Schizophrenia (SCZ) and bipolar disorder (BIP) share substantial common-variant liability but differ in cognitive, comorbidity, and treatment response. Here we decomposed these disorders into schizophrenia-predominant, bipolar-predominant, and shared psychosis dimensions to test whether these components show distinct pleiotropic and biological profiles than the original disorder. Using the largest available SCZ and BIP GWAS, we applied bidirectional mtCOJO and Genomic SEM to derive SCZcondBIP, BIPcondSCZ, and PSY-shared and validated them using inter-component genetic correlations, FinnGen psychiatric endpoints, and Genomic SEM latent factors. We then characterized each component across cognitive, cardiometabolic, and immune traits, followed by genomic risk-locus discovery, pathway analysis, developmental expression profiling, and drug-target enrichment. The three components showed marked divergence. SCZcondBIP was negatively correlated with cognition, education, metabolic syndrome, C-reactive protein, and neutrophil percentage, whereas BIPcondSCZ showed the opposite cognitive profile and shifted toward positive cardiometabolic and immune correlations. PSY-shared remained positively correlated with education and negatively correlated with cognitive task performance, immune and metabolic traits. We identified 248 consensus genomic risk loci, including 81 not detected in the input disorder GWAS. Biologically, PSY-shared was enriched for synaptic signalling, ion-channel, and neurodevelopmental pathways; SCZcondBIP primarily implicated synaptic-signalling and cellular-homeostasis pathways; and BIPcondSCZ showed weaker but distinct enrichment for synaptic-vesicular biology. Drug-target enrichment further separated the components, with strong antipsychotic enrichment for PSY-shared and distinct non-antipsychotic signals for the conditional factors. These findings show that SCZ and BIP genetic risk is best understood as biologically distinguishable shared and disorder-predominant dimensions that differentially map onto cognitive, cardiometabolic, immune, and molecular architecture. These findings provide a framework for evaluating whether component-specific polygenic scores improve stratification of cognitive, cardiometabolic, and inflammatory heterogeneity across severe psychiatric illness.

Population-scale proteomics is driving precision medicine by enabling systematic drug target identification and robust biomarker discovery. Comparable, well-powered studies across diverse global populations are essential to elucidate shared and population-specific disease pathways. Mass-spectrometry-based and multiplexed affinity-based assays have emerged as complementary, leading strategies for quantifying proteomic variation in large-scale population-based studies. With the objective of performing a comprehensive comparative evaluation of the latest assays for each of these technologies, we compare three leading affinity-based and mass-spectrometry-based proteomic platforms (SomaScan11K, Olink Explore HT, Orbitrap Astral with Seer Proteograph [MS-Seer]) in a multi-ethnic Asian cohort, to inform biomarker discovery and functional genomic studies of global populations. We find limited overlap of 1,740 proteins out of 12,825 total proteins quantified across the three platforms, with modest correlations (0.34–0.10). SomaScan had lower missingness (< 1

Early identification of individuals at risk of developing psychosis enables timely intervention and better clinical outcomes. Current approach relies on clinical assessments, such as the Comprehensive Assessment for At‑Risk Mental States (CAARMS), which provides an Ultra‑High‑Risk (UHR) classification but have limited predictive precision. Artificial intelligence (AI) models integrating neuroimaging, clinical, genetic, and biomolecular data show promise for improving prognostic accuracy, yet their clinical applicability remains uncertain. We systematically evaluated the contribution of six data modalities, ranging from socio‑environmental factors to polygenic risk scores, using data from 56 UHR participants in the Longitudinal Youth at‑Risk Study (LYRIKS) cohort. Modalities were iteratively included and excluded to generate candidate models tested on two tasks: (1) predicting transition to psychosis within 12 months and (2) predicting remission from UHR status within the same period. Statistical significance was assessed via permutation testing. Of the 56 UHR participants, 12 developed psychosis and 26 achieved remission from UHR status. The model combining CAARMS, socio-environmental risk factors, and social functioning (CAARMS + RISK + HiSoC) achieved the best overall performance for both predicting conversion (MCC = 0.71, SP = 0.93, SE = 0.78) and remission (MCC = 0.64, SP = 0.82, SE = 0.77), and was the only model to significantly outperform null models in the conversion task. Behavioural and socio‑environmental modalities demonstrate strong predictive value for forecasting outcomes in UHR within multimodal AI frameworks. While molecular and genetic modalities remain promising, further advances are needed to translate their high‑dimensional complexity into clinically useful predictive tools.

Abstract Tandem repeats (TRs) are implicated in over 70 Mendelian disorders and likely contribute to the “missing heritability” of complex traits and diseases, yet TR variations in Asian populations remain poorly characterized. Here, we constructed an Asian-specific SG10K-TR catalog by leveraging the SG10K_Health Dataset, comprising 916,274 autosomal TR loci genotyped in 9,490 individuals of Chinese (5,528), Malay (1,824), Indian (2,108), and other ancestries (30). Using a novel integrative measure for both repeat length and frequency variations, TRDDS, we found that population-level TR variations are selectively constrained in coding and promoter regions, whereas the enrichment of TRs with high population diversity was observed in regulatory sites with low chromatin accessibility and pathways related to neuronal functions. We also identified candidate TRs under selection that predominantly targets neuronal and synaptic architecture. Analysis of linkage disequilibrium (LD) patterns revealed that TRs are often poorly tagged by small variants, although we identified 123 candidate functional TRs that may underlie association signals previously attributed to nearby noncoding SNPs. Finally, TR-based GWAS of six anthropometric and lipid traits identified ten loci with genome-wide significant associations, including two novel loci for BMI ( LINC02817 ) and height ( UNC45B ), and a TR variant as causal candidate for a known GWAS locus at HMGCR for LDL. Together, this study establishes a critical Asian-specific TR resource and highlights the fundamental role of TR diversity in driving evolutionary neuroplasticity and shaping the genetic architecture of complex traits.

The Singapore National Precision Medicine (NPM) program is a three-phase whole-of-nation effort designed to develop scalable, evidence-based solutions for precision health tailored to Asia's diverse populations. Here we present NPM phase II (2020-2025), highlighting how large-scale precision medicine initiatives can drive new research insights, enable healthcare innovations and create economic value. Key achievements include the PRECISE-SG100K population dataset, which reflects Singapore's unique multi-ancestry Asian population; the successful translation of research findings into mainstream national healthcare using familial hypercholesterolemia as a use case; and the establishment of strategic public-private partnerships with diverse industry sectors. We also outline phase III (2025-2031), which aims to establish whole-genome sequencing as a foundational element of a patient's lifelong healthcare record for 10% of the population. NPM provides a model for how smaller countries, despite limited populations and finite resources, can contribute meaningfully to global scientific movements while preserving strategic independence and addressing national health challenges.

Genomic insights into psychiatric disorders remain heavily skewed toward European populations. In European-ancestry studies, educational attainment is typically negatively genetically correlated with major depression but paradoxically positively correlated with schizophrenia, raising the question of whether these relationships generalize across ancestries. We investigated whether this cross-trait architecture extends to East Asian ancestry (EAS). Using EAS GWAS summary statistics for major depressive disorder (MDD), schizophrenia (SZ), and educational attainment (EDU), we applied multi-trait (MTAG) and pleiotropy-informed (PLEIO) analyses to characterize shared genetic architecture across these traits. Across MTAG and PLEIO analyses, we identified 32 unique genome-wide significant loci (p < 5 x 10-8), including seven novel loci revealed in depression analysis that overlapped schizophrenia-associated signals, consistent with shared cross-trait architecture. Results reinforce a convergent risk architecture for affective and psychotic disorders in this population. Fine-mapping analyses prioritized variants mapping to candidate genes, including serine/threonine kinase VRK2, nominating targets for future follow-up. Cross-trait analyses supported a positive genetic relationship between EDU and MDD (rg = 0.308, p = 9.63 x 10-17) in East Asian data, contrasting to the negative correlation typically observed in European ancestry. These findings suggest that the genetic relationship between educational attainment and psychiatric risk may not be fully transferable across ancestries. In an independent cohort of individuals at ultra-high risk for psychosis, MTAG-derived polygenic risk scores improved case-control discrimination relative to single-trait GWAS-based scores. These results underscore the importance of ancestry-specific genomic frameworks for interpreting cross-trait psychiatric architecture and improving polygenic prediction. ### Competing Interest Statement The authors have declared no competing interest. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes All data produced in the present study are available upon reasonable request to the authors Singapore Ministry of Health’s National Medical Research, NMRC/CG1/005/2021-IMH, NMRC/TCR/003/2008

Background: Memory influences decision-making regarding eating behaviour. However, marketing utilises our memory to exploit our consumptive habits. Whether we have an inherent bias towards memorising visual cues associated with high-calorie foods and how such bias contributes to our health outcomes remains to be determined. We therefore aimed to develop and validate a computerized, image-based Food Memory Bias Task (FMBT) for use in multiethnic Asian populations. Methods: In the development phase involving 172 participants, 91.9% rated the instrument as user-friendly and we optimised the instrument format into 12 visual food cues and 30-min delay period to avoid a ceiling effect. In the validation phase involving 184 multi-ethnic Asian participants (49.3 (14.5) years old, 44.6% male, 71.7% Chinese), FMBT memory score was associated with the brief assessment of cognition memory (β(p) = 0.23 (5.1 × 10−6)), and general cognition score ‘g’ (β(p) = 0.30 (6.5 × 10−9)). Results: Higher memory bias was associated with lower memory score in the low-calorie bias group (r=−0.33, p=0.002), but not in the hig h-calorie bias group. In this present study, despite randomly administering two versions of the test across visits, the FMBT bias score demonstrates a significant learning effect and low reproducibility (r(p) = −0.14 (0.11) across visits). Food memory bias score is associated with higher food approach traits according to the Adult-Eating Behaviour Questionnaire (β(p) = 0.23(0.015)), independent of age, sex, and ethnicity. Conclusion: FMBT provides an opportunity to address epidemiological questions on how food memory bias influences dietary habit and cardiometabolic health, and how it is in turn shaped by our built environment.

Importance A major challenge in the management of psychotic disorders is the lack of objective biomarkers. While angiotensin-converting enzyme ( ACE ) has been identified as a risk gene for schizophrenia, a major knowledge gap remains regarding whether and how this genetic impact results in biological dysfunction through its protein product. Objective To compare ACE levels, enzymatic activity, and genetic variations between patients with early-stage psychosis and healthy controls and between patient subgroups with and without treatment-resistant psychosis. Design, Setting, and Participants This cross-sectional study included patients aged 13 to 35 years with early-stage psychosis (onset within the prior 2 years) who were recruited at the Johns Hopkins Schizophrenia Center in Baltimore, Maryland, between October 30, 2013, and February 9, 2018. Data were analyzed from June 1, 2018, to June 22, 2026. Exposure Early-stage psychosis. Main Outcomes and Measures The main outcomes were ACE protein levels in cerebrospinal fluid and serum, genomewide and ACE gene–specific polygenic risk scores, and ACE canonical enzymatic activity in serum. Results The study included data from 200 participants (mean [SD] age, 22.8 [4.1] years; 129 [64.5%] male), including 78 healthy controls and 122 patients with early-stage psychosis. ACE protein levels were significantly lower in both cerebrospinal fluid (Cohen d = −1.16; P = .005) and serum (Cohen d = −0.92; P < .001) in patients compared with controls, with significant correlations between biofluids ( r = 0.34; P = .04). In patients with psychosis, higher ACE genetic burden for schizophrenia was associated with lower ACE protein levels ( r = −0.35; P = .002) but not with ACE canonical (renin-angiotensin system) enzymatic activity. Patients with treatment-resistant cases had lower serum ACE protein levels than patients with non–treatment-resistant cases (Cohen d = −0.50; P = .03). There was no difference in canonical enzymatic activity between these 2 patient subgroups. Conclusions and Relevance This cross-sectional study found that ACE protein levels were significantly lower in cerebrospinal fluid and serum in patients with early-stage psychosis compared with controls. Between patient subgroups, ACE protein levels were significantly lower in serum in patients with treatment-resistant psychosis compared with those with non–treatment-resistant psychosis. A contrast was observed between ACE protein levels and catalytic function, where ACE genetic variations showed a significant correlation with protein levels but not with enzymatic activity.

Genomewide association studies (GWAS) of brain scans are complicated by the large number and high collinearity of the available image-derived phenotypes (IDPs). Here, we present DIMPLE-GWAS (Dimensionality reduction and Integrated Multi-Phenotype Landscape Explorer for GWAS), a dimensionality-reduction framework designed to identify latent genetic architecture across high-dimensional pleiotropic phenotypes. This approach, applied to ∼4000 IDPs from ∼33K European ancestry participants in the UK Biobank, yielded 25 biologically interpretable latent phenotypes; this structure was validated in the independent ABCD cohort. The DIMPLE-GWAS clusters demonstrated substantially greater heritability than the input IDPs and yielded greater power for locus discovery, including 104 genomewide-significant loci not reported in prior GWAS of individual IDPs. These genetically defined phenotypes only partially aligned with conventional brain atlas boundaries based on gyral, cytoarchitectonic, or functional features. Instead, they revealed distinct patterns of brain organization and novel genetic relationships with neurologic, psychiatric, cognitive, and behavioral phenotypes.

Background: Schizophrenia (SCZ) and bipolar disorder (BD) share clinical and genetic features, which complicates differential diagnosis. Separating shared from disorder-associated liability may identify distinguishing genetic features. Methods: We applied genome-wide association study (GWAS)-by-subtraction to SCZ (n = 130,644) and BD (n = 780,742) summary statistics and derived the SCZ-BD factor. Polygenic risk score (PRS) performance was evaluated in the UK Biobank (UKB; European) and the Korean Multicenter Psychiatric Cohort (East Asian). Multi-omics analyses prioritized SCZ-BD-associated genes and proteins. Brain imaging analyses evaluated associations between the SCZ-BD PRS and regional brain volumes. Single-cell RNA sequencing analyses characterized the cell-type-specific and developmental expression patterns of the prioritized genes. Findings: The SCZ-BD factor explained 57% of SCZ genetic variance and identified 35 loci, including seven previously unreported. In UKB, the SCZ-BD PRS outperformed the SCZ PRS (odds ratio, 3.38 vs 1.57) in distinguishing SCZ from BD. Eleven genes and proteins were prioritized, including HCG4, ZKSCAN3, and CACNA2D1, which were not identified in the SCZ GWAS. CACNA2D1 showed convergent multi-omics and knockout mouse support. The SCZ-BD PRS was associated with seven brain regions, and prioritized genes showed neuronal enrichment and developmental expression. Interpretation: Modeling SCZ liability after accounting for BD-related liability may identify associations not captured by the conventional SCZ GWAS. The PRS and molecular candidates require further validation before clinical or therapeutic application.

Schizophrenia (SCZ) and bipolar disorder (BIP) share substantial common-variant liability but differ in clinical course, cognition, and treatment response. We previously resolved this overlap into three cognitively divergent components, SCZcondBIP (negatively correlated with cognition and education), BIPcondSCZ (positively correlated with both), and PSY-shared (negative with cognition, positive with education), representing distinct neurocognitive axes. Here, we applied directional pleiotropic meta-analysis (PLEIO) to integrate these components with cognitive task performance and educational attainment, yielding 818 consensus loci of which 514 (63%) were shared across all three components and 220 were component-specific, including 99 novel loci absent from individual GWAS. Each component was partitioned into concordant (aligning with the expected cognitive direction, e.g., increased cognition with reduced SCZcondBIP risk) and discordant (reverse pattern, e.g., increased cognition with increased SCZcondBIP risk) locus sets. Pathway analyses revealed marked biological divergence: SCZcondBIP concordant loci were enriched for neurodevelopmental pathways and discordant loci for cellular-homeostasis pathways, suggesting two separable processes: an early developmental-regulatory branch linking cognitive disadvantage to increased risk, and a homeostatic/metabolic-stress branch enabling cognitive advantage despite increased risk. BIPcondSCZ concordant loci, where increased cognition co-occurs with increased bipolar risk, were enriched for synaptic pathways. Unlike neurodevelopmental pathways, which emerged as primary cognitive determinants in SCZcondBIP, synaptic pathways may principally mediate disease liability without substantially impacting cognition, explaining preserved or enhanced performance alongside increased bipolar risk. Discordant loci, where decreased cognition co-occurs with decreased bipolar risk; implicated cellular-homeostasis pathways through mitochondria mediated pathways, consistent with mitochondrial dysfunction and reduced cellular resilience contributing to bipolar pathophysiology and progressive cognitive decline. PSY-shared concordant loci showed nominal synaptic and cellular-homeostasis enrichment, while discordant loci implicated neuroimmune and vesicular processes, directionally consistent with disorder-specific partitions. Schizophrenia and bipolar disorder genetic risk comprises biologically coherent disorder-specific and shared dimensions; integrating these with cognition exposes directional pleiotropic architecture across divergent developmental, synaptic, and metabolic pathways, providing a mechanistic framework for understanding cognitive heterogeneity in severe psychiatric illness.

With recent advances in high-throughput proteomic technologies, population-scale plasma proteomics datasets, often linked to extensive genetic and phenotypic information, have become increasingly accessible. Yet the relationships between circulating protein levels, brain imaging phenotypes, and risk for neurological and psychiatric disorders remain largely unexplored. Proteome-wide association studies offer a promising approach for elucidating biological mechanisms that connect genetic variation to complex brain-related traits and diseases. In this study, we integrated protein quantitative trait loci (pQTLs) from the two largest plasma proteomic resources (the UK Biobank Pharma Proteomics Project [UKB-PPP] and Ferkingstad et al. [deCODE]) with genome-wide association studies of brain imaging-derived phenotypes in UK Biobank using Mendelian randomization and colocalization analyses. We identified 120 cis and 20 trans associations between plasma proteins and imaging phenotypes and validated these findings using brain tissue-derived proteomic and transcriptomic datasets. Multivariable Mendelian randomization revealed eleven plasma proteins (coding genes APOE, ARL3, MICB, NSF, RHOC, RSPO3, ENPP2, BTN2A1, EIF2AK3, MRVI1, and OPLAH ) with significant direct effects on the risk of Alzheimer's disease, Parkinson's disease, multiple sclerosis, bipolar disorder, and schizophrenia. Single-cell expression and pathway enrichment analyses further revealed cell-type-specific effects and distinct biological processes underlying these protein- disease associations. Together, these findings demonstrate robust links between plasma protein variation and brain structure, delineate protein-disease pathways, and highlight the cellular and molecular mechanisms that contribute to neurobiological diversity and pathology.

We compared three leading affinity-based and mass-spectrometry-based proteomic platforms (SomaScan11K, Olink Explore HT, Orbitrap Astral with Seer Proteograph [MS-Seer]) in a multi-ethnic Asian cohort, to inform biomarker discovery and functional genomic studies of global populations. We found limited overlap (1,740 proteins) across the three platforms, with modest correlations (0.34-0.10). SomaScan had lower missingness (<1%) and CV (<10%), compared to Olink (51%, 23%) and MS-Seer (12%, 19%). The new assays in Olink Explore HT (absent in Explore3072) primarily drove its higher missingness and CV. The number of phenotypic associations varied by trait, while the number of genetic associations ( cis- pQTLs at P<5e-08) were similar for Olink and SomaScan. Protein levels differed between ethnicities, with SomaScan identifying more ethnicity-differentiated proteins than Olink (FDR<0.05). Finally, SomaScan ANML normalization attenuated biologically relevant associations in our study. These findings underscore the importance of platform evaluation and data normalization strategies for application in large-scale, diverse population cohorts. ### Competing Interest Statement L.D.W works for Alnylam Pharmaceuticals and holds stocks as part of employment. O.B works for Bayer AG and does not hold stocks as part of employment. Z.D works for Boehringer Ingelheim Pharma GmbH & Co. KG and does not hold stocks as part of employment. J.F works for Novo Nordisk A/S and holds minor share portions as part of employment. The rest of the authors declare no competing interests. NMRC Singapore, NMRC/StaR/0028/2017, MOH-000271-00, NMRC/PRECISE/2020

The burden of cardiovascular disease is rising in the Asia-Pacific region, in contrast to falling cardiovascular disease mortality rates in Europe and North America. Here we perform quantification of 883 metabolites by untargeted mass spectroscopy in 8,124 Asian adults and investigate their relationships with carotid intima media thickness, a marker of atherosclerosis. Plasma concentrations of 3beta-hydroxy-5-cholestenoate (3BH5C), a cholesterol metabolite, were inversely associated with carotid intima media thickness, and Mendelian randomization studies supported a causal relationship between 3BH5C and coronary artery disease. The observed effect size was 5- to 6-fold higher in Asians than Europeans. Colocalization analyses indicated the presence of a shared causal variant between 3BH5C plasma levels and messenger RNA and protein expression of ferredoxin-1 (FDX1), a protein that is essential for sterol and bile acid synthesis. We validated FDX1 as a regulator of 3BH5C synthesis in hepatocytes and macrophages and demonstrated its role in cholesterol efflux in macrophages and aortic smooth muscle cells, using knockout and overexpression models. Sadhu et al. identify and functionally validate ferredoxin-1 (FDX1) as a determinant of cholesterol metabolism and cardiovascular risk in Asian populations.

BACKGROUND:Identifying biomarkers that predict social and cognitive outcomes in individuals at ultra-high risk (UHR) for psychosis remains a key challenge in preventive psychiatry. While genetic factors contribute to psychosis vulnerability, specific markers that predict individual trajectories of functional decline or resilience are still unclear. METHODS:In a 24-month longitudinal study involving UHR (n = 45) and healthy control participants (n = 54), we investigated for the first time the predictive causal relationship between key immunological genes (FABP5 family and immunoglobulins) and social-cognitive outcomes. Participants completed comprehensive assessments at baseline and four 6-month intervals. We used regression modelling and dynamic Bayesian network analysis to identify predictive relationships between gene expression and behavioral outcomes over time. RESULTS:FABP5 family genes (FABP5P1, FABP5P11, FABP5P9) significantly predicted verbal memory (β = 0.233, p = 0.002); working memory (β = 0.225, p = 0.004), and social skills (β =-0·190, p < 0.029), respectively, at 24 months in the UHR group. Immunoglobulin-related genes showed distinct effects: FCGR2B predicted object recognition ability (β = 0.233, p = 0.014), while GOT2 inversely predicted planning ability (β = -0.147, p = 0.067). Network analysis revealed UHR-specific temporal dependencies absent in controls, with FCGRT emerging as a central node linking genetic markers to changes in processing speed and perceptual closure. CONCLUSIONS:This study provides the first evidence that FABP5 and immunoglobulin-related genetic markers can predict social-cognitive trajectories in individuals at risk for psychosis. These findings support the use of genetic profiling for early identification and highlight new opportunities for personalized preventive strategies in psychiatry.