude snapshots rather than forecasts. A proteomic risk score could transform pre-clinical obesity from a vague state of heightened risk into a stratified condition where patients know, with quantified probability, what their five-, ten-, and fifteen-year futures hold. High-risk individuals could be prioritized for intensive lifestyle intervention, pharmacotherapy such as incretin-based drugs, or heightened surveillance, while lower-risk individuals might be spared unnecessary treatment and anxiety. Because the scores rely on plasma proteins measurable with existing affinity-based platforms, the path from research to clinical deployment is shorter than for many emerging omics technologies, though cost, standardization, and prospective validation in diverse populations remain hurdles.

The study’s limitations are those inherent to its design. It was conducted within the UK Biobank, a cohort that skews toward middle-aged British participants and may not generalize to other ancestries or health systems, and the authors note that the magnitude of improvement over extended clinical models varied across outcomes and horizons rather than being uniformly large. The scores were also developed and evaluated within the same data framework, so external validation in independent cohorts will be essential before routine clinical use. Nevertheless, the work represents a substantive step toward precision prevention in obesity medicine. Rather than waiting for diabetes or hypertension to declare itself, physicians may one day read the proteomic tea leaves in a single blood draw, identifying which of their patients with obesity are quietly heading toward disease and intervening while prevention is still possible. In a world where more than a billion people live with obesity, that kind of foresight could reshape public health on a global scale.

Subject of Research: Proteomic risk scores for early prediction of cardiometabolic diseases in pre-clinical obesity

Article Title: Proteomic risk scores facilitate early prediction of cardiometabolic diseases in pre-clinical obesity

Article References: Liu, Y., Dong, D., Cheskin, L. J., Jiang, L., Cao, G., Speakman, J. R., & Wen, D. (2026). Proteomic risk scores facilitate early prediction of cardiometabolic diseases in pre-clinical obesity. Genome Medicine. https://doi.org/10.1186/s13073-026-01780-1

Image Credits: AI Generated

DOI: 10.1186/s13073-026-01780-1

Keywords: proteomics, pre-clinical obesity, type 2 diabetes, hypertension, hyperlipidemia, UK Biobank, risk prediction, elastic-net regression, polygenic risk scores, metabolomics, precision prevention, cardiometabolic disease

The scale of the initiative is striking. The collaborative research center, funded by the German Research Foundation under the designation SFB/TRR 379, will recruit 900 adults and 200 children and adolescents across three study sites in Frankfurt, Mannheim/Heidelberg, and Aachen, along with a satellite cohort of 150 forensic patients. A matched control group of 300 adults and 100 youth will provide comparison data. Participants between the ages of 10 and 40 will be followed annually for at least 3.5 years, with a planned total duration of 12 years, allowing researchers to track how aggressive behavior emerges, stabilizes, or fades across critical developmental transitions. The study is registered in the German Clinical Trials Register under DRKS00035731.

What makes the approach genuinely novel is its transdiagnostic framing. Rather than treating aggression as a feature of individual disorders, the investigators anchor their measurements in the National Institute of Mental Health’s Research Domain Criteria framework, specifically the negative valence system, which encompasses acute threat and frustrative non-reward, and the cognitive system, which covers cognitive control and self-regulation. The scientific consensus, the authors argue, is that aggressive behavior is largely determined by the processing of threat, frustration, and cognitive control. When sensitivity to threat-related cues is heightened, or neutral cues are misread as threatening, the acute threat system may over-activate and elicit anger or fear; whether that escalates into aggression depends on whether prefrontal control regions can counteract the limbic surge.

Functional neuroimaging research has lent substantial support to this imbalance model. In impulsive or reactive aggression, studies suggest attenuated top-down control from the prefrontal cortex paired with excessive bottom-up activity in limbic and reward regions. Abnormalities in fronto-limbic networks have been documented in borderline personality disorder and conduct disorder, while impairments in emotion processing and regulation networks contribute to heightened aggressiveness in ADHD, autism spectrum disorder, and disruptive behavior disorders. Children and adolescents with conduct disorder show altered fronto-limbic regulation, a pattern also observed in individuals with high trait aggression. The new study will test whether these system-level dysfunctions cut across diagnostic categories, forming shared mechanisms rather than disorder-specific quirks.

Genetics features prominently in the design. The heritability of aggressiveness has been estimated at 50 to 60 percent, rising even higher in individuals with severe aggressive behavior and high levels of callous-unemotional traits. Traditional candidate gene studies focusing on serotonergic and dopaminergic systems, stress hormones, and oxytocin have fallen short, because roughly half of the variance in aggressive behavior appears to reflect complex polygenic influences spanning thousands of polymorphisms. The team will therefore construct an aggression polygenic risk score from genome-wide data and examine its genetic correlation with psychiatric disorder risk scores from large consortia. A genome-wide analysis of more than 700 violent offenders has already suggested complex interactions with childhood stress, trauma, gender, and hormones, and a recent synthesis highlighted RBFOX1, an RNA-binding protein affecting the splicing and stability of thousands of gene transcripts, as a replicated candidate for aggressive behavior.

Hormones form another pillar. Testosterone concentrations are elevated in both female and male patients with borderline personality disorder, and the hormone is thought to modulate aggression by influencing key brain regions for social threat processing, including the amygdala, hypothalamus, and periaqueductal gray. The protocol calls for annual saliva sampling of testosterone, estradiol, progesterone, cortisol, and oxytocin, with additional challenge-based measurements taken before and after participants view violent film sequences inside the MRI scanner. Notably, the role of female-specific sex hormones in aggression has barely been studied, a gap the researchers intend to close. Reduced estrogen has been found in females with conduct disorder, and cortisol awakening responses were positively related to anger and aggressiveness in female patients with borderline personality disorder.

Environmental adversity receives equal weight. Adverse childhood experiences, including abuse and neglect as well as early exposure to harsh urban environments, are well-established risk factors for aggressive behavior and psychopathology, with the type and timing of adversity shaping later risk in distinct ways. The study will assess childhood trauma retrospectively and capture daily life prospectively through ecological momentary assessment, a smartphone-based system run on the open-source JTrack platform developed at Forschungszentrum Jülich. Participants will answer prompts twice a week on a randomized sparse schedule, and in intensive two-week bursts of five prompts per day following each MRI session, recording provocations, anger, stress, substance use, and social context as they unfold in real time. Paper diaries administered every four days will log conflicts and aggressive outbursts continuously throughout the funding period.

The neuroimaging protocol is equally comprehensive. Every participant will undergo 3-Tesla MRI including T1- and T2-weighted anatomical scans, multidirectional diffusion tensor imaging, resting-state functional MRI, and a naturalistic aggression-perception task in which they watch validated violent movie clips drawn from films such as Saving Private Ryan and Leon: The Professional, with age-appropriate clips from a German children’s television series used for participants under 16. Non-violent clips matched for duration and action intensity serve as controls. Beyond the brain, the study will even profile the gut microbiome, collecting annual stool samples for metagenomic sequencing alongside dietary questionnaires, reflecting growing interest in the microbiome’s connection to behavior.

The central hypothesis is bold: that combining questionnaire-based characterization of the negative valence and cognitive systems with genetic, hormonal, and brain measures will reveal up to six distinct clusters, or biosignatures, of aggression that cut across diagnostic categories. Power calculations indicate the planned sample of 1,100 participants is more than adequate, with simulation studies showing that even 160 individuals can suffice to identify four clusters when effect sizes are moderate to large. The researchers will use k-means, hierarchical, and Gaussian mixture clustering, validated through hold-out methods, and later apply machine-learning frameworks to predict individual trajectories and clinical outcomes, prioritizing features with effect sizes above 0.5 to maximize predictive accuracy.

The stakes extend well beyond taxonomy. Aggressive behavior in patients worsens treatment outcomes and carries heavy social, economic, and personal costs, yet current pharmaceutical and psychotherapeutic interventions remain insufficient, partly because aggression itself is rarely the explicit target of treatment. There is also a persistent gender bias: males show higher rates of direct physical aggression, but women predominantly display verbal and indirect aggression that is often overlooked, and the relative risk of offending can actually be higher in female patients compared to female controls. By following participants across adolescence and adulthood, tracking sex-specific developmental trajectories, and measuring everything from polygenic risk to momentary frustration, the consortium hopes to finally predict who is at risk, who is resilient, and precisely which neurobiological levers might one day be pulled to intervene before aggression becomes entrenched.

Subject of Research: Transdiagnostic neurobiological mechanisms of aggression in mental disorders across the lifespan

Article Title: Neuropsychobiological transdiagnostic risk mechanisms for aggression in mental disorders across the life span

Article References: Habel, U., Reif, A., Herpertz, S., Konrad, K., Freitag, C., Banaschewski, T., Frodl, T., Barnes-Scheufler, C., Berg, L., Bertsch, K., Chechko, N., Chiocchetti, A. G., Ecker, C., Ende, G., Fritsch, A., Gramegna, C., Hanke, M., Holz, N., Hüpen, P., … Zweerings, J. (2026). Neuropsychobiological transdiagnostic risk mechanisms for aggression in mental disorders across the life span. BMC Psychiatry, 26(1), Article 750. https://doi.org/10.1186/s12888-026-08359-5

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08359-5

Keywords: aggression, mental disorders, transdiagnostic research, RDoC, neuroimaging, polygenic risk scores, testosterone, childhood adversity, borderline personality disorder, longitudinal study, forensic psychiatry, gut microbiome

The research team drew on 245,388 whole-genome sequences from the All of Us Research Program, one of the largest and most ancestrally diverse biomedical data resources ever assembled in the United States, and combined them with data from the UK Biobank, the dominant resource in European-ancestry genomics. From these combined resources, the investigators developed multiancestry polygenic risk scores for 32 traits and diseases, ranging from quantitative blood biomarkers to common chronic conditions. Their central question was deceptively simple: when building a risk score for a given population, is it better to use the largest possible pooled dataset, or to match the training data to the ancestry of the people being predicted?

The answer, the study shows, depends on context. For many traits, the sheer diversity of the All of Us cohort improved prediction accuracy, particularly for participants from populations that have historically been under-represented in genetic research, including people of African, Hispanic and Indigenous American ancestry. This finding matters because the performance gap between ancestry groups has been a major obstacle to equitable clinical deployment of polygenic scores. When the training data include more individuals who share ancestry segments with the target population, the score captures more of the variants and effect sizes relevant to that population, and prediction improves.

But the study also delivered a cautionary result. Maximizing sample size by meta-analyzing the All of Us and UK Biobank data was not universally optimal. For traits with lower polygenicity, meaning traits influenced by a smaller number of variants with larger effects, training exclusively on All of Us data performed best for participants of African ancestry. The explanation lies in ancestry-enriched genetic effects: variants or effect sizes that are specific to, or much more common in, particular ancestry groups. When such effects are averaged together with data from a very different population, the pooled score can dilute the very signals that matter most for the target group. In other words, more data is not always better data if the additional data come from a population whose genetic architecture differs meaningfully from the people being predicted.

To quantify these effects, the researchers examined how individual-level prediction accuracy decayed as a function of ancestry divergence between the discovery genome-wide association study and the target individual. Consistent with theoretical expectations and prior work, accuracy declined roughly linearly with increasing genetic distance from the discovery population. However, this decay was substantially attenuated when the training data themselves were multiancestry. A score trained across multiple ancestry groups retained more of its predictive power as individuals became more distant from any single discovery population, suggesting that multiancestry training acts as a form of insurance against the portability problem that has plagued single-ancestry scores.

Methodologically, the study was rigorous and comprehensive. The team used REGENIE for genome-wide association analyses, METAL for meta-analysis, and the Bayesian shrinkage frameworks PRS-CS and PRS-CSx for score construction, the latter designed specifically for cross-population prediction. They estimated SNP-based heritability and polygenicity with SBayesS and GCTB, quantified heritability with LD Score Regression, and assessed cross-ancestry genetic correlations with Popcorn. This combination allowed them to connect observed differences in score performance to underlying properties of trait architecture, such as heritability, polygenicity and the degree to which effect sizes are shared or divergent across ancestries.

One of the most striking illustrations of ancestry-enriched effects came from blood panel traits. For several hematological measures, multiancestry meta-analysis scores showed improved accuracy at the individual level in African ancestry participants, driven by variants that are far more common in African-ancestry populations than in European-ancestry populations. A classic example in human genetics is the regulatory variant in the Duffy antigen receptor for chemokines gene, which strongly influences neutrophil counts and is nearly fixed in many African-ancestry populations but rare elsewhere. Variants of this kind are invisible to predominantly European discovery studies, yet they can carry substantial predictive weight. Including diverse populations in discovery is therefore not merely a matter of fairness; it uncovers biology that would otherwise be missed entirely.

The findings arrive at a moment when polygenic risk scores are beginning to move into clinical settings, with several chronic disease scores already being evaluated for implementation in diverse US populations. Earlier work, including a widely cited 2019 analysis by Martin and colleagues, warned that clinical use of current scores could exacerbate health disparities because their accuracy is so uneven across ancestry groups. The new study provides a practical roadmap for mitigating that risk. It suggests that institutions building clinical scores should evaluate multiple training strategies per trait, considering the trait’s genetic architecture, the ancestry composition of the intended population, and the availability of ancestrally matched discovery data, rather than defaulting to the largest available meta-analysis.

The study also underscores the strategic value of programs like All of Us, which was designed from the outset to reflect the diversity of the United States, with more than half of its participants coming from under-represented racial and ethnic backgrounds. The results demonstrate concretely that this design choice pays scientific dividends: the diversity of the cohort is not just an ethical feature but a source of predictive power that a homogeneous biobank of equivalent size could not replicate. At the same time, the authors note that individual prediction accuracy still declines with ancestry divergence, and that even the best multiancestry strategies leave gaps for populations that remain thinly sampled, such as Indigenous and Pacific Islander groups. Continued investment in globally representative genomic resources remains essential.

Looking forward, the study’s framework, with its systematic comparison of single-ancestry, meta-analyzed and cross-population score construction across dozens of traits, offers a template for future work. The authors have released their PRS-CS and PRS-CSx weights, analysis code and figure-generation scripts via Zenodo, enabling other researchers to reproduce and extend the findings. As polygenic prediction edges closer to routine clinical use, the message of this research is clear: equitable genetic risk prediction will not emerge automatically from bigger data. It will require deliberate attention to who is included in discovery studies, how their genetic architecture differs across traits, and which training strategy genuinely serves each population best. Diversity, in genomics as in so many domains, is not just a matter of representation but of scientific and clinical performance.

Subject of Research: Multiancestry polygenic risk score development and evaluation using the All of Us Research Program and UK Biobank

Article Title: All of Us diversity and scale yield context-dependent improvements in polygenic prediction

Article References: Tsuo, K., Shi, Z., Ge, T., Mandla, R., Hou, K., Ding, Y., Pasaniuc, B., Wang, Y., & Martin, A. R. (2026). All of Us diversity and scale yield context-dependent improvements in polygenic prediction. Nature Genetics. https://doi.org/10.1038/s41588-026-02734-4

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02734-4

Keywords: polygenic risk scores, All of Us Research Program, UK Biobank, genetic diversity, genome-wide association studies, ancestry-enriched effects, health disparities, precision medicine, multiancestry meta-analysis, genetic architecture, Nature Genetics, biobanks

The core difficulty that SPLENDID addresses stems from how human genetic variation is actually structured. Genetic ancestry does not come in neat, mutually exclusive boxes. Decades of population genetics have shown that human variation varies smoothly across geography and history, and many people, particularly those with recent admixture from multiple continental sources, sit at points in ancestry space that no pre-specified category can capture. When existing multi-ancestry prediction methods force every individual into a labeled group, they make two kinds of errors at once. Individuals whose ancestry blends several sources are assigned awkwardly to one group or another, discarding information, and the boundaries between groups themselves are arbitrary, changing with the reference panels and clustering algorithms used to define them. As the SPLENDID authors note, clinical decisions are rarely, if ever, made on the basis of an ancestry label, so a prediction framework that depends on such labels is poorly matched to real-world practice.

SPLENDID, whose name reflects its penalized-regression foundation, works directly with individual-level genotype data at biobank scale and treats ancestry as a continuum throughout the modeling pipeline. Rather than building separate ancestry-specific scores and then deciding which one to hand to each patient, the method fits a unified penalized regression model in which genetic effects are allowed to vary smoothly as a function of continuous ancestry information, typically captured through principal components of the genotype matrix. The framework uses modern sparse-regression machinery, including L0L1-type penalties that encourage the selection of a parsimonious set of variants, together with ensemble learning strategies that combine multiple candidate models tuned across the ancestry continuum. This produces one model, not many, and that model can be applied to any incoming individual without first asking which population bucket they belong to.

The mathematical machinery behind this idea is demanding, because biobank-scale genotyping data routinely contain hundreds of thousands of individuals and millions of variants. Fitting penalized models in which variant effects interact smoothly with ancestry dimensions requires efficient coordinate-descent and discrete-optimization algorithms that can operate on ultrahigh-dimensional data. The authors, including Tony Chen and Xihong Lin of Harvard T.H. Chan School of Public Health, Haoyu Zhang of the National Cancer Institute and Rahul Mazumder of MIT’s Sloan School of Management, built SPLENDID on computational tools of the kind developed for fast best-subset selection and large-scale sparse regression, making the method practical for datasets such as the UK Biobank and the All of Us Research Program. The software is implemented for R and released openly with tutorials, lowering the barrier for other groups to adopt the approach.

The empirical evaluation is extensive. In simulation studies designed to mimic the genetic architecture and demographic history of real human populations, SPLENDID outperformed a battery of existing polygenic prediction methods, including GWAS-based approaches and multi-ancestry tools such as PRS-CSx, CT-SLEB, PROSPER and iPGS. Notably, when the team forced other methods to produce a single pooled-ancestry score rather than ancestry-specific ones, prediction accuracy dropped sharply, particularly for the GWAS-based approaches. This comparison underscores the central argument of the paper: existing frameworks implicitly assume that ancestry labels are available and reliable, and when that assumption fails, their performance collapses in ways that SPLENDID’s continuum-based design avoids by construction.

The real-world tests were conducted on two of the largest and most demographically diverse genetic resources in existence. In the All of Us Research Program, whose more than 224,000 genotyped participants include substantial representation of African, Hispanic, Asian and Indigenous American ancestries, and in the UK Biobank, with roughly 340,000 analyzed participants, SPLENDID delivered significantly higher prediction accuracy than competing methods across nine continuous traits examined. The improvements were most pronounced for non-European and admixed individuals, exactly the groups for whom conventional scores perform worst. In admixed UK Biobank samples, identified through probabilistic classification against the 1000 Genomes Project reference panel, SPLENDID maintained accuracy where label-dependent methods faltered, validating the premise that treating ancestry as a continuum captures genuine genetic heterogeneity that discrete labels erase.

One striking illustration of why continuous modeling matters comes from the biology of blood cell traits. The paper examines individual-level effects of a variant, rs1213375, on mean corpuscular hemoglobin across the genetic ancestry spectrum, an example that connects to well-known ancestry-linked variation in hematological parameters such as the influence of Duffy status on neutrophil counts in people of African ancestry. Effect sizes of genetic variants are known from recent work to be broadly conserved across human groups, but fine-scale population structure and local ancestry still modulate how variants express themselves. A model that can represent these gradual shifts, rather than jumping between group-specific estimates at arbitrary boundaries, is better positioned to translate genomic discovery into accurate predictions for everyone.

The broader significance of this work lies in health equity. Polygenic risk scores are beginning to enter clinical workflows, with validated assays for conditions ranging from cardiovascular disease to breast cancer and with large consortia such as PRIMED dedicated to reducing disparities in polygenic risk assessment. If these tools systematically underperform in non-European populations, the benefits of genomic medicine will flow disproportionately to people already advantaged by existing research infrastructure, widening health disparities rather than narrowing them. Analyses spanning the Global Biobank Meta-analysis Initiative and studies of portability across nine ancestry groups have repeatedly documented the scope of this problem. SPLENDID’s contribution is to show that part of the solution does not require waiting for larger non-European datasets alone; it also requires statistical frameworks that use the full richness of ancestry information already present in biobanks.

There are practical implications as well. Because SPLENDID produces a single model, health systems deploying it do not need to maintain separate score pipelines for separate populations, nor do they need to assign patients an ancestry label before applying a score, a step that raises both logistical and ethical concerns given the fraught relationship between genetic ancestry and social identity. Scholars have cautioned repeatedly that genetic ancestry must be handled carefully in science and society, and a method that renders explicit labeling unnecessary sidesteps many of those pitfalls. The method’s demonstrated scalability also means it can be retrained as biobanks grow and diversify, and its dependence on individual-level data makes it well suited to trusted-access environments such as All of Us and the UK Biobank, where controlled analyses are the norm.

The path from a Nature Methods paper to routine clinical use is never short, and the authors are careful to frame SPLENDID as a tool for robust risk prediction across diverse populations rather than a finished clinical product. Validation for additional traits, disease endpoints and health systems will be needed, and continued investment in diversifying genomic datasets remains essential, because no statistical method can fully compensate for reference data that under-represent much of humanity. Still, the results reported across more than half a million participants in two flagship biobanks represent a meaningful advance: evidence that acknowledging the continuous, blended nature of human genetic ancestry, rather than forcing it into categories, yields measurably better predictions for the people who need them most. In a field where each incremental gain in accuracy can translate into earlier screening and better prevention, SPLENDID’s continuum-based approach offers a template for building genomic risk tools that work equitably across the full breadth of human diversity.

Subject of Research: A penalized regression framework, SPLENDID, that treats genetic ancestry as a continuum to improve polygenic risk prediction across diverse biobank populations.

Article Title: SPLENDID incorporates continuous genetic ancestry in biobank-scale data to improve polygenic risk prediction across diverse populations

Article References: Chen, T., Zhang, H., Mazumder, R., & Lin, X. (2026). SPLENDID incorporates continuous genetic ancestry in biobank-scale data to improve polygenic risk prediction across diverse populations. Nature Methods. https://doi.org/10.1038/s41592-026-03235-2

Image Credits: AI Generated

DOI: 10.1038/s41592-026-03235-2

Keywords: polygenic risk scores, genetic ancestry, SPLENDID, biobank-scale data, All of Us Research Program, UK Biobank, penalized regression, health disparities, admixed populations, Nature Methods, precision medicine, statistical genetics