Identification of Hub Genes Involved in Early-onset Schizophrenia: From Genetic Susceptibility to Predicted Regulated Gene Expression

Identification of Hub Genes Involved in Early-onset Schizophrenia: From Genetic Susceptibility to Predicted Regulated Gene Expression | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Identification of Hub Genes Involved in Early-onset Schizophrenia: From Genetic Susceptibility to Predicted Regulated Gene Expression Yawen Jen, Sung-Liang Yu, Po-Chang Hsiao, Po-Hsiu Kuo, Chih-Min Liu, and 15 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5833160/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract BACKGROUND Despite a high heritability of schizophrenia (SZ), only limited variance was attributed to gene loci or the polygenic risk score in genome-wide association studies (GWAS). Early-onset SZ, a more homogeneous SZ subtype, may aid in bridging the genotype-phenotype gap and the identification of its hub genes is critical for early intervention in clinical practice. We aimed to examine the gene expression risk score (GeRS) in patients from both multiplex and simplex families to identify hub genes for early-onset SZ, and perform enrichment analysis to understand the biological functions of the hub genes. METHODS Based on the GWAS genotype data from patients with SZ in multiplex families (223 early-onset and 372 late-onset) and those from simplex families (matched for sex and onset age), GeRSs for SZ (SZ-GeRSs) were estimated using the SNP-expression prediction model derived from existing brain tissues of patients with psychiatric disorders. Module-based SZ-GeRS was summed over genes from empirically derived gene clusters, network analysis was conducted to identify hub genes, and enrichment analysis was used for functional mapping. RESULTS Among the 13 modules from existing coexpression analyses of postmortem brains of patients with psychiatric disorders, the meta-analysis revealed that associations with early-onset SZ existed for the GeRS of module 10 in subset, M10 sub -GeRS (adjusted odds ratio [aOR] = 1.38, 95% CI = 1.22–1.57), and six hub genes, M10 hub -GeRS (aOR = 1.22, 95% CI = 1.07–1.39), after adjustment for covariates. Functional mapping of the genes revealed their enrichment in excitatory neurons and immune-regulatory pathways. CONCLUSIONS GeRS for SZ helps identify six hub genes for early-onset schizophrenia, and the enrichment analysis sheds light on their possible roles in the pathophysiology. These findings will enhance the understanding of SZ etiology and may contribute to early screening and personalized prevention efforts. Schizophrenia Polygenic risk score Gene expression risk scores Early onset Coexpression analyses Network analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Schizophrenia (SZ) is a leading cause of disability worldwide (1), with a median lifetime prevalence of 6.35 per 1000 people (2). Despite a heritability of 80% (3, 4), only limited variance in SZ was attributed to gene loci or the polygenic risk score (PRS) in genome-wide association studies (GWASs) (5-7). SZ is a heterogeneous disorder characterized by psychotic symptoms (delusions and hallucinations) and cognitive impairments (8). Given that a substantial proportion of the heritability of SZ remains unexplained, varying levels of gene expression resulting from the interaction of genes and environmental factors are postulated to further contribute to the occurrence of SZ, although collecting postmortem brain tissues for gene expression analysis is challenging (9, 10). Instead, predicting genetically regulated gene expression in the brain by integrating an existing tissue-specific single-nucleotide polymorphism (SNP)-expression prediction model with GWAS-based genotype data, i.e., transcriptome imputation, is an alternative approach (11, 12). Studies applying the genetically regulated gene expression risk score (GeRS) for SZ have revealed their associations with SZ or SZ-related traits, including clinical symptomatology and patients’ cognitive performance (12-15). However, only a few genes with potential biological mechanisms that could aid in clinical implementation have been identified in these studies. Early-onset SZ, occurring in about 8.2% of cases diagnosed before age 18 (16), is a distinct and homogeneous subtype of SZ that serves as a critical phenotype for dissecting its genetic etiology. This subtype is associated with more severe symptoms (17), a greater genetic burden (18, 19), and a higher risk of treatment resistance (20) compared to adult-onset SZ. Findings from GWAS for SZ have identified genetic loci of suggestive significance in early-onset SZ (21, 22), highlighting its potential as a phenotype for uncovering genes relevant to early intervention. However, the genetic basis of early-onset SZ remains largely unknown (23). To elucidate the genetic underpinnings of early-onset SZ, we aimed to examine the gene expression risk score (GeRS) in patients from both multiplex and simplex families to identify hub genes for early-onset SZ, and perform enrichment analysis to understand the biological functions of the hub genes. Methods And Materials Study workflow The study workflow is shown in Additional file 1 : Fig ure S1 . We first predicted genetically regulated gene expression levels in the dorsolateral prefrontal cortex (DLPFC) using the PrediXcan (https://github.com/hakyimlab/PrediXcan), based on the GWAS genotype data from patients with SZ in multiplex families. The SNP-expression prediction model was based on the CommonMind Consortium (CMC; including 254 SZ, 52 bipolar disorder, and 279 controls) (https://predictdb.org/) (12). To extract disease-associated gene expression changes, we then performed a transcriptome-wide association study (TWAS) applying S-PrediXcan (24) with the same SNP-expression prediction model and the latest GWAS summary statistics from the SZ cohort of the Psychiatric Genomics Consortium (PGC; 76,755 patients with SZ and 243,648 controls) (7). After that, we calculated GeRS for each gene by weighting its predicted expression based on the corresponding effect size from the TWAS for SZ. Furthermore, we constructed the module-based GeRS by summing the GeRS of genes listed in the modules obtained from the published weighted gene coexpression network analysis (WGCNA) results (participants including 150 SZ, 94 bipolar disorder, 87 major depression, 50 autism, 17 alcoholism, and 293 controls) (25). Finally, we evaluated the relationship between the module-based GeRS and early-onset SZ in the multiplex sample and explored the hub genes within the early-onset-related module using Ingenuity Pathway Analysis (IPA®, QIAGEN Redwood City, CA, USA; www.qiagen.com/ingenuity). We then repeated the analysis in the simplex sample of patients with SZ from trio families. Participants This study included two samples: a multiplex sample and a simplex sample. The multiplex sample was derived from the Taiwan Schizophrenia Linkage Study (TSLS), which recruited Han Chinese families with more than two siblings affected by SZ and their first-degree relatives across Taiwan from 1998–2002 (26, 27) (see details in the Additional file 1 ). After quality control, 595 patients and 570 unaffected relatives from 314 families, with age-at-onset information for the first psychotic episode obtained via the Diagnostic Interview for Genetic Studies (DIGS) (28), were included. To examine the findings derived from the multiplex sample in patients with different genetic loadings, i.e., the simplex sample, we also included patients who had individual genotype data and age at onset information from another independent sample, the Schizophrenia Trio Genomic Research in Taiwan (S-TOGET). Among the 2,923 patients with SZ from simplex families in S-TOGET, 1,649 had genotype data, and 1,635 of them also had information on age at onset. More detailed information about the S-TOGET cohort has been described previously (29, 30). To minimize sociodemographic differences between the TSLS and S-TOGET samples, sex- and onset-age-matched subsamples were selected for comparable analysis ( Additional file 1: Table S1 ). Both studies were approved by the Institutional Review Board of National Taiwan University Hospital, and all participants provided written informed consent (see details in the Additional file 1 ). Measurements The age at onset and basic sociodemographic characteristics were extracted from the DIGS, with the former obtained from the psychosis section that inquired about the age at onset of the first psychotic episode. We classified age at onset as binary, defining early-onset SZ as psychotic symptoms emerging before age 18, based on our prior research (30). Furthermore, cognitive function was assessed using the Continuous Performance Test (CPT) for sustained attention (31) and the Wisconsin Card Sorting Test (WCST) for executive function (32). CPT sensitivity index (d′), WCST perseverative errors, and categories achieved were included as cognitive indices. Further details on measurements are provided in the Additional file 1 . Genotyping, quality control, and imputation Genotyping for the multiplex and simplex samples was performed using the PsychChip array version 1.1 (Illumina, San Diego, CA), containing 588,628 SNPs. Standard quality control (QC) procedures (33) were followed (see details in the Additional file 1 ). For the multiplex sample, 1,224 participants and 266,899 SNPs that passed QC were used for genotype imputation, and principal components (PCs) were calculated to account for ancestry in subsequent analyses. The genomic inflation factor (λ) of this final sample was 1.07, and the corresponding QQ plot is shown in Additional file 1: Figure S 2. After genotype imputation, the post-imputation QC filtered for the variants included poor imputation quality (INFO score < 0.8) and low minor allele frequency (MAF < 0.1). Ultimately, 4,853,692 SNPs were included in later analyses. For the simplex sample, the same QC procedures resulted in 293,643 SNPs that passed QC, with imputation yielding 5,154,665 SNPs. Details of the QC process have been described previously (29, 30). Calculation of PRS PRSs for schizophrenia (SZ-PRS) were calculated for both samples using the p-value thresholding method, utilizing the latest version of summary statistics from the PGC SZ cohort (7). All the SZ-PRS were normalized to z scores for easier interpretation. The best-performing p-value threshold for the SZ-PRS, which explained the most variance in early-onset SZ, was selected for further analysis. Additional details are available in the Additional file 1 . Deriving the effect sizes of predicted gene expression The tissue-specific TWAS for SZ was performed by S-PrediXcan (24), which integrates GWAS summary data of patients with SZ with a SNP-expression prediction model to identify genes associated with SZ (see see details in the Additional file 1 ). The existing SNP-expression prediction model from CMC (12), which was derived from DLPFC brain tissue samples, was used to perform the TWAS for SZ. Predicting gene expression using individual genotypes To determine the genetically regulated expression predicted for a SNP, hereafter referred to as predicted gene expression, we used PrediXcan (11) to incorporate the genotypes of individual patients and the corresponding SNP-expression prediction model. The DLPFC SNP-expression prediction model from CMC was used to predict gene expression, and only those genotypes that passed standard quality control procedures were subjected to such prediction. Calculation of SZ-GeRS For each gene, its predicted SZ-GeRS was calculated by weighting the predicted gene expression for each patient using the effect size from the TWAS for SZ described in the preceding section. In addition to the SZ-GeRS of individual genes, we constructed two other types of SZ-GeRS by summing genes to identify biologically meaningful relationships with early-onset SZ. The first type of gene summary SZ-GeRS was a module-based SZ-GeRS, where we utilized gene lists from 13 coexpression modules related to psychiatric disorders identified in a previous WGCNA (25). These modules were derived from postmortem brain tissues of patients with psychotic disorders compared with healthy controls, and we summed the individual SZ-GeRS over the genes listed in those modules. The other type of gene summary SZ-GeRS was constructed by summing over the hub genes of a module. Hub genes were identified via network analysis using IPA, selecting genes regulating the expression of four or more other genes in the top-ranking network. All the SZ-GeRS results were then normalized to z scores for easy interpretation. Furthermore, we utilized the gene lists from the newly derived coexpression modules identified by the same research group using narrowed samples (32) to recalculate the module-based GeRS and assess the reproducibility of the early-onset-related module and its hub genes identified in these procedures (see details in the Additional file 1 ). Enrichment analyses To understand possible biological functions of the hub genes, we conducted enrichment analyses using the GENE2FUNC function from the functional mapping and annotation platform (FUMA, https://fuma.ctglab.nl/) to annotate them in the biological context. The gene expression datasets from the BrainSpin consisted of brain tissue samples across various developmental stages. We used the Molecular Signatures Database (MsigDB) (34) to assess enrichment in chemical and genetic perturbation gene sets. Additionally, cell type enrichment analysis was conducted using the PsychSCREEN gene portal (https://psychscreen.beta.wenglab.org/psychscreen/gene), designed to explore single-cell expression of genes in psychiatric disease. The Schizophrenia Bipolar Disorder Multi-Omics Sequencing dataset from PsychENCODE was utilized for this analysis. Statistical analyses Associations between risk scores (including SZ-PRS and all types of predicted SZ-GeRS) and early-onset SZ were assessed using a mixed-effects logistic regression model for the multiplex sample, with family index as a random effect. Covariates included sex, education level, and the first four ancestry PCs. For the simplex sample, a logistic regression model was used with identical covariates. The proportion of explained variance was estimated by calculating the change in Nagelkerke's pseudo-R 2 between the null model (containing only the covariates) and the full model (containing covariates + risk scores) (35). We then conducted a meta-analysis of both samples using the R package “metafor” (36), employing a restricted maximum likelihood estimator with inverse-variance weighting. The SZ-PRS and PCs were computed using the PLINK (37), whereas regression analyses and Nagelkerke's pseudo-R 2 calculations were conducted using the R packages “lme4” and “rcompanion” (version 4.2.2), respectively. Results Distribution of sociodemographic characteristics and neurocognitive performance In the multiplex sample of 595 patients with SZ from 314 multiplex families, 223 patients with early-onset SZ (≤ 18 years of age) were younger, had a lower education level, and had lower undegraded CPT d´ than 372 patients with late-onset SZ (> 18 years of age), with no differences in terms of degraded CPT d´ and two indices of the WCST (Table 1 ). In the simplex sample of 595 patients with SZ matched 1:1 by sex and onset age to the multiplex sample, the early- and late-onset patients presented similar differences in sociodemographic and neurocognitive performance. Table 1 Distributions of sociodemographic characteristics and neurocognitive performance indices in patients with schizophrenia grouped by early- and late-onset in the multiplex and simplex samples, respectively. Multiplex sample Simplex sample Variables Early-onset patients a (N = 223) Late-onset patients a (N = 372) Group Comparison b Early-onset patients (N = 223) Late-onset patients (N = 372) Group Comparison c Mean (S.D.) Mean (S.D.) p Mean (S.D.) Mean (S.D.) p Male sex, n % 131 (58.7%) 224 (60.2%) 0.72 131 (58.7%) 224 (60.2%) 0.72 Age, years 30.72 (7.69) 34.63 (6.72) <0.001** 30.72 (7.61) 35.94 (7.47) < 0.001** Age at onset, years 15.99 (2.13) 24.17 (5.09) < 0.001** 16.08 (1.89) 24.03 (4.95) < 0.001** Education level, years 9.68 (2.49) 11.07 (2.83) < 0.001** 11.41 (2.40) 12.21 (2.48) 0.001** CPT indices d Undegraded CPT d´ -2.74 (2.15) -2.26 (2.03) 0.018* -2.16 (2.07) -1.70 (1.77) 0.019* Degraded CPT d´ -2.56 (1.69) -2.51 (1.47) 0.782 -1.46 (1.66) -1.27 (1.71) 0.276 WCST indices e Categories achieved -1.09 (0.72) -1.04 (0.84) 0.476 -0.90 (0.98) -0.88 (0.96) 0.882 Perseverative errors 1.23 (1.74) 1.39 (1.82) 0.361 1.15 (1.70) 1.28 (1.72) 0.457 *p value < .05; **<0.01 a 106 individuals in the early-onset group (47.5%) and 250 individuals in the late-onset group (67.2%) were siblings from multiplex families with SZ. b The group comparisons were conducted using a mixed-effect logistic regression model with family index as a random effect. c The group comparisons were conducted using logistic regression. d The adjusted z scores were derived by means of standardizing the raw scores with adjustments for sex, age and education against a community sample of 345 individuals (31). e The adjusted z scores were derived by means of standardizing the raw scores with adjustments for sex, age and education against a group of 392 healthy controls (32). Association of PRS for SZ with early-onset SZ Regarding the extent to which the difference in onset age was attributed to genetic predisposition, the SZ-PRS at the p -value threshold of 0.1 was associated with early-onset SZ ( p = 0.03) and explained the most variance of early-onset SZ (pseudo-R 2 = 1.11%) in the multiplex sample (Fig. 1 A), whereas the SZ-PRS at the p -value threshold of 0.01 was not associated with early-onset SZ (p = 0.14) but explained the most variance of early-onset SZ (pseudo-R 2 = 0.50%) in the simplex sample (Fig. 1 B). We then compared the SZ-PRS scores of the two groups of patients with those of unaffected relatives using the same method of PRS derivation. Among both the multiplex (Fig. 1 C) and simplex (Fig. 1 D) samples, patients with early-onset SZ had the highest average SZ-PRS score, followed by patients with late-onset SZ and then their unaffected relatives. Predicted GeRS for SZ Before building the GeRS for our samples, we obtained a weighting system for the predicted expression of individual SNPs through TWAS of SZ in an existing database to identify the associations between the cis -genetic component of expression and SZ. Among the 10,358 genes included in the existing SNP-expression prediction model from the CMC (12) for the DLPFC, the expression levels of 10,306 genes identified in the GWAS of PGC SZ cohort using S-PrediXcan were successfully predicted, with the prediction performance per gene ranging from 0.8–81.5% and a mean of 8% (s.d.=11.1%). A comparison of the predicted expression levels of these genes in the PGC for patients with SZ and their controls revealed that 46 genes showed associations with SZ that reached genome-wide significance (p < 5×10 − 8 ) ( Additional file 2: Supplementary Data 1 ). We then utilized the effect size of the preceding TWAS of SZ as the weight of the predicted expression of individual SNPs to derive the predicted SZ-GeRS for our samples. Figure S3 depicts the genome-wide association between the predicted SZ-GeRS and early-onset SZ in the multiplex sample, with 7834 genes predicted to be expressed; none of them showed an association that reached genome-wide significance. When the same analysis was repeated in the simplex sample, 7987 genes were predicted to be expressed, with none reaching genome-wide significance ( Additional file 1: Figure S4 ). Module-based GeRS for SZ and early-onset SZ To capture the possible contribution of the combination of expression levels of a group of genes, we turned to 13 modules derived from coexpression analysis in a previous study using postmortem brain tissues of patients with five major psychiatric disorders versus those of healthy controls (25). When the association between individual module-based SZ-GeRS and early-onset SZ was assessed in the multiplex sample, the SZ-GeRS based on module 10 (M10-GeRS) explained the most variance in early-onset SZ (pseudo-R 2 = 2.4%, p = 0.005; Fig. 2 A). Among the 161 genes included in the original module 10, 94 had SNP-expression weights in the DLPFC, 57 of which were predicted to have genetically regulated expression in our multiplex sample ( Additional file 1: Table S2 ) . We then examined the same 94 genes located in the DLPFC of module 10 in the simplex sample and found that 61 genes were predicted to have genetically regulated expression, with 56 of them also appearing in the multiplex samples. Hence, these 56 genes were considered a subset of M10 (M10 sub ), and the corresponding SZ-GeRS was denoted as M10 sub -GeRS. Patients with early-onset SZ presented significantly greater standardized M10 sub -GeRS values than did those with late-onset SZ in both the multiplex sample (Fig. 2 B) and the simplex sample ( Fig. 2 C ) . Hub genes for early-onset SZ To identify possible hub genes, we conducted a network analysis of the 56 genes of M10 sub -GeRS in the multiplex sample using the IPA and revealed four networks with functionally connected genes, in which only one network had an IPA score > 40, i.e., a score of 75 for Network 1 ( Additional file 2: Supplementary Data 2) . Focusing on this network, we identified RUVBL2, COPS6, TUBA4A, PSMB5, PSMD2 , and LRPPRC as hub genes that were correlated with the expression of more than four genes within the network (Fig. 3 and Additional file 2: Supplementary Data 3) . When these six hub genes were used to derive M10 hub -GeRS, the distribution of early-onset SZ tended to be greater than that of late-onset SZ in both the multiplex and simplex samples, although the difference reached statistical significance only in the latter ( Additional file 1: Figure S8) . Meta-analysis of the association We then conducted a series of multivariable logistic regression analyses, adjusting for sex, educational level, and four ancestry PCs for the multiplex sample (Fig. 4 A) and the simplex sample (Fig. 4 B), respectively. After accounting for SZ-PRS, the variance explained by M10 sub -GeRS (or M10 hub -GeRS) was 2.5% (or 0.7%) in the multiplex sample and 3.6% (or 1.3%) in the simplex sample. In terms of the association of M10 sub -GeRS (or M10 hub -GeRS) with early-onset SZ, the corresponding adjusted ORs [aORs] in Model 3 (or Model 4) for the multiplex sample (Fig. 4 A) and the simplex sample (Fig. 4 B) were similar in direction and magnitude, although some did not reach statistical significance, i.e., M10 hub -GeRS in Model 4 for the multiplex sample. Hence, we aggregated both the multiplex and the simplex samples to conduct a meta-analysis to estimate summary effects (Fig. 4 C). The meta-analysis revealed that the aORs of SZ-PRS (Model 0), M10 sub -GeRS (Model 1), and M10 hub -GeRS (Model 3) were significantly greater than one. Nevertheless, even after adjustment for SZ-PRS and the covariates, significant associations with early-onset SZ existed for M10 sub -GeRS (aOR = 1.38, 95% CI = 1.22–1.57) and M10 hub -GeRS (aOR = 1.22, 95% CI = 1.07–1.39). That is, reducing the genes in gene-summing GeRS from 56 (M10 sub -GeRS) to 6 (M10 hub -GeRS) did not diminish the association. To examine the robustness of the module-based GeRS, we used another set of modules derived from 3 diagnostic groups (SZ, bipolar disorder, and autism) (38) to reconstruct individual module-based GeRS (more detail in the Additional file 1: Supplementary Note ). Intriguingly, 2 of the 8 genes designated as hub genes using module 9 in these newly derived modules were also defined as hub genes in the originally identified early-onset-related module (i.e., module 10). Functional mapping of the hub genes The known functions and previously established genetic association of the six hub genes, including PSMB5 (39–42), PSMD2 (43, 44), TUBA4A (45–47), LRPPRC (48, 49), RUVBL2 (50, 51), and COPS6 (52, 53), are summarized in Table 2 . Briefly, they are mainly involved with cellular processes crucial for neurological function and implicated in various neurodevelopmental and neurodegenerative disorders, including proteasomal protein degradation, microtubule dynamics, mitochondrial function, chromatin remodeling, and protein degradation. Table 2 Background on the six hub genes identified from module 10, including brief descriptions of their genetic functions and previously established genetic associations. Gene Known genetic functions and relationships PSMB5 PSMB5 encodes the proteasome subunit beta type-5. The proteasome system is involved in regulating major cellular pathways, including those leading to inflammation (39), and has been identified as a canonical pathway associated with SZ (40, 41) and treatment resistance in major depressive disorder (42). PSMD2 PSMD2 , which encodes a non-ATPase subunit of the 26S proteasome, is crucial for regulating protein degradation and cellular processes (43). Recent in vivo studies show that neuronal membrane proteasomes regulate neuronal circuit activity, potentially involving PSMD2, and learning-induced behavioral plasticity, thereby highlighting the role of PSMD2 in neuronal function (44). TUBA4A TUBA4A encodes alpha tubulin 4A, a key component of microtubules regulating axonal transport in the central nervous system (45). TUBA4A demonstrates ubiquitous expression across all cell types, with its highest levels observed in the brain (46), and has been associated with amyotrophic lateral sclerosis and frontotemporal dementia (47). LRPPRC LRPPRC , which encodes leucine-rich pentatricopeptide repeat-containing protein, plays a crucial role in mitochondrial function and is involved in regulating mRNA stability and polyadenylation and coordinating mitochondrial translation (48). LRPPRC deficiency is linked with ATP synthase deficiency and severe neurodegenerative disorders such as Leigh syndrome, underscoring its critical role in brain function (49). RUVBL2 RUVBL2 encodes RuvBL2 (RuvB-like 2), which is part of various ATP-dependent chromatin remodeling complexes that are involved in the regulation of gene expression (50). RUVBL2 predominantly manifests in the SOX2-positive compartment of cerebral organoids and mid-gestation fetal brain tissue, with its chemical inactivation leading to precursor cell displacement and apoptosis and genetic variants linked to neurodevelopmental impairments (51). COPS6 COPS6 encodes the sixth subunit of the photomorphogenic 9 (COP9) signalsome (CSN), which is involved in the ubiquitin-mediated protein degradation process (52) and is further implicated in neurodegenerative diseases (53). We then performed gene-to-function analyses for those hub genes using the FUMA to understand which biological pathways or gene sets might be involved (Fig. 5 ). PSMB5 and COPS6 showed high expression from early prenatal stages through middle adulthood, while RUVBL2 and PSMD2 had elevated expression only during early prenatal stages, diminishing after infancy (Fig. 5 A). In contrast, TUBA4A exhibited increased expression after infancy. Also, several chemical and genetic perturbation gene sets showed significant enrichment (Fig. 5 B), including those related to psychiatric disorders duration, B lymphocytes regulatory network, and perineuronal oligodendrocytes in the BA9 brain region. Additionally, we utilized brain single-cell expression from psychENCODE, including samples with SZ, bipolar disorder, and controls, to better understand the expression patterns of these hub genes across various cell types ( Additional file 1: Figure S9 ). We found most hub genes were highly enriched in excitatory neurons compared to other cell types except LRPPRC . Discussion Despite antipsychotics being generally used to treat the symptoms of SZ, many individuals with early-onset SZ exhibit poor responses and resistance to treatment. The antipsychotics currently available were developed based on serendipitous clinical observations of the antipsychotic properties of drugs initially used for other indications (54). Identifying new drug targets by utilizing genomic information to understand the pathophysiology of early-onset schizophrenia is essential. Here, we applied the GeRS approach to early-onset SZ and identified six hub genes, including RUVBL2, COPS6, TUBA4A, PSMB5, PSMB2 , and LRPPRC . Functional mapping and enrichment analysis revealed potential biological mechanisms involved in early-onset SZ, e.g., expression changes in excitatory neurons and immune-regulatory processes. Our findings provide clues for downstream investigations into the etiology of early-onset SZ and may potentially be applied in early interventions. Several strategies were implemented to ensure the robustness of our findings, particularly 1) examining the module-based GeRS in both the multiplex and the simplex samples of patients with SZ and then aggregating both the multiplex and the simplex samples to conduct a meta-analysis to obtain estimates of summary effects, 2) adopting the SNP-expression prediction model derived from brain tissues of patients with psychiatric disorders to avoid potential bias resulting from nontrait-related tissues in different organs (13, 55), 3) and utilizing existing brain tissue gene expression data to map the identified genes to their biological context. Despite the differing familial loadings in the multiplex and the simplex samples of patients with SZ, the differences in neurocognitive performance and SZ-PRS between early-onset SZ and late-onset SZ remained and were in line with previous findings, e.g., early-onset SZ displaying more neurocognitive impairment (56, 57) and higher SZ-PRS (58, 59) than late-onset SZ. This, in part, justified our conduction of meta-analysis to estimate the summary effects of the identified genes. Although none of the individual SNPs in SZ-GeRS were significantly associated with early-onset SZ in this study, similar to previous studies (15, 60), we summarized genes from empirically derived gene clusters to obtain a functionally meaningful SZ-GeRS (25, 38). The modules derived from the postmortem brain tissues of patients with five psychiatric disorders (25) led to greater variance explained by the M10 hub -GeRS than that explained by the modules from patients with three psychiatric disorders and two specific brain regions (38). There are several possibilities, including the following: 1) many mental disorders of different diagnostic categories have shared genetic susceptibility; 2) preselection of brain regions may overlook contributions from other regions; 3) the gene list of module 10 has been found to overlap with a neuronal mitochondria gene-enriched module previously associated with fundamental neuronal processes and psychiatric disease (25); and 4) module 10 has been used to enrich the coexpression module profiling from the brain tissues of mouse models harboring SZ and autism-associated mutations, identifying its association with neuronal energetics and firing rate (61). Our findings support the postulation that GeRS has better portability across ancestry groups for predicting traits compared to conventional PRS (62, 63), based on the hypothesis that GeRS, derived from quantitative-trait locus variants potentially shared among different ancestral groups, is associated with shared biological mechanisms. Moreover, we observed that adding module-based SZ-GeRS to the model containing SZ-PRS and potential confounders would substantially increase r 2 and maintain the significant pattern of main effects in both the multiplex and the simplex samples. This implies that module-based SZ-GeRS and SZ-PRS each provided their respective contributions to early-onset SZ and that module-based SZ-GeRS might be a gene-based complement to SZ-PRS, potentially improving the ability to distinguish between early-onset and late-onset SZ. Our findings align with the fact that GeRS and PRS quantify independent genetic effects on complex traits (14, 62) and that integrating information from functional GeRS can improve risk prediction for these traits (63, 64). Based on the 56 genes of one subset of M10, M10 sub , this study utilized network analysis to narrow down to six hub genes that were predicted to highly regulate the expression of other genes within the coexpression network of module 10, and hence their aberrant expressions might impact downstream biological functions and lead to the onset of the illness (65). This possibility is further supported by subsequent explorations of the hub genes, from literature review to functional mapping of gene sets in developmental temporality and specific diseases or traits, and to single-cell gene expression profiles. These enrichment analyses provide further insight that these hub genes might have diverse expression patterns across developmental stages. Their coordination during brain development may reflect an atypical “transcriptional program” that deviates from normal development (66). This deviation could contribute to abnormal neural processes, potentially facilitating early-onset SZ. Additionally, several enriched gene sets link these genes to neurodevelopmental processes, such as the number of perineuronal oligodendrocytes in the BA9 region, and immune dysregulation, particularly in the B lymphocyte regulatory network. Reduction of perineuronal oligodendrocytes, as reported in a recent systematic review of postmortem brain studies (67), has been associated with major psychiatric disorders, and immune system dysfunction and inflammatory processes have long been associated with these disorders (68). The enriched gene set linked to the duration of psychiatric disorders might account in part for the clinical characteristics of early-onset SZ, i.e., being associated with more severe symptoms and treatment resistance, and hence resulting in a longer duration of illness. Moreover, most hub genes were enriched in excitatory neurons, indicating these cell types are likely central to the relationship between gene expression and psychiatric disorders. This support recent idea (69) that genetic risk variants affect genes with measurable expression changes in excitatory neurons of individuals with SZ. These hub genes have significant clinical relevance due to their dual roles as both susceptibility genes and modifier genes for SZ. While most antipsychotics are designed to target susceptibility genes, early-onset SZ involves modifier genes. This distinction may explain why patients with early-onset SZ often exhibit poor responses to existing antipsychotics, resulting in treatment resistance. Hence, our findings provide valuable insights for identifying potential drug targets and improving treatment for the earliest stages of this illness. Our study has several limitations. First, as this study was based on a case-only design, we potentially identified only the 'mixed genes' for SZ, which have dual functions as susceptibility genes and modifier genes (23, 70), but overlooked purely modifier genes for early-onset SZ. Second, the small sample size of this study might render individual SNPs included in M10 sub -GeRS and M10 hub -GeRS not being able to reach transcriptome-wide significance. Nevertheless, the combination of these genes was able to show strong associations with early-onset SZ. Last, the aberrant gene expression, as incorporated in the module-derived GeRS, was based on prediction; future in vitro and animal studies are needed to elucidate their roles in the biological mechanisms of early-onset SZ. In conclusion, we found that a module-based GeRS approach that incorporates predicted gene expression resulting from the interaction of genes and environmental factors may help capture further contributions to early-onset SZ. Pathway and enrichment analyses revealed that the identified hub genes might be involved in deviation from typical brain development and expression changes in excitatory neurons. Our findings will enhance the understanding of SZ etiology and may contribute to early screening and personalized prevention efforts. Declarations Acknowledgements and Disclosures This study was supported by grants from the Ministry of Science and Technology, Taiwan (107-2314-B-002-214-MY3; 109-2314-B-002-172-MY3); the National Health Research Institutes, Taiwan (NP-109-SP-07, NP-110-SP-02, and NP-111-SP-02); the NIH/NHGRI grant U54HG003067; NIMH grants R01 MH085521 and R01 MH085560; the Gerber Foundation; the Sidney R. Baer, Jr. Foundation; NARSAD: The Brain and Behavior Research Foundation; and the Stanley Center for Psychiatric Research. Dr. Neale reports serving as a member of the Scientific Advisory Board of Deep Genomics and as a consultant for Camp4 Therapeutics Corporation, Merck & Co., Takeda Pharmaceutical, and Avanir Pharmaceuticals, Inc. All the other authors report no biomedical financial interests or potential conflicts of interest. Competing Interests Dr. Neale reports serving as a member of the Scientific Advisory Board of Deep Genomics and as a consultant for Camp4 Therapeutics Corporation, Merck & Co., Takeda Pharmaceutical, and Avanir Pharmaceuticals, Inc. All the other authors report no biomedical financial interests or potential conflicts of interest. Author Contribution W.J.C. was responsible for the conception, design of the study, and funding acquisition. Y.J. performed genotype data preprocessing, developed analytical schemes, and conducted statistical analyses. S.-L.Y. provided technical and material support for DNA experiments. P.-C.H. provided technical support. C.-M.L., C.-C.L., T.-J.H., M.H.H., Y.-L.C., Y.-T.L., H.-G.H., and W.J.C. managed participant recruitment and data acquisition. H.H., Y.A.F., P.-H.K., C.K.H., and Y.-F.L. contributed to the interpretation of the data. S.V.F., B.N., S.J.G., and M.T.T. were responsible for conceptualization, resources, and supervision. Y.J. and W.J.C. implemented the literature review and data interpretation. The first draft was written by Y.J., and W.J.C. provided critical revisions and supervised the manuscript. All authors reviewed, commented on, and approved the final version of the manuscript. Data Availability The datasets used and analyzed in the current study are not publicly available due to conditions in the participant consent and other ethical restrictions. However, the data that support the findings of this study are available from the corresponding author upon reasonable request. References GBD 2016 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017; 390: 1211–1259. Moreno-Küstner B, Martín C, Pastor L. Prevalence of psychotic disorders and its association with methodological issues. A systematic review and meta-analyses. PLoS ONE 2018; 13: e0195687. Sullivan PF, Kendler KS, Neale MC. Schizophrenia as a complex trait: evidence from a meta-analysis of twin studies. Arch Gen Psychiatry 2003; 60: 1187–1192. Hilker R, Helenius D, Fagerlund B, Skytthe A, Christensen K, Werge TM, et al. Heritability of schizophrenia and schizophrenia spectrum based on the nationwide Danish Twin Register. Biol Psychiatry 2018; 83: 492–498. Schizophrenia Working Group of the Psychiatric Genomics Consortium. Biological insights from 108 schizophrenia-associated genetic loci. Nature 2014; 511: 421–427. Lam M, Chen C-Y, Li Z, Martin AR, Bryois J, Ma X, et al. Comparative genetic architectures of schizophrenia in East Asian and European populations. Nat Genet 2019; 51: 1670–1678. Trubetskoy V, Pardiñas AF, Qi T, Panagiotaropoulou G, Awasthi S, Bigdeli TB, et al. Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature 2022; 604: 502–508. Owen MJ, Sawa A, Mortensen PB. Schizophrenia. Lancet 2016; 388: 86–97. Jaffe AE. Postmortem human brain genomics in neuropsychiatric disorders—how far can we go? Curr Opin Neurobiol 2016; 36: 107–111. Zhu Y, Wang L, Yin Y, Yang E. Systematic analysis of gene expression patterns associated with postmortem interval in human tissues. Sci Rep 2017; 7: 5435. Gamazon ER, Wheeler HE, Shah KP, Mozaffari SV, Aquino-Michaels K, Carroll RJ, et al. A gene-based association method for mapping traits using reference transcriptome data. Nat Genet 2015; 47: 1091–1098. Huckins LM, Dobbyn A, Ruderfer DM, Hoffman G, Wang W, Pardiñas AF, et al. Gene expression imputation across multiple brain regions provides insights into schizophrenia risk. Nat Genet 2019; 51: 659–674. Gusev A, Mancuso N, Won H, Kousi M, Finucane HK, Reshef Y, et al. Transcriptome-wide association study of schizophrenia and chromatin activity yields mechanistic disease insights. Nat Genet 2018; 50: 538–548. Rodriguez-López J, Arrojo M, Paz E, Páramo M, Costas J. Identification of relevant hub genes for early intervention at gene coexpression modules with altered predicted expression in schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry 2020; 98: 109815. Prohens L, Rodríguez N, Segura À-G, Martínez-Pinteño A, Olivares-Berjaga D, Martínez I, et al. Gene expression imputation provides clinical and biological insights into treatment-resistant schizophrenia polygenic risk. Psychiatry Res 2024; 332: 115722. Solmi M, Radua J, Olivola M, Croce E, Soardo L, Salazar de Pablo G, et al. Age at onset of mental disorders worldwide: large-scale meta-analysis of 192 epidemiological studies. Mol Psychiatry 2022; 27: 281–295. Coulon N, Godin O, Bulzacka E, Dubertret C, Mallet J, Fond G, et al. Early and very early-onset schizophrenia compared with adult-onset schizophrenia: French FACE-SZ database. Brain Behav 2020; 10: e01495. Chen WJ. Taiwan Schizophrenia Linkage Study: Lessons learned from endophenotype-based genome-wide linkage scans and perspective. Am J Med Genet B Neuropsychiatr Genet 2013; 162B: 636–647. Birnbaum R, Weinberger DR. Genetic insights into the neurodevelopmental origins of schizophrenia. Nat Rev Neurosci 2017; 18: 727–740. Iasevoli F, Razzino E, Altavilla B, Avagliano C, Barone A, Ciccarelli M, et al. Relationships between early age at onset of psychotic symptoms and treatment resistant schizophrenia. Early Interv Psychiatry 2022; 16: 352–362. Bergen SE, O'Dushlaine CT, Lee PH, Fanous AH, Ruderfer DM, Ripke S, et al. Genetic modifiers and subtypes in schizophrenia: Investigations of age at onset, severity, sex and family history. Schizophr Res 2014; 154: 48–53. Woolston AL, Hsiao P-C, Kuo P-H, Wang S-H, Lien Y-J, Liu C-M, et al. Genetic loci associated with an earlier age at onset in multiplex schizophrenia. Sci Rep 2017; 7: 6486. Zhan N, Sham PC, So HC, Lui SSY. The genetic basis of onset age in schizophrenia: evidence and models. Front Genet 2023; 14: 1163361. Barbeira AN, Dickinson SP, Bonazzola R, Zheng J, Wheeler HE, Torres JM, et al. Exploring the phenotypic consequences of tissue specific gene expression variation inferred from GWAS summary statistics. Nat Commun 2018; 9: 1825. Gandal MJ, Haney JR, Parikshak NN, Leppa V, Ramaswami G, Hartl C, et al. Shared molecular neuropathology across major psychiatric disorders parallels polygenic overlap. Science 2018; 359: 693–697. Hwu H-G, Faraone SV, Liu C-M, Chen WJ, Liu S-K, Shieh M-H, et al. Taiwan Schizophrenia Linkage Study: The field study. Am J Med Genet B Neuropsychiatr Genet 2005; 134B: 30–36. Faraone SV, Hwu H-G, Liu C-M, Chen WJ, Tsuang M-M, Liu S-K, et al. Genome scan of Han Chinese schizophrenia families from Taiwan: confirmation of linkage to 10q22.3. Am J Psychiatry 2006; 163: 1760–1766. Nurnberger JI, Jr., Blehar MC, Kaufmann CA, York-Cooler C, Simpson SG, Harkavy-Friedman J, et al. Diagnostic Interview for Genetic Studies: Rationale, unique features, and training. Arch Gen Psychiatry 1994; 51: 849–859. Rees E, Kirov G, Walters JT, Richards AL, Howrigan D, Kavanagh DH, et al. Analysis of exome sequence in 604 trios for recessive genotypes in schizophrenia. Transl Psychiatry 2015; 5: e607. Wang SH, Hsiao PC, Yeh LL, Liu CM, Liu CC, Hwang TJ, et al. Advanced paternal age and early-onset of schizophrenia in sporadic cases: not confounded by parental polygenic risk to schizophrenia. Biol Psychiatry 2019; 86: 56–64. Chen WJ, Hsiao CK, Hsiao LL, Hwu HG. Performance of the Continuous Performance Test among community samples. Schizophr Bull 1998; 24: 163–174. Lin S-H, Liu C-M, Hwang T-J, Hsieh MH, Hsiao P-C, Faraone SV, et al. Performance on the Wisconsin Card Sorting Test in families of schizophrenia patients with different familial loadings. Schizophr Bull 2013; 39: 537–546. Peterson RE, Kuchenbaecker K, Walters RK, Chen CY, Popejoy AB, Periyasamy S, et al. Genome-wide association studies in ancestrally diverse populations: Opportunities, methods, pitfalls, and recommendations. Cell 2019; 179: 589–603. Liberzon A, Birger C, Thorvaldsdóttir H, Ghandi M, Mesirov JP, Tamayo P. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 2015; 1: 417–425. Nagelkerke NJD. A note on a general definition of the coefficient of determination. Biometrika 1991; 78: 691–692. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw 2010; 36: 1–48. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, et al. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575. Gandal MJ, Zhang P, Hadjimichael E, Walker RL, Chen C, Liu S, et al. Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science 2018; 362: eaat8127. Ebstein F, Kloetzel PM, Krüger E, Seifert U. Emerging roles of immunoproteasomes beyond MHC class I antigen processing. Cell Mol Life Sci 2012; 69: 2543–2558. Bousman CA, Chana G, Glatt SJ, Chandler SD, Lucero GR, Tatro E, et al. Preliminary evidence of ubiquitin proteasome system dysregulation in schizophrenia and bipolar disorder: convergent pathway analysis findings from two independent samples. Am J Med Genet B Neuropsychiatr Genet 2010; 153B: 494–502. Lee YH, Kim JH, Song GG. Pathway analysis of a genome-wide association study in schizophrenia. Gene 2013; 525: 107–115. Minelli A, Magri C, Barbon A, Bonvicini C, Segala M, Congiu C, et al. Proteasome system dysregulation and treatment resistance mechanisms in major depressive disorder. Transl Psychiatry 2015; 5: e687. Ying K, Wang C, Liu S, Kuang Y, Tao Q, Hu X. Diverse Ras-related GTPase DIRAS2, downregulated by PSMD2 in a proteasome-mediated way, inhibits colorectal cancer proliferation by blocking NF-κB signaling. Int J Biol Sci 2022; 18: 1039–1050. He HY, Ahsan A, Bera R, McLain N, Faulkner R, Ramachandran KV, et al. Neuronal membrane proteasomes regulate neuronal circuit activity in vivo and are required for learning-induced behavioral plasticity. Proc Natl Acad Sci USA 2023; 120: e2216537120. Breuss MW, Leca I, Gstrein T, Hansen AH, Keays DA. Tubulins and brain development; The origins of functional specification. Mol Cell Neurosci 2017; 84: 58–67. Rustici G, Kolesnikov N, Brandizi M, Burdett T, Dylag M, Emam I, et al. ArrayExpress update–trends in database growth and links to data analysis tools. Nucleic Acids Res 2013; 41: D987-990. Van Schoor E, Vandenbulcke M, Bercier V, Vandenberghe R, van der Zee J, Van Broeckhoven C, et al. Frontotemporal lobar degeneration case with an N-terminal TUBA4A mutation exhibits reduced TUBA4A levels in the brain and TDP-43 pathology. Biomolecules 2022; 12: 440. Mourier A, Ruzzenente B, Brandt T, Kühlbrandt W, Larsson NG. Loss of LRPPRC causes ATP synthase deficiency. Hum Mol Genet 2014; 23: 2580–2592. Cuillerier A, Honarmand S, Cadete VJJ, Ruiz M, Forest A, Deschênes S, et al. Loss of hepatic LRPPRC alters mitochondrial bioenergetics, regulation of permeability transition and trans-membrane ROS diffusion. Hum Mol Genet 2017; 26: 3186–3201. Zhang G, Li S, Cheng KW, Chou TF. AAA ATPases as therapeutic targets: Structure, functions, and small-molecule inhibitors. Eur J Med Chem 2021; 219: 113446. Melliou S, Sangster KT, Kao J, Zarrei M, Lam KHB, Howe J, et al. Regionally defined proteomic profiles of human cerebral tissue and organoids reveal conserved molecular modules of neurodevelopment. Cell Rep 2022; 39: 110846. Wei N, Deng XW. The COP9 signalosome. Annu Rev Cell Dev Biol 2003; 19: 261–286. Zheng Q, Huang T, Zhang L, Zhou Y, Luo H, Xu H, et al. Dysregulation of ubiquitin-proteasome system in neurodegenerative diseases. Front Aging Neurosci 2016; 8: 303. Lago SG, Bahn S. The druggable schizophrenia genome: from repurposing opportunities to unexplored drug targets. npj Genom Med 2022; 7: 25. Wainberg M, Sinnott-Armstrong N, Mancuso N, Barbeira AN, Knowles DA, Golan D, et al. Opportunities and challenges for transcriptome-wide association studies. Nat Genet 2019; 51: 592–599. Lien Y-J, Hsiao P-C, Liu C-M, Faraone SV, Tsuang MT, Hwu H-G, et al. A genome-wide linkage scan for distinct subsets of schizophrenia characterized by age at onset and neurocognitive deficits. PLoS ONE 2011; 6: e24103. Bortolato B, Miskowiak KW, Köhler CA, Vieta E, Carvalho AF. Cognitive dysfunction in bipolar disorder and schizophrenia: a systematic review of meta-analyses. Neuropsychiatr Dis Treat 2015; 11: 3111–3125. Deng M, Liu Z, Zhang W, Wu Z, Cao H, Yang J, et al. Associations between polygenic risk, negative symptoms, and functional connectome topology during a working memory task in early-onset schizophrenia. Schizophrenia 2022; 8: 54. Sada-Fuente E, Aranda S, Papiol S, Heilbronner U, Moltó MD, Aguilar EJ, et al. Common genetic variants contribute to heritability of age at onset of schizophrenia. Transl Psychiatry 2023; 13: 201. Fabbri C, Leggio GM, Drago F, Serretti A. Imputed expression of schizophrenia-associated genes and cognitive measures in patients with schizophrenia. Mol Genet Genomic Med 2022; 10: e1942. Gordon A, Forsingdal A, Klewe IV, Nielsen J, Didriksen M, Werge T, et al. Transcriptomic networks implicate neuronal energetic abnormalities in three mouse models harboring autism and schizophrenia-associated mutations. Mol Psychiatry 2021; 26: 1520–1534. Hu X, Qiao D, Kim W, Moll M, Balte PP, Lange LA, et al. Polygenic transcriptome risk scores for COPD and lung function improve cross-ethnic portability of prediction in the NHLBI TOPMed program. Am J Hum Genet 2022; 109: 857–870. Liang Y, Pividori M, Manichaikul A, Palmer AA, Cox NJ, Wheeler HE, et al. Polygenic transcriptome risk scores (PTRS) can improve portability of polygenic risk scores across ancestries. Genome Biol 2022; 23: 23. Pain O, Glanville KP, Hagenaars S, Selzam S, Fürtjes A, Coleman JRI, et al. Imputed gene expression risk scores: a functionally informed component of polygenic risk. Hum Mol Genet 2021; 30: 727–738. He T, Chan KCC. Measuring boundedness for protein complex identification in PPI networks. IEEE/ACM Trans Comput Biol Bioinform 2019; 16: 967–979. Dear R, Wagstyl K, Seidlitz J, Markello RD, Arnatkevičiūtė A, Anderson KM, et al. Cortical gene expression architecture links healthy neurodevelopment to the imaging, transcriptomics and genetics of autism and schizophrenia. Nat Neurosci 2024; 27: 1075–1086. Liu SH, Du Y, Chen L, Cheng Y. Glial cell abnormalities in major psychiatric diseases: A systematic review of postmortem brain studies. Mol Neurobiol 2022; 59: 1665–1692. Chen S, Tan Y, Tian L. Immunophenotypes in psychosis: is it a premature inflamm-aging disorder? Mol Psychiatry 2024; 29: 2834–2848. Ruzicka WB, Mohammadi S, Fullard JF, Davila-Velderrain J, Subburaju S, Tso DR, et al. Single-cell multi-cohort dissection of the schizophrenia transcriptome. Science 2024; 384: eadg5136. Fanous AH, Kendler KS. Genetic heterogeneity, modifier genes, and quantitative phenotypes in psychiatric illness: searching for a framework. Mol Psychiatry 2005; 10: 6–13. Additional Declarations Competing interest reported. Dr. Neale reports serving as a member of the Scientific Advisory Board of Deep Genomics and as a consultant for Camp4 Therapeutics Corporation, Merck & Co., Takeda Pharmaceutical, and Avanir Pharmaceuticals, Inc. All the other authors report no biomedical financial interests or potential conflicts of interest. Supplementary Files Additionalfile1JenetalSupplementaryinformationGMclean.docx Additionalfile2JenetalSupplementaryDataGMclean.xlsx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5833160","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":402444515,"identity":"80d1a728-7ad9-4c05-bea5-10e3697f23e2","order_by":0,"name":"Yawen Jen","email":"","orcid":"","institution":"Center for Neuropsychiatric Research, National Health Research Institutes","correspondingAuthor":false,"prefix":"","firstName":"Yawen","middleName":"","lastName":"Jen","suffix":""},{"id":402444516,"identity":"c3531ae8-1ef4-4917-afc3-7dd2ecfd5b19","order_by":1,"name":"Sung-Liang Yu","email":"","orcid":"","institution":"Department of Clinical Laboratory Sciences and Medical Biotechnology, College of Medicine, National Taiwan University","correspondingAuthor":false,"prefix":"","firstName":"Sung-Liang","middleName":"","lastName":"Yu","suffix":""},{"id":402444517,"identity":"dbc28972-b587-4f0f-8a4c-ef0cbf0393b3","order_by":2,"name":"Po-Chang Hsiao","email":"","orcid":"","institution":"Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University","correspondingAuthor":false,"prefix":"","firstName":"Po-Chang","middleName":"","lastName":"Hsiao","suffix":""},{"id":402444518,"identity":"9bac5c05-04e2-4a4f-82fe-d6ce6c2f621a","order_by":3,"name":"Po-Hsiu Kuo","email":"","orcid":"","institution":"Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University","correspondingAuthor":false,"prefix":"","firstName":"Po-Hsiu","middleName":"","lastName":"Kuo","suffix":""},{"id":402444519,"identity":"39b43a94-de70-4076-bf6f-b889317538b3","order_by":4,"name":"Chih-Min Liu","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chih-Min","middleName":"","lastName":"Liu","suffix":""},{"id":402444520,"identity":"ac66f04f-32b5-4003-a336-be9e1f85c725","order_by":5,"name":"Chen-Chung Liu","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chen-Chung","middleName":"","lastName":"Liu","suffix":""},{"id":402444521,"identity":"5d3df063-23b8-48c9-ba02-b7fed0b5fdd6","order_by":6,"name":"Tzung-Jeng Hwang","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Tzung-Jeng","middleName":"","lastName":"Hwang","suffix":""},{"id":402444522,"identity":"10342637-b9c8-48c8-81f9-30e101f47ff4","order_by":7,"name":"Ming H. Hsieh","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ming","middleName":"H.","lastName":"Hsieh","suffix":""},{"id":402444523,"identity":"7c592e0a-c02c-440c-90f1-7be875d54289","order_by":8,"name":"Yi-Ling Chien","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yi-Ling","middleName":"","lastName":"Chien","suffix":""},{"id":402444524,"identity":"46af9374-286f-4760-9405-4d26146dd559","order_by":9,"name":"Yi-Ting Lin","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yi-Ting","middleName":"","lastName":"Lin","suffix":""},{"id":402444525,"identity":"fa36fb21-62ac-44c7-8dd7-bfb74b18d113","order_by":10,"name":"Hailiang Huang","email":"","orcid":"","institution":"Stanley Center for Psychiatric Research, Broad Institute of MIT and Harvard","correspondingAuthor":false,"prefix":"","firstName":"Hailiang","middleName":"","lastName":"Huang","suffix":""},{"id":402444526,"identity":"9936c4bc-c77c-49fa-a8b9-f3c18b34ffd7","order_by":11,"name":"Yen-Chen Anne Feng","email":"","orcid":"","institution":"Institute of Health Data Analytics and Statistics, College of Public Health, National Taiwan University","correspondingAuthor":false,"prefix":"","firstName":"Yen-Chen","middleName":"Anne","lastName":"Feng","suffix":""},{"id":402444527,"identity":"499c1d77-23d3-4fae-a2d4-a42002ca2aae","order_by":12,"name":"Chuhsing K. Hsiao","email":"","orcid":"","institution":"Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University","correspondingAuthor":false,"prefix":"","firstName":"Chuhsing","middleName":"K.","lastName":"Hsiao","suffix":""},{"id":402444528,"identity":"8f698786-0bea-43d1-8d86-2ec83bd77b79","order_by":13,"name":"Yen-Feng Lin","email":"","orcid":"","institution":"Center for Neuropsychiatric Research, National Health Research Institutes","correspondingAuthor":false,"prefix":"","firstName":"Yen-Feng","middleName":"","lastName":"Lin","suffix":""},{"id":402444529,"identity":"957e6ca1-1f16-4401-a0f7-dcb8f14ba097","order_by":14,"name":"Stephen V. Faraone","email":"","orcid":"","institution":"Departments of Psychiatry and Behavioral Sciences, Neuroscience and Physiology, and Public Health and Preventive Medicine, SUNY Upstate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"V.","lastName":"Faraone","suffix":""},{"id":402444530,"identity":"11581757-b26f-4e49-a5fe-bee5fb60df6e","order_by":15,"name":"Benjamin Neale","email":"","orcid":"","institution":"Stanley Center for Psychiatric Research, Broad Institute of MIT and Harvard","correspondingAuthor":false,"prefix":"","firstName":"Benjamin","middleName":"","lastName":"Neale","suffix":""},{"id":402444531,"identity":"94487588-e9c4-4547-8b71-a9ae5b68c0a0","order_by":16,"name":"Stephen J. Glatt","email":"","orcid":"","institution":"Departments of Psychiatry and Behavioral Sciences, Neuroscience and Physiology, and Public Health and Preventive Medicine, SUNY Upstate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"J.","lastName":"Glatt","suffix":""},{"id":402444532,"identity":"24f3dd94-60fa-4499-a58a-615e7177ab45","order_by":17,"name":"Ming T. Tsuang","email":"","orcid":"","institution":"Center for Behavioral Genomics, Department of Psychiatry, University of California San Diego","correspondingAuthor":false,"prefix":"","firstName":"Ming","middleName":"T.","lastName":"Tsuang","suffix":""},{"id":402444533,"identity":"13129e6b-d4d4-4c56-89e6-0f6bec2187b0","order_by":18,"name":"Hai-Gwo Hwu","email":"","orcid":"","institution":"National Taiwan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hai-Gwo","middleName":"","lastName":"Hwu","suffix":""},{"id":402444534,"identity":"f4e4b577-d66c-48d9-b649-e08de2dcc25d","order_by":19,"name":"Wei J. Chen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxUlEQVRIiWNgGAWjYFCCA2xAwobBgIGxgYGBjXgtaQgtPEToAWk5DNQCZRPUYnDw8LMHH3eczzOXPtzA8KHsMIO9RAIBLQeOmRvOPHO72LIvsYFxxrnDDDyEtJgdOMMmzdt2O3HDGcYGZt42oBZpYrT8bTsH0fKXaC2MbQcgWhiJ0WJ/4JiZZG9bMljLwZ5z6Tw89x/g1yI54/AziZ9tdkAt7A8f/CizlmPvOYBfC4MEkgIQk4iY5G8grGYUjIJRMApGOAAAeAdIX+tUg7oAAAAASUVORK5CYII=","orcid":"","institution":"Center for Neuropsychiatric Research, National Health Research Institutes","correspondingAuthor":true,"prefix":"","firstName":"Wei","middleName":"J.","lastName":"Chen","suffix":""}],"badges":[],"createdAt":"2025-01-15 09:38:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5833160/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5833160/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":74282903,"identity":"de974d24-d563-409b-8a93-cc5b4fc9accb","added_by":"auto","created_at":"2025-01-20 15:52:32","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":373720,"visible":true,"origin":"","legend":"\u003cp\u003eProportion of variance in early-onset SZ explained by SZ-PRSs at various p value thresholds (x-axis), with the y-axis showing the change in Nagelkerke pseudo-R² for the multiplex (A) and simplex (B) samples. Changes in R² were estimated by comparing the model fit (covariates plus SZ-PRS) vs. the null model (only covariates). The mixed-effects logistic regression model, with family index as a random effect, is used in the multiplex sample, and standard logistic regression is used in the simplex sample. The significance of SZ-PRS is set at p \u0026lt; 0.0083 (0.05/6, Bonferroni correction). Density plots showing the best-performing SZ-PRS distributions for early-onset SZ (red), late-onset SZ (blue), and their unaffected relatives (green) in the multiplex (C) and simplex (D) samples. The SZ-PRS follows a normal distribution, scaled to a mean of 0 and a standard deviation of 1.\u003c/p\u003e","description":"","filename":"F1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/818f20e3ea4544822f38ce7e.jpg"},{"id":74282917,"identity":"dc983e02-7f92-40c1-b86b-5fd6df3b48a9","added_by":"auto","created_at":"2025-01-20 15:52:33","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":397415,"visible":true,"origin":"","legend":"\u003cp\u003ePredictive utility of individual module-based GeRS, derived from WGCNA gene lists based on cerebral cortex expression, for early-onset SZ in the multiplex sample (A). The X-axis shows modules 1 to 13, and the Y-axis indicates the change in Nagelkerke pseudo-R² by comparing the model fit (covariates plus SZ-PRS) to the null model (only covariates). A mixed-effects logistic regression with family index as a random effect was used. Violin plots of M10\u003csub\u003esub\u003c/sub\u003e-GeRS according to onset status for the multiplex (B) and simplex (C) samples. The M10\u003csub\u003esub\u003c/sub\u003e-GeRS were based on a subset of module 10 consisting of 56 genes predicted to have genetically regulated expression in DLPFC, identified in both the multiplex and simplex samples. The M10\u003csub\u003esub\u003c/sub\u003e-GeRS follows a normal distribution, scaled to a mean of 0 and a standard deviation of 1.\u003c/p\u003e","description":"","filename":"F2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/a6942b717f39b7369c16f8f6.jpg"},{"id":74282901,"identity":"e77ea47e-94f9-4e20-b205-7fedbf8ed133","added_by":"auto","created_at":"2025-01-20 15:52:32","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":274045,"visible":true,"origin":"","legend":"\u003cp\u003eThe 56 genes of M10\u003csub\u003esub\u003c/sub\u003e-GeRS listed in module-10 forming the top gene network, as identified using Ingenuity Pathway Analysis (IPA). Gray shading: genes from module 10 subjected to the network analysis. Straight lines represent direct relationships between genes, whereas broken lines signify indirect relationships. Hub genes are colored in red (i.e., genes in regulatory relationships with at least 4 other genes).\u003c/p\u003e","description":"","filename":"F3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/333bd5cd8e8b804451d52ff5.jpg"},{"id":74282909,"identity":"2a452a10-79d2-47ea-8d59-087917bccaec","added_by":"auto","created_at":"2025-01-20 15:52:33","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":345971,"visible":true,"origin":"","legend":"\u003cp\u003eUtility of M10-GeRS and SZ-PRS for distinguishing early- from late-onset SZ was assessed in the multiplex (A) and simplex (B) samples, with summary effects (C) obtained via meta-analysis using a random-effects model. Risk scores were initially tested individually for their associations with early-onset SZ (Models 0 to 3). Advanced models incorporating gene summary GeRS and SZ-PRS data were tested in Models 3 and 4. SZ-PRS: the best-performing SZ-PRS in each sample; M10\u003csub\u003esub\u003c/sub\u003e-GeRS: the gene summary GeRS based on a subset of module 10 consisting of 56 genes predicted to have genetically regulated expression in DLPFC, identified in both the multiplex and simplex samples; M10\u003csub\u003ehub\u003c/sub\u003e-GeRS: the gene summary GeRS based on 6 hub genes identified by the IPA top network; aOR: adjusted odds ratio from mixed-effect logistic regression with family index as a random effect in the multiplex sample and logistic regression in the simplex sample. Models adjusted for sex, education, and 4 PCs; Pseudo-R\u003csup\u003e2\u003c/sup\u003e: increase in R² estimated by comparing the model fit to the null model (only covariates).\u003c/p\u003e","description":"","filename":"F4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/515bcae338b05265f1ad3f8d.jpg"},{"id":74282908,"identity":"f6e66dd9-0ddc-4429-959c-4566f81461e3","added_by":"auto","created_at":"2025-01-20 15:52:32","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":384864,"visible":true,"origin":"","legend":"\u003cp\u003eEnrichment analyses of hub genes conducted by FUMA. Gene expression heatmap for hub genes across the 11 general developmental stages of brain samples from BrainSpan, with genes ordered by expression level (A). The average expression level (log2 transformed expression value) of the genes is between 0 (blue) and 5.67 (red). Enrichment of hub genes in gene sets in FUMA using molecular signatures database (B). The -log10 adjusted P-value indicates the statistical significance of enrichment after applying the Benjamini-Hochberg (BH) correction for multiple testing. The BH method controls the false discovery rate, reducing the likelihood of false positives arising from testing multiple gene sets for enrichment.\u003c/p\u003e","description":"","filename":"F5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/860f7685c01a8f443475fcd3.jpg"},{"id":74438701,"identity":"faba2138-20e3-4587-830f-469f721dc731","added_by":"auto","created_at":"2025-01-22 09:47:29","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2928317,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/24e349cc-c147-4c69-9f09-5c5ca730019d.pdf"},{"id":74282916,"identity":"b0755d5d-cd89-4cc2-9f60-eb139f219041","added_by":"auto","created_at":"2025-01-20 15:52:33","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":3514329,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile1JenetalSupplementaryinformationGMclean.docx","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/8c5399c489c12ff11d466ec8.docx"},{"id":74282905,"identity":"524a49f8-6dca-4b38-9d06-a5bb54659111","added_by":"auto","created_at":"2025-01-20 15:52:32","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1195675,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile2JenetalSupplementaryDataGMclean.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5833160/v1/4c735618d8db2abcad97c3dd.xlsx"}],"financialInterests":"Competing interest reported. Dr. Neale reports serving as a member of the Scientific Advisory Board of Deep Genomics and as a consultant for Camp4 Therapeutics Corporation, Merck \u0026 Co., Takeda Pharmaceutical, and Avanir Pharmaceuticals, Inc. All the other authors report no biomedical financial interests or potential conflicts of interest.","formattedTitle":"Identification of Hub Genes Involved in Early-onset Schizophrenia: From Genetic Susceptibility to Predicted Regulated Gene Expression","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSchizophrenia (SZ)\u0026nbsp;is\u0026nbsp;a leading\u0026nbsp;cause of disability worldwide\u0026nbsp;(1), with a median lifetime\u0026nbsp;prevalence of 6.35 per 1000 people\u0026nbsp;(2).\u0026nbsp;Despite\u0026nbsp;a heritability of 80%\u0026nbsp;(3, 4),\u0026nbsp;only limited variance in SZ was attributed to gene loci or the polygenic risk score (PRS) in genome-wide association studies (GWASs)\u0026nbsp;(5-7).\u0026nbsp;SZ\u0026nbsp;is a heterogeneous disorder characterized by psychotic symptoms (delusions and hallucinations) and cognitive impairments (8). Given that a substantial proportion of the heritability of SZ remains unexplained, varying levels of gene expression resulting from the interaction of genes and environmental factors are postulated to further contribute to the occurrence of SZ, although collecting \u003cem\u003epostmortem\u003c/em\u003e brain tissues for gene expression analysis is challenging\u0026nbsp;(9, 10).\u0026nbsp;Instead, predicting genetically regulated gene expression in the brain by integrating an existing tissue-specific single-nucleotide polymorphism (SNP)-expression prediction model with GWAS-based genotype data,\u003cem\u003e\u0026nbsp;\u003c/em\u003ei.e., transcriptome imputation, is an alternative approach\u0026nbsp;(11, 12). Studies applying the genetically regulated gene expression risk score (GeRS) for SZ have revealed their associations with SZ or SZ-related traits, including clinical symptomatology and patients\u0026rsquo; cognitive performance\u0026nbsp;(12-15). However,\u0026nbsp;only a few genes with potential biological mechanisms that could aid in clinical implementation have been identified in these studies.\u003c/p\u003e\n\u003cp\u003eEarly-onset SZ, occurring in about 8.2% of cases diagnosed before age 18 (16), is a distinct and homogeneous subtype of SZ that serves as a critical phenotype for dissecting its genetic etiology. This subtype is associated with more severe symptoms (17), a greater genetic burden (18, 19), and a higher risk of treatment resistance (20) compared to adult-onset SZ. Findings from GWAS for SZ have identified genetic loci of suggestive significance in early-onset SZ (21, 22), highlighting its potential as a phenotype for uncovering genes relevant to early intervention. However, the genetic basis of early-onset SZ remains largely unknown (23). To elucidate the genetic underpinnings of early-onset SZ, we aimed to examine the gene expression risk score (GeRS) in patients from both multiplex and simplex families to identify hub genes for early-onset SZ, and perform enrichment analysis to understand the biological functions of the hub genes.\u003c/p\u003e"},{"header":"Methods And Materials","content":"\u003cp\u003e\u003cstrong\u003eStudy workflow\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study workflow is shown in \u003cstrong\u003eAdditional file 1\u003c/strong\u003e\u003cstrong\u003e: Fig\u003c/strong\u003e\u003cstrong\u003eure\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eS1\u003c/strong\u003e. We first predicted genetically regulated gene expression levels in the dorsolateral prefrontal cortex (DLPFC) using the PrediXcan (https://github.com/hakyimlab/PrediXcan), based on the GWAS genotype data from patients with SZ in multiplex families. The SNP-expression prediction model was based on the CommonMind Consortium (CMC; including 254 SZ, 52 bipolar disorder, and 279 controls) (https://predictdb.org/) (12). To extract disease-associated gene expression changes, we then performed a transcriptome-wide association study (TWAS) applying S-PrediXcan (24) with the same SNP-expression prediction model and the latest GWAS summary statistics from the SZ cohort of the Psychiatric Genomics Consortium (PGC; 76,755 patients with SZ and 243,648 controls) (7). After that, we calculated GeRS for each gene by weighting its predicted expression based on the corresponding effect size from the TWAS for SZ. Furthermore, we constructed the module-based GeRS by summing the GeRS of genes listed in the modules obtained from the published weighted gene coexpression network analysis (WGCNA) results (participants including 150 SZ, 94 bipolar disorder, 87 major depression, 50 autism, 17 alcoholism, and 293 controls) (25). Finally, we evaluated the relationship between the module-based GeRS and early-onset SZ in the multiplex sample and explored the hub genes within the early-onset-related module using Ingenuity Pathway Analysis (IPA\u0026reg;, QIAGEN Redwood City, CA, USA; www.qiagen.com/ingenuity). We then repeated the analysis in the simplex sample of patients with SZ from trio families.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eParticipants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study included two samples: a multiplex sample and a simplex sample. The multiplex sample was derived from the Taiwan Schizophrenia Linkage Study (TSLS), which recruited Han Chinese families with more than two siblings affected by SZ and their first-degree relatives across Taiwan from 1998\u0026ndash;2002 (26, 27) (see details in the\u003cstrong\u003e\u0026nbsp;Additional file 1\u003c/strong\u003e). After quality control, 595 patients and 570 unaffected relatives from 314 families, with age-at-onset information for the first psychotic episode obtained via the Diagnostic Interview for Genetic Studies (DIGS)\u0026nbsp;(28), were included.\u003c/p\u003e\n\u003cp\u003eTo examine the findings derived from the multiplex sample in patients with different genetic loadings, i.e., the simplex sample, we also included patients who had individual genotype data and age at onset information from another independent sample, the Schizophrenia Trio Genomic Research in Taiwan (S-TOGET). Among the 2,923 patients with SZ from simplex families in S-TOGET, 1,649 had genotype data, and 1,635 of them also had information on age at onset. More detailed information about the S-TOGET cohort has been described previously (29, 30).\u003c/p\u003e\n\u003cp\u003eTo minimize sociodemographic differences between the TSLS and S-TOGET samples, sex- and onset-age-matched subsamples were selected for comparable analysis (\u003cstrong\u003eAdditional file 1:\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eTable S1\u003c/strong\u003e). Both studies were approved by the Institutional Review Board of National Taiwan University Hospital, and all participants provided written informed consent (see details in the \u003cstrong\u003eAdditional file 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMeasurements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe age at onset and basic sociodemographic characteristics were extracted from the DIGS, with the former obtained from the psychosis section that inquired about the age at onset of the first psychotic episode. We classified age at onset as binary, defining early-onset SZ as psychotic symptoms emerging before age 18, based on our prior research (30).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Furthermore, cognitive function was assessed using the Continuous Performance Test (CPT) for sustained attention (31) and the Wisconsin Card Sorting Test (WCST) for executive function (32). CPT sensitivity index (d\u0026prime;), WCST perseverative errors, and categories achieved were included as cognitive indices. Further details on measurements are provided in the \u003cstrong\u003eAdditional file 1\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenotyping, quality control, and imputation\u003c/strong\u003e\u003cbr\u003eGenotyping for the multiplex and simplex samples was performed using the PsychChip array version 1.1 (Illumina, San Diego, CA), containing 588,628 SNPs. Standard quality control (QC) procedures (33) were followed (see details in the \u003cstrong\u003eAdditional file 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eFor the multiplex sample, 1,224 participants and 266,899 SNPs that passed QC were used for genotype imputation, and principal components (PCs) were calculated to account for ancestry in subsequent analyses. The genomic inflation factor (\u0026lambda;) of this final sample was 1.07, and the corresponding QQ plot is shown in \u003cstrong\u003eAdditional file 1:\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eFigure S\u003c/strong\u003e\u003cstrong\u003e2.\u0026nbsp;\u003c/strong\u003eAfter genotype imputation, the post-imputation QC filtered for the variants included poor imputation quality (INFO score \u0026lt; 0.8) and low minor allele frequency (MAF \u0026lt; 0.1). Ultimately, 4,853,692 SNPs were included in later analyses.\u003c/p\u003e\n\u003cp\u003eFor the simplex sample, the same QC procedures resulted in 293,643 SNPs that passed QC, with imputation yielding 5,154,665 SNPs. Details of the QC process have been described previously (29, 30).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCalculation\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eof PRS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePRSs for schizophrenia (SZ-PRS) were calculated for both samples using the p-value thresholding method, utilizing the latest version of summary statistics from the PGC SZ cohort (7). All the SZ-PRS were normalized to z scores for easier interpretation. The best-performing p-value threshold for the SZ-PRS, which explained the most variance in early-onset SZ, was selected for further analysis. Additional details are available in the \u003cstrong\u003eAdditional file 1\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeriving the effect sizes of predicted gene expression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe tissue-specific TWAS for SZ was performed by S-PrediXcan (24), which integrates GWAS summary data of patients with SZ with a SNP-expression prediction model to identify genes associated with SZ (see see details in the \u003cstrong\u003eAdditional file 1\u003c/strong\u003e). The existing SNP-expression prediction model from CMC\u0026nbsp;(12), which was derived from DLPFC brain tissue samples, was used to perform the TWAS for SZ.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePredicting gene expression using individual genotypes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo determine the genetically regulated expression predicted for a SNP, hereafter referred to as predicted gene expression, we used PrediXcan (11) to incorporate the genotypes of individual patients and the corresponding SNP-expression prediction model. The DLPFC SNP-expression prediction model from CMC was used to predict gene expression, and only those genotypes that passed standard quality control procedures were subjected to such prediction.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCalculation of SZ-GeRS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor each gene, its predicted SZ-GeRS was calculated by weighting the predicted gene expression for each patient using the effect size from the TWAS for SZ described in the preceding section. In addition to the SZ-GeRS of individual genes, we constructed two other types of SZ-GeRS by summing genes to identify biologically meaningful relationships with early-onset SZ.\u003c/p\u003e\n\u003cp\u003eThe first type of gene summary SZ-GeRS was a module-based SZ-GeRS, where we utilized gene lists from 13 coexpression modules related to psychiatric disorders identified in a previous WGCNA (25). These modules were derived from postmortem brain tissues of patients with psychotic disorders compared with healthy controls, and we summed the individual SZ-GeRS over the genes listed in those modules. The other type of gene summary SZ-GeRS was constructed by summing over the hub genes of a module. Hub genes were identified via network analysis using IPA, selecting genes regulating the expression of four or more other genes in the top-ranking network. All the SZ-GeRS results were then normalized to z scores for easy interpretation.\u003c/p\u003e\n\u003cp\u003eFurthermore, we utilized the gene lists from the newly derived coexpression modules identified by the same research group using narrowed samples (32) to recalculate the module-based GeRS and assess the reproducibility of the early-onset-related module and its hub genes identified in these procedures (see details in the \u003cstrong\u003eAdditional file 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEnrichment analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo understand possible biological functions of the hub genes, we conducted enrichment analyses using the GENE2FUNC function from the functional mapping and annotation platform (FUMA, https://fuma.ctglab.nl/) to annotate them in the biological context. The gene expression datasets from the BrainSpin consisted of brain tissue samples across various developmental stages. We used the Molecular Signatures Database (MsigDB) (34) to assess enrichment in chemical and genetic perturbation gene sets.\u003c/p\u003e\n\u003cp\u003eAdditionally, cell type enrichment analysis was conducted using the PsychSCREEN gene portal (https://psychscreen.beta.wenglab.org/psychscreen/gene), designed to explore single-cell expression of genes in psychiatric disease. The Schizophrenia Bipolar Disorder Multi-Omics Sequencing dataset from PsychENCODE was utilized for this analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAssociations between risk scores (including SZ-PRS and all types of predicted SZ-GeRS) and early-onset SZ were assessed using a mixed-effects logistic regression model for the multiplex sample, with family index as a random effect. Covariates included sex, education level, and the first four ancestry PCs. For the simplex sample, a logistic regression model was used with identical covariates. The proportion of explained variance was estimated by calculating the change in Nagelkerke\u0026apos;s pseudo-R\u003csup\u003e2\u003c/sup\u003e between the null model (containing only the covariates) and the full model (containing covariates + risk scores) (35). We then conducted a meta-analysis of both samples using the R package \u0026ldquo;metafor\u0026rdquo;\u0026nbsp;(36), employing a restricted maximum likelihood estimator with inverse-variance weighting.\u003c/p\u003e\n\u003cp\u003eThe SZ-PRS and PCs were computed using the PLINK (37), whereas regression analyses and Nagelkerke\u0026apos;s pseudo-R\u003csup\u003e2\u003c/sup\u003e calculations were conducted using the R packages \u0026ldquo;lme4\u0026rdquo; and \u0026ldquo;rcompanion\u0026rdquo; (version 4.2.2), respectively.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003eDistribution of sociodemographic characteristics and neurocognitive performance\u003c/h2\u003e \u003cp\u003eIn the multiplex sample of 595 patients with SZ from 314 multiplex families, 223 patients with early-onset SZ (\u0026le;\u0026thinsp;18 years of age) were younger, had a lower education level, and had lower undegraded CPT d\u0026acute; than 372 patients with late-onset SZ (\u0026gt;\u0026thinsp;18 years of age), with no differences in terms of degraded CPT d\u0026acute; and two indices of the WCST (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). In the simplex sample of 595 patients with SZ matched 1:1 by sex and onset age to the multiplex sample, the early- and late-onset patients presented similar differences in sociodemographic and neurocognitive performance.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDistributions of sociodemographic characteristics and neurocognitive performance indices in patients with schizophrenia grouped by early- and late-onset in the multiplex and simplex samples, respectively.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"43\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c15\" colnum=\"15\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c16\" colnum=\"16\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c17\" colnum=\"17\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c18\" colnum=\"18\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c19\" colnum=\"19\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c20\" colnum=\"20\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c21\" colnum=\"21\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c22\" colnum=\"22\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c23\" colnum=\"23\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c24\" colnum=\"24\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c25\" colnum=\"25\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c26\" colnum=\"26\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c27\" colnum=\"27\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c28\" colnum=\"28\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c29\" colnum=\"29\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c30\" colnum=\"30\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c31\" colnum=\"31\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c32\" colnum=\"32\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c33\" colnum=\"33\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c34\" colnum=\"34\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c35\" colnum=\"35\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c36\" colnum=\"36\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c37\" colnum=\"37\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c38\" colnum=\"38\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c39\" colnum=\"39\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c40\" colnum=\"40\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c41\" colnum=\"41\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c42\" colnum=\"42\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c43\" colnum=\"43\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"22\" nameend=\"c23\" namest=\"c2\"\u003e \u003cp\u003eMultiplex sample\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c27\" namest=\"c24\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"14\" nameend=\"c41\" namest=\"c28\"\u003e \u003cp\u003eSimplex sample\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c43\" namest=\"c42\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c9\" namest=\"c2\"\u003e \u003cp\u003eEarly-onset patients\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;223)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c12\" namest=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c18\" namest=\"c13\"\u003e \u003cp\u003eLate-onset patients\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;372)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c20\" namest=\"c19\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c23\" namest=\"c21\"\u003e \u003cp\u003eGroup Comparison\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c26\" namest=\"c24\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c32\" namest=\"c27\"\u003e \u003cp\u003eEarly-onset patients\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;223)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c35\" namest=\"c33\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c39\" namest=\"c36\"\u003e \u003cp\u003eLate-onset patients\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;372)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c40\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003eGroup Comparison\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e(S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c15\" namest=\"c14\"\u003e \u003cp\u003e(S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c17\" namest=\"c16\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c21\" namest=\"c18\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c24\" namest=\"c22\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c28\" namest=\"c25\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c30\" namest=\"c29\"\u003e \u003cp\u003e(S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c33\" namest=\"c31\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c36\" namest=\"c34\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c38\" namest=\"c37\"\u003e \u003cp\u003e(S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c40\" namest=\"c39\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale sex, n %\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e131\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(58.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e224\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(60.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e131\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(58.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e224\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(60.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge, years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e30.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(7.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e34.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(6.72)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e\u0026lt;0.001**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e30.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(7.61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e35.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(7.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge at onset, years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e15.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(2.13)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e24.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(5.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e16.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(1.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e24.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(4.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation level, years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e9.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(2.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e11.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(2.83)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e11.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(2.40)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e12.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(2.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e0.001**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCPT indices\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUndegraded CPT d\u0026acute;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e-2.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(2.15)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e-2.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(2.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e0.018*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e-2.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(2.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e-1.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(1.77)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e0.019*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDegraded\u0026nbsp;CPT d\u0026acute;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e-2.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(1.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e-2.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(1.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e-1.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(1.66)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e-1.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(1.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e0.276\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWCST indices\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCategories achieved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e-1.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(0.72)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e-1.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(0.84)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e0.476\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e-0.90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(0.98)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e-0.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(0.96)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e0.882\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerseverative errors\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e(1.74)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c11\" namest=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c14\" namest=\"c12\"\u003e \u003cp\u003e1.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c16\" namest=\"c15\"\u003e \u003cp\u003e(1.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c19\" namest=\"c17\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c22\" namest=\"c20\"\u003e \u003cp\u003e0.361\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c25\" namest=\"c23\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c29\" namest=\"c26\"\u003e \u003cp\u003e1.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c31\" namest=\"c30\"\u003e \u003cp\u003e(1.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c34\" namest=\"c32\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c36\" namest=\"c35\"\u003e \u003cp\u003e1.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c37\"\u003e \u003cp\u003e(1.72)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c40\" namest=\"c38\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c42\" namest=\"c41\"\u003e \u003cp\u003e0.457\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c43\" namest=\"c43\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"43\" nameend=\"c43\" namest=\"c1\"\u003e \u003cp\u003e*p value\u0026thinsp;\u0026lt;\u0026thinsp;.05; **\u0026lt;0.01\u003c/p\u003e \u003cp\u003e\u003csup\u003ea\u003c/sup\u003e 106 individuals in the early-onset group (47.5%) and 250 individuals in the late-onset group (67.2%) were siblings from multiplex families with SZ.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"43\" nameend=\"c43\" namest=\"c1\"\u003e \u003cp\u003e\u003csup\u003eb\u003c/sup\u003e The group comparisons were conducted using a mixed-effect logistic regression model with family index as a random effect.\u003c/p\u003e \u003cp\u003e\u003csup\u003ec\u003c/sup\u003e The group comparisons were conducted using logistic regression.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"43\" nameend=\"c43\" namest=\"c1\"\u003e \u003cp\u003e\u003csup\u003ed\u003c/sup\u003e The adjusted z scores were derived by means of standardizing the raw scores with adjustments for sex, age and education against a community sample of 345\u003c/p\u003e \u003cp\u003eindividuals (31).\u003c/p\u003e \u003cp\u003e\u003csup\u003ee\u003c/sup\u003e The adjusted z scores were derived by means of standardizing the raw scores with adjustments for sex, age and education against a group of 392 healthy controls (32).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eAssociation of PRS for SZ with early-onset SZ\u003c/h2\u003e \u003cp\u003eRegarding the extent to which the difference in onset age was attributed to genetic predisposition, the SZ-PRS at the \u003cem\u003ep\u003c/em\u003e-value threshold of 0.1 was associated with early-onset SZ (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03) and explained the most variance of early-onset SZ (pseudo-R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;1.11%) in the multiplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003eA), whereas the SZ-PRS at the \u003cem\u003ep\u003c/em\u003e-value threshold of 0.01 was not associated with early-onset SZ (p\u0026thinsp;=\u0026thinsp;0.14) but explained the most variance of early-onset SZ (pseudo-R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.50%) in the simplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). We then compared the SZ-PRS scores of the two groups of patients with those of unaffected relatives using the same method of PRS derivation. Among both the multiplex (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003eC) and simplex (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003eD) samples, patients with early-onset SZ had the highest average SZ-PRS score, followed by patients with late-onset SZ and then their unaffected relatives.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePredicted GeRS for SZ\u003c/h3\u003e\n\u003cp\u003eBefore building the GeRS for our samples, we obtained a weighting system for the predicted expression of individual SNPs through TWAS of SZ in an existing database to identify the associations between the \u003cem\u003ecis\u003c/em\u003e-genetic component of expression and SZ. Among the 10,358 genes included in the existing SNP-expression prediction model from the CMC (12) for the DLPFC, the expression levels of 10,306 genes identified in the GWAS of PGC SZ cohort using S-PrediXcan were successfully predicted, with the prediction performance per gene ranging from 0.8\u0026ndash;81.5% and a mean of 8% (s.d.=11.1%). A comparison of the predicted expression levels of these genes in the PGC for patients with SZ and their controls revealed that 46 genes showed associations with SZ that reached genome-wide significance (p\u0026thinsp;\u0026lt;\u0026thinsp;5\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e) (\u003cb\u003eAdditional file 2: Supplementary Data 1\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eWe then utilized the effect size of the preceding TWAS of SZ as the weight of the predicted expression of individual SNPs to derive the predicted SZ-GeRS for our samples. \u003cb\u003eFigure S3\u003c/b\u003e depicts the genome-wide association between the predicted SZ-GeRS and early-onset SZ in the multiplex sample, with 7834 genes predicted to be expressed; none of them showed an association that reached genome-wide significance. When the same analysis was repeated in the simplex sample, 7987 genes were predicted to be expressed, with none reaching genome-wide significance (\u003cb\u003eAdditional file 1: Figure S4\u003c/b\u003e).\u003c/p\u003e\n\u003ch3\u003eModule-based GeRS for SZ and early-onset SZ\u003c/h3\u003e\n\u003cp\u003eTo capture the possible contribution of the combination of expression levels of a group of genes, we turned to 13 modules derived from coexpression analysis in a previous study using \u003cem\u003epostmortem\u003c/em\u003e brain tissues of patients with five major psychiatric disorders versus those of healthy controls (25). When the association between individual module-based SZ-GeRS and early-onset SZ was assessed in the multiplex sample, the SZ-GeRS based on module 10 (M10-GeRS) explained the most variance in early-onset SZ (pseudo-R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;2.4%, p\u0026thinsp;=\u0026thinsp;0.005; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Among the 161 genes included in the original module 10, 94 had SNP-expression weights in the DLPFC, 57 of which were predicted to have genetically regulated expression in our multiplex sample (\u003cb\u003eAdditional file 1: Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e)\u003c/b\u003e. We then examined the same 94 genes located in the DLPFC of module 10 in the simplex sample and found that 61 genes were predicted to have genetically regulated expression, with 56 of them also appearing in the multiplex samples. Hence, these 56 genes were considered a subset of M10 (M10\u003csub\u003esub\u003c/sub\u003e), and the corresponding SZ-GeRS was denoted as M10\u003csub\u003esub\u003c/sub\u003e-GeRS. Patients with early-onset SZ presented significantly greater standardized M10\u003csub\u003esub\u003c/sub\u003e-GeRS values than did those with late-onset SZ in both the multiplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eB) and the simplex sample \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e\n\u003ch3\u003eHub genes for early-onset SZ\u003c/h3\u003e\n\u003cp\u003eTo identify possible hub genes, we conducted a network analysis of the 56 genes of M10\u003csub\u003esub\u003c/sub\u003e-GeRS in the multiplex sample using the IPA and revealed four networks with functionally connected genes, in which only one network had an IPA score\u0026thinsp;\u0026gt;\u0026thinsp;40, i.e., a score of 75 for Network 1 (\u003cb\u003eAdditional file 2: Supplementary Data 2)\u003c/b\u003e. Focusing on this network, we identified \u003cem\u003eRUVBL2, COPS6, TUBA4A, PSMB5, PSMD2\u003c/em\u003e, and \u003cem\u003eLRPPRC\u003c/em\u003e as hub genes that were correlated with the expression of more than four genes within the network (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003e \u003cb\u003eand Additional file 2: Supplementary Data 3)\u003c/b\u003e. When these six hub genes were used to derive M10\u003csub\u003ehub\u003c/sub\u003e-GeRS, the distribution of early-onset SZ tended to be greater than that of late-onset SZ in both the multiplex and simplex samples, although the difference reached statistical significance only in the latter (\u003cb\u003eAdditional file 1: Figure S8)\u003c/b\u003e.\u003c/p\u003e\n\u003ch3\u003eMeta-analysis of the association\u003c/h3\u003e\n\u003cp\u003eWe then conducted a series of multivariable logistic regression analyses, adjusting for sex, educational level, and four ancestry PCs for the multiplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) and the simplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003eB), respectively. After accounting for SZ-PRS, the variance explained by M10\u003csub\u003esub\u003c/sub\u003e-GeRS (or M10\u003csub\u003ehub\u003c/sub\u003e-GeRS) was 2.5% (or 0.7%) in the multiplex sample and 3.6% (or 1.3%) in the simplex sample. In terms of the association of M10\u003csub\u003esub\u003c/sub\u003e-GeRS (or M10\u003csub\u003ehub\u003c/sub\u003e-GeRS) with early-onset SZ, the corresponding adjusted ORs [aORs] in Model 3 (or Model 4) for the multiplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) and the simplex sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003eB) were similar in direction and magnitude, although some did not reach statistical significance, i.e., M10\u003csub\u003ehub\u003c/sub\u003e-GeRS in Model 4 for the multiplex sample.\u003c/p\u003e \u003cp\u003eHence, we aggregated both the multiplex and the simplex samples to conduct a meta-analysis to estimate summary effects (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). The meta-analysis revealed that the aORs of SZ-PRS (Model 0), M10\u003csub\u003esub\u003c/sub\u003e-GeRS (Model 1), and M10\u003csub\u003ehub\u003c/sub\u003e-GeRS (Model 3) were significantly greater than one. Nevertheless, even after adjustment for SZ-PRS and the covariates, significant associations with early-onset SZ existed for M10\u003csub\u003esub\u003c/sub\u003e-GeRS (aOR\u0026thinsp;=\u0026thinsp;1.38, 95% CI\u0026thinsp;=\u0026thinsp;1.22\u0026ndash;1.57) and M10\u003csub\u003ehub\u003c/sub\u003e-GeRS (aOR\u0026thinsp;=\u0026thinsp;1.22, 95% CI\u0026thinsp;=\u0026thinsp;1.07\u0026ndash;1.39). That is, reducing the genes in gene-summing GeRS from 56 (M10\u003csub\u003esub\u003c/sub\u003e-GeRS) to 6 (M10\u003csub\u003ehub\u003c/sub\u003e-GeRS) did not diminish the association.\u003c/p\u003e \u003cp\u003eTo examine the robustness of the module-based GeRS, we used another set of modules derived from 3 diagnostic groups (SZ, bipolar disorder, and autism) (38) to reconstruct individual module-based GeRS (more detail in the \u003cb\u003eAdditional file 1: Supplementary Note\u003c/b\u003e). Intriguingly, 2 of the 8 genes designated as hub genes using module 9 in these newly derived modules were also defined as hub genes in the originally identified early-onset-related module (i.e., module 10).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eFunctional mapping of the hub genes\u003c/h2\u003e \u003cp\u003eThe known functions and previously established genetic association of the six hub genes, including \u003cem\u003ePSMB5\u003c/em\u003e (39\u0026ndash;42), \u003cem\u003ePSMD2\u003c/em\u003e (43, 44), \u003cem\u003eTUBA4A\u003c/em\u003e (45\u0026ndash;47), \u003cem\u003eLRPPRC\u003c/em\u003e (48, 49), \u003cem\u003eRUVBL2\u003c/em\u003e (50, 51), and \u003cem\u003eCOPS6\u003c/em\u003e (52, 53), are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Briefly, they are mainly involved with cellular processes crucial for neurological function and implicated in various neurodevelopmental and neurodegenerative disorders, including proteasomal protein degradation, microtubule dynamics, mitochondrial function, chromatin remodeling, and protein degradation.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBackground on the six hub genes identified from module 10, including brief descriptions of their genetic functions and previously established genetic associations.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGene\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eKnown genetic functions and relationships\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003ePSMB5\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003ePSMB5\u003c/em\u003e encodes the proteasome subunit beta type-5. The proteasome system is involved in regulating major cellular pathways, including those leading to inflammation (39), and has been identified as a canonical pathway associated with SZ (40, 41) and treatment resistance in major depressive disorder (42).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003ePSMD2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003ePSMD2\u003c/em\u003e, which encodes a non-ATPase subunit of the 26S proteasome, is crucial for regulating protein degradation and cellular processes (43). Recent in vivo studies show that neuronal membrane proteasomes regulate neuronal circuit activity, potentially involving PSMD2, and learning-induced behavioral plasticity, thereby highlighting the role of PSMD2 in neuronal function (44).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eTUBA4A\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eTUBA4A\u003c/em\u003e encodes alpha tubulin 4A, a key component of microtubules regulating axonal transport in the central nervous system (45). \u003cem\u003eTUBA4A\u003c/em\u003e demonstrates ubiquitous expression across all cell types, with its highest levels observed in the brain (46), and has been associated with amyotrophic lateral sclerosis and frontotemporal dementia (47).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eLRPPRC\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eLRPPRC\u003c/em\u003e, which encodes leucine-rich pentatricopeptide repeat-containing protein, plays a crucial role in mitochondrial function and is involved in regulating mRNA stability and polyadenylation and coordinating mitochondrial translation (48). LRPPRC deficiency is linked with ATP synthase deficiency and severe neurodegenerative disorders such as Leigh syndrome, underscoring its critical role in brain function (49).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eRUVBL2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eRUVBL2\u003c/em\u003e encodes RuvBL2 (RuvB-like 2), which is part of various ATP-dependent chromatin remodeling complexes that are involved in the regulation of gene expression (50). \u003cem\u003eRUVBL2\u003c/em\u003e predominantly manifests in the SOX2-positive compartment of cerebral organoids and mid-gestation fetal brain tissue, with its chemical inactivation leading to precursor cell displacement and apoptosis and genetic variants linked to neurodevelopmental impairments (51).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eCOPS6\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eCOPS6\u003c/em\u003e encodes the sixth subunit of the photomorphogenic 9 (COP9) signalsome (CSN), which is involved in the ubiquitin-mediated protein degradation process (52) and is further implicated in neurodegenerative diseases (53).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWe then performed gene-to-function analyses for those hub genes using the FUMA to understand which biological pathways or gene sets might be involved (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003e). \u003cem\u003ePSMB5\u003c/em\u003e and \u003cem\u003eCOPS6\u003c/em\u003e showed high expression from early prenatal stages through middle adulthood, while \u003cem\u003eRUVBL2\u003c/em\u003e and \u003cem\u003ePSMD2\u003c/em\u003e had elevated expression only during early prenatal stages, diminishing after infancy (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). In contrast, \u003cem\u003eTUBA4A\u003c/em\u003e exhibited increased expression after infancy. Also, several chemical and genetic perturbation gene sets showed significant enrichment (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), including those related to psychiatric disorders duration, B lymphocytes regulatory network, and perineuronal oligodendrocytes in the BA9 brain region.\u003c/p\u003e \u003cp\u003eAdditionally, we utilized brain single-cell expression from psychENCODE, including samples with SZ, bipolar disorder, and controls, to better understand the expression patterns of these hub genes across various cell types (\u003cb\u003eAdditional file 1: Figure S9\u003c/b\u003e). We found most hub genes were highly enriched in excitatory neurons compared to other cell types except \u003cem\u003eLRPPRC\u003c/em\u003e.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eDespite antipsychotics being generally used to treat the symptoms of SZ, many individuals with early-onset SZ exhibit poor responses and resistance to treatment. The antipsychotics currently available were developed based on serendipitous clinical observations of the antipsychotic properties of drugs initially used for other indications (54). Identifying new drug targets by utilizing genomic information to understand the pathophysiology of early-onset schizophrenia is essential. Here, we applied the GeRS approach to early-onset SZ and identified six hub genes, including \u003cem\u003eRUVBL2, COPS6, TUBA4A, PSMB5, PSMB2\u003c/em\u003e, and \u003cem\u003eLRPPRC\u003c/em\u003e. Functional mapping and enrichment analysis revealed potential biological mechanisms involved in early-onset SZ, e.g., expression changes in excitatory neurons and immune-regulatory processes. Our findings provide clues for downstream investigations into the etiology of early-onset SZ and may potentially be applied in early interventions.\u003c/p\u003e \u003cp\u003eSeveral strategies were implemented to ensure the robustness of our findings, particularly 1) examining the module-based GeRS in both the multiplex and the simplex samples of patients with SZ and then aggregating both the multiplex and the simplex samples to conduct a meta-analysis to obtain estimates of summary effects, 2) adopting the SNP-expression prediction model derived from brain tissues of patients with psychiatric disorders to avoid potential bias resulting from nontrait-related tissues in different organs (13, 55), 3) and utilizing existing brain tissue gene expression data to map the identified genes to their biological context.\u003c/p\u003e \u003cp\u003eDespite the differing familial loadings in the multiplex and the simplex samples of patients with SZ, the differences in neurocognitive performance and SZ-PRS between early-onset SZ and late-onset SZ remained and were in line with previous findings, e.g., early-onset SZ displaying more neurocognitive impairment (56, 57) and higher SZ-PRS (58, 59) than late-onset SZ. This, in part, justified our conduction of meta-analysis to estimate the summary effects of the identified genes.\u003c/p\u003e \u003cp\u003eAlthough none of the individual SNPs in SZ-GeRS were significantly associated with early-onset SZ in this study, similar to previous studies (15, 60), we summarized genes from empirically derived gene clusters to obtain a functionally meaningful SZ-GeRS (25, 38). The modules derived from the postmortem brain tissues of patients with five psychiatric disorders (25) led to greater variance explained by the M10\u003csub\u003ehub\u003c/sub\u003e-GeRS than that explained by the modules from patients with three psychiatric disorders and two specific brain regions (38). There are several possibilities, including the following: 1) many mental disorders of different diagnostic categories have shared genetic susceptibility; 2) preselection of brain regions may overlook contributions from other regions; 3) the gene list of module 10 has been found to overlap with a neuronal mitochondria gene-enriched module previously associated with fundamental neuronal processes and psychiatric disease (25); and 4) module 10 has been used to enrich the coexpression module profiling from the brain tissues of mouse models harboring SZ and autism-associated mutations, identifying its association with neuronal energetics and firing rate (61).\u003c/p\u003e \u003cp\u003eOur findings support the postulation that GeRS has better portability across ancestry groups for predicting traits compared to conventional PRS (62, 63), based on the hypothesis that GeRS, derived from quantitative-trait locus variants potentially shared among different ancestral groups, is associated with shared biological mechanisms. Moreover, we observed that adding module-based SZ-GeRS to the model containing SZ-PRS and potential confounders would substantially increase r\u003csup\u003e2\u003c/sup\u003e and maintain the significant pattern of main effects in both the multiplex and the simplex samples. This implies that module-based SZ-GeRS and SZ-PRS each provided their respective contributions to early-onset SZ and that module-based SZ-GeRS might be a gene-based complement to SZ-PRS, potentially improving the ability to distinguish between early-onset and late-onset SZ. Our findings align with the fact that GeRS and PRS quantify independent genetic effects on complex traits (14, 62) and that integrating information from functional GeRS can improve risk prediction for these traits (63, 64).\u003c/p\u003e \u003cp\u003eBased on the 56 genes of one subset of M10, M10\u003csub\u003esub\u003c/sub\u003e, this study utilized network analysis to narrow down to six hub genes that were predicted to highly regulate the expression of other genes within the coexpression network of module 10, and hence their aberrant expressions might impact downstream biological functions and lead to the onset of the illness (65). This possibility is further supported by subsequent explorations of the hub genes, from literature review to functional mapping of gene sets in developmental temporality and specific diseases or traits, and to single-cell gene expression profiles. These enrichment analyses provide further insight that these hub genes might have diverse expression patterns across developmental stages. Their coordination during brain development may reflect an atypical \u0026ldquo;transcriptional program\u0026rdquo; that deviates from normal development (66). This deviation could contribute to abnormal neural processes, potentially facilitating early-onset SZ. Additionally, several enriched gene sets link these genes to neurodevelopmental processes, such as the number of perineuronal oligodendrocytes in the BA9 region, and immune dysregulation, particularly in the B lymphocyte regulatory network. Reduction of perineuronal oligodendrocytes, as reported in a recent systematic review of postmortem brain studies (67), has been associated with major psychiatric disorders, and immune system dysfunction and inflammatory processes have long been associated with these disorders (68). The enriched gene set linked to the duration of psychiatric disorders might account in part for the clinical characteristics of early-onset SZ, i.e., being associated with more severe symptoms and treatment resistance, and hence resulting in a longer duration of illness. Moreover, most hub genes were enriched in excitatory neurons, indicating these cell types are likely central to the relationship between gene expression and psychiatric disorders. This support recent idea (69) that genetic risk variants affect genes with measurable expression changes in excitatory neurons of individuals with SZ.\u003c/p\u003e \u003cp\u003eThese hub genes have significant clinical relevance due to their dual roles as both susceptibility genes and modifier genes for SZ. While most antipsychotics are designed to target susceptibility genes, early-onset SZ involves modifier genes. This distinction may explain why patients with early-onset SZ often exhibit poor responses to existing antipsychotics, resulting in treatment resistance. Hence, our findings provide valuable insights for identifying potential drug targets and improving treatment for the earliest stages of this illness.\u003c/p\u003e \u003cp\u003eOur study has several limitations. First, as this study was based on a case-only design, we potentially identified only the 'mixed genes' for SZ, which have dual functions as susceptibility genes and modifier genes (23, 70), but overlooked purely modifier genes for early-onset SZ. Second, the small sample size of this study might render individual SNPs included in M10\u003csub\u003esub\u003c/sub\u003e-GeRS and M10\u003csub\u003ehub\u003c/sub\u003e-GeRS not being able to reach transcriptome-wide significance. Nevertheless, the combination of these genes was able to show strong associations with early-onset SZ. Last, the aberrant gene expression, as incorporated in the module-derived GeRS, was based on prediction; future \u003cem\u003ein vitro\u003c/em\u003e and animal studies are needed to elucidate their roles in the biological mechanisms of early-onset SZ.\u003c/p\u003e \u003cp\u003eIn conclusion, we found that a module-based GeRS approach that incorporates predicted gene expression resulting from the interaction of genes and environmental factors may help capture further contributions to early-onset SZ. Pathway and enrichment analyses revealed that the identified hub genes might be involved in deviation from typical brain development and expression changes in excitatory neurons. Our findings will enhance the understanding of SZ etiology and may contribute to early screening and personalized prevention efforts.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e \u003cb\u003eAcknowledgements and Disclosures\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThis study was supported by grants from the Ministry of Science and Technology, Taiwan (107-2314-B-002-214-MY3; 109-2314-B-002-172-MY3); the National Health Research Institutes, Taiwan (NP-109-SP-07, NP-110-SP-02, and NP-111-SP-02); the NIH/NHGRI grant U54HG003067; NIMH grants R01 MH085521 and R01 MH085560; the Gerber Foundation; the Sidney R. Baer, Jr. Foundation; NARSAD: The Brain and Behavior Research Foundation; and the Stanley Center for Psychiatric Research.\u003c/p\u003e \u003cp\u003eDr. Neale reports serving as a member of the Scientific Advisory Board of Deep Genomics and as a consultant for Camp4 Therapeutics Corporation, Merck \u0026amp; Co., Takeda Pharmaceutical, and Avanir Pharmaceuticals, Inc. All the other authors report no biomedical financial interests or potential conflicts of interest.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDr. Neale reports serving as a member of the Scientific Advisory Board of Deep Genomics and as a consultant for Camp4 Therapeutics Corporation, Merck \u0026amp; Co., Takeda Pharmaceutical, and Avanir Pharmaceuticals, Inc. All the other authors report no biomedical financial interests or potential conflicts of interest.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eW.J.C. was responsible for the conception, design of the study, and funding acquisition. Y.J. performed genotype data preprocessing, developed analytical schemes, and conducted statistical analyses. S.-L.Y. provided technical and material support for DNA experiments. P.-C.H. provided technical support. C.-M.L., C.-C.L., T.-J.H., M.H.H., Y.-L.C., Y.-T.L., H.-G.H., and W.J.C. managed participant recruitment and data acquisition. H.H., Y.A.F., P.-H.K., C.K.H., and Y.-F.L. contributed to the interpretation of the data. S.V.F., B.N., S.J.G., and M.T.T. were responsible for conceptualization, resources, and supervision. Y.J. and W.J.C. implemented the literature review and data interpretation. The first draft was written by Y.J., and W.J.C. provided critical revisions and supervised the manuscript. All authors reviewed, commented on, and approved the final version of the manuscript.\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe datasets used and analyzed in the current study are not publicly available due to conditions in the participant consent and other ethical restrictions. However, the data that support the findings of this study are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGBD 2016 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990\u0026ndash;2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017; 390: 1211\u0026ndash;1259.\u003c/li\u003e\n\u003cli\u003eMoreno-K\u0026uuml;stner B, Mart\u0026iacute;n C, Pastor L. Prevalence of psychotic disorders and its association with methodological issues. A systematic review and meta-analyses. PLoS ONE 2018; 13: e0195687.\u003c/li\u003e\n\u003cli\u003eSullivan PF, Kendler KS, Neale MC. Schizophrenia as a complex trait: evidence from a meta-analysis of twin studies. Arch Gen Psychiatry 2003; 60: 1187\u0026ndash;1192.\u003c/li\u003e\n\u003cli\u003eHilker R, Helenius D, Fagerlund B, Skytthe A, Christensen K, Werge TM, et al. Heritability of schizophrenia and schizophrenia spectrum based on the nationwide Danish Twin Register. Biol Psychiatry 2018; 83: 492\u0026ndash;498.\u003c/li\u003e\n\u003cli\u003eSchizophrenia Working Group of the Psychiatric Genomics Consortium. Biological insights from 108 schizophrenia-associated genetic loci. Nature 2014; 511: 421\u0026ndash;427.\u003c/li\u003e\n\u003cli\u003eLam M, Chen C-Y, Li Z, Martin AR, Bryois J, Ma X, et al. Comparative genetic architectures of schizophrenia in East Asian and European populations. Nat Genet 2019; 51: 1670\u0026ndash;1678.\u003c/li\u003e\n\u003cli\u003eTrubetskoy V, Pardi\u0026ntilde;as AF, Qi T, Panagiotaropoulou G, Awasthi S, Bigdeli TB, et al. Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature 2022; 604: 502\u0026ndash;508.\u003c/li\u003e\n\u003cli\u003eOwen MJ, Sawa A, Mortensen PB. Schizophrenia. Lancet 2016; 388: 86\u0026ndash;97.\u003c/li\u003e\n\u003cli\u003eJaffe AE. Postmortem human brain genomics in neuropsychiatric disorders\u0026mdash;how far can we go? Curr Opin Neurobiol 2016; 36: 107\u0026ndash;111.\u003c/li\u003e\n\u003cli\u003eZhu Y, Wang L, Yin Y, Yang E. Systematic analysis of gene expression patterns associated with postmortem interval in human tissues. Sci Rep 2017; 7: 5435.\u003c/li\u003e\n\u003cli\u003eGamazon ER, Wheeler HE, Shah KP, Mozaffari SV, Aquino-Michaels K, Carroll RJ, et al. A gene-based association method for mapping traits using reference transcriptome data. Nat Genet 2015; 47: 1091\u0026ndash;1098.\u003c/li\u003e\n\u003cli\u003eHuckins LM, Dobbyn A, Ruderfer DM, Hoffman G, Wang W, Pardi\u0026ntilde;as AF, et al. Gene expression imputation across multiple brain regions provides insights into schizophrenia risk. Nat Genet 2019; 51: 659\u0026ndash;674.\u003c/li\u003e\n\u003cli\u003eGusev A, Mancuso N, Won H, Kousi M, Finucane HK, Reshef Y, et al. Transcriptome-wide association study of schizophrenia and chromatin activity yields mechanistic disease insights. Nat Genet 2018; 50: 538\u0026ndash;548.\u003c/li\u003e\n\u003cli\u003eRodriguez-L\u0026oacute;pez J, Arrojo M, Paz E, P\u0026aacute;ramo M, Costas J. Identification of relevant hub genes for early intervention at gene coexpression modules with altered predicted expression in schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry 2020; 98: 109815.\u003c/li\u003e\n\u003cli\u003eProhens L, Rodr\u0026iacute;guez N, Segura \u0026Agrave;-G, Mart\u0026iacute;nez-Pinte\u0026ntilde;o A, Olivares-Berjaga D, Mart\u0026iacute;nez I, et al. Gene expression imputation provides clinical and biological insights into treatment-resistant schizophrenia polygenic risk. Psychiatry Res 2024; 332: 115722.\u003c/li\u003e\n\u003cli\u003eSolmi M, Radua J, Olivola M, Croce E, Soardo L, Salazar de Pablo G, et al. Age at onset of mental disorders worldwide: large-scale meta-analysis of 192 epidemiological studies. Mol Psychiatry 2022; 27: 281\u0026ndash;295.\u003c/li\u003e\n\u003cli\u003eCoulon N, Godin O, Bulzacka E, Dubertret C, Mallet J, Fond G, et al. Early and very early-onset schizophrenia compared with adult-onset schizophrenia: French FACE-SZ database. Brain Behav 2020; 10: e01495.\u003c/li\u003e\n\u003cli\u003eChen WJ. Taiwan Schizophrenia Linkage Study: Lessons learned from endophenotype-based genome-wide linkage scans and perspective. Am J Med Genet B Neuropsychiatr Genet 2013; 162B: 636\u0026ndash;647.\u003c/li\u003e\n\u003cli\u003eBirnbaum R, Weinberger DR. Genetic insights into the neurodevelopmental origins of schizophrenia. Nat Rev Neurosci 2017; 18: 727\u0026ndash;740.\u003c/li\u003e\n\u003cli\u003eIasevoli F, Razzino E, Altavilla B, Avagliano C, Barone A, Ciccarelli M, et al. Relationships between early age at onset of psychotic symptoms and treatment resistant schizophrenia. Early Interv Psychiatry 2022; 16: 352\u0026ndash;362.\u003c/li\u003e\n\u003cli\u003eBergen SE, O'Dushlaine CT, Lee PH, Fanous AH, Ruderfer DM, Ripke S, et al. Genetic modifiers and subtypes in schizophrenia: Investigations of age at onset, severity, sex and family history. Schizophr Res 2014; 154: 48\u0026ndash;53.\u003c/li\u003e\n\u003cli\u003eWoolston AL, Hsiao P-C, Kuo P-H, Wang S-H, Lien Y-J, Liu C-M, et al. Genetic loci associated with an earlier age at onset in multiplex schizophrenia. Sci Rep 2017; 7: 6486.\u003c/li\u003e\n\u003cli\u003eZhan N, Sham PC, So HC, Lui SSY. The genetic basis of onset age in schizophrenia: evidence and models. Front Genet 2023; 14: 1163361.\u003c/li\u003e\n\u003cli\u003eBarbeira AN, Dickinson SP, Bonazzola R, Zheng J, Wheeler HE, Torres JM, et al. Exploring the phenotypic consequences of tissue specific gene expression variation inferred from GWAS summary statistics. Nat Commun 2018; 9: 1825.\u003c/li\u003e\n\u003cli\u003eGandal MJ, Haney JR, Parikshak NN, Leppa V, Ramaswami G, Hartl C, et al. Shared molecular neuropathology across major psychiatric disorders parallels polygenic overlap. Science 2018; 359: 693\u0026ndash;697.\u003c/li\u003e\n\u003cli\u003eHwu H-G, Faraone SV, Liu C-M, Chen WJ, Liu S-K, Shieh M-H, et al. Taiwan Schizophrenia Linkage Study: The field study. Am J Med Genet B Neuropsychiatr Genet 2005; 134B: 30\u0026ndash;36.\u003c/li\u003e\n\u003cli\u003eFaraone SV, Hwu H-G, Liu C-M, Chen WJ, Tsuang M-M, Liu S-K, et al. Genome scan of Han Chinese schizophrenia families from Taiwan: confirmation of linkage to 10q22.3. Am J Psychiatry 2006; 163: 1760\u0026ndash;1766.\u003c/li\u003e\n\u003cli\u003eNurnberger JI, Jr., Blehar MC, Kaufmann CA, York-Cooler C, Simpson SG, Harkavy-Friedman J, et al. Diagnostic Interview for Genetic Studies: Rationale, unique features, and training. Arch Gen Psychiatry 1994; 51: 849\u0026ndash;859.\u003c/li\u003e\n\u003cli\u003eRees E, Kirov G, Walters JT, Richards AL, Howrigan D, Kavanagh DH, et al. Analysis of exome sequence in 604 trios for recessive genotypes in schizophrenia. Transl Psychiatry 2015; 5: e607.\u003c/li\u003e\n\u003cli\u003eWang SH, Hsiao PC, Yeh LL, Liu CM, Liu CC, Hwang TJ, et al. Advanced paternal age and early-onset of schizophrenia in sporadic cases: not confounded by parental polygenic risk to schizophrenia. Biol Psychiatry 2019; 86: 56\u0026ndash;64.\u003c/li\u003e\n\u003cli\u003eChen WJ, Hsiao CK, Hsiao LL, Hwu HG. Performance of the Continuous Performance Test among community samples. Schizophr Bull 1998; 24: 163\u0026ndash;174.\u003c/li\u003e\n\u003cli\u003eLin S-H, Liu C-M, Hwang T-J, Hsieh MH, Hsiao P-C, Faraone SV, et al. Performance on the Wisconsin Card Sorting Test in families of schizophrenia patients with different familial loadings. Schizophr Bull 2013; 39: 537\u0026ndash;546.\u003c/li\u003e\n\u003cli\u003ePeterson RE, Kuchenbaecker K, Walters RK, Chen CY, Popejoy AB, Periyasamy S, et al. Genome-wide association studies in ancestrally diverse populations: Opportunities, methods, pitfalls, and recommendations. Cell 2019; 179: 589\u0026ndash;603.\u003c/li\u003e\n\u003cli\u003eLiberzon A, Birger C, Thorvaldsd\u0026oacute;ttir H, Ghandi M, Mesirov JP, Tamayo P. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 2015; 1: 417\u0026ndash;425.\u003c/li\u003e\n\u003cli\u003eNagelkerke NJD. A note on a general definition of the coefficient of determination. Biometrika 1991; 78: 691\u0026ndash;692.\u003c/li\u003e\n\u003cli\u003eViechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw 2010; 36: 1\u0026ndash;48.\u003c/li\u003e\n\u003cli\u003ePurcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, et al. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559\u0026ndash;575.\u003c/li\u003e\n\u003cli\u003eGandal MJ, Zhang P, Hadjimichael E, Walker RL, Chen C, Liu S, et al. Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science 2018; 362: eaat8127.\u003c/li\u003e\n\u003cli\u003eEbstein F, Kloetzel PM, Kr\u0026uuml;ger E, Seifert U. Emerging roles of immunoproteasomes beyond MHC class I antigen processing. Cell Mol Life Sci 2012; 69: 2543\u0026ndash;2558.\u003c/li\u003e\n\u003cli\u003eBousman CA, Chana G, Glatt SJ, Chandler SD, Lucero GR, Tatro E, et al. Preliminary evidence of ubiquitin proteasome system dysregulation in schizophrenia and bipolar disorder: convergent pathway analysis findings from two independent samples. Am J Med Genet B Neuropsychiatr Genet 2010; 153B: 494\u0026ndash;502.\u003c/li\u003e\n\u003cli\u003eLee YH, Kim JH, Song GG. Pathway analysis of a genome-wide association study in schizophrenia. Gene 2013; 525: 107\u0026ndash;115.\u003c/li\u003e\n\u003cli\u003eMinelli A, Magri C, Barbon A, Bonvicini C, Segala M, Congiu C, et al. Proteasome system dysregulation and treatment resistance mechanisms in major depressive disorder. Transl Psychiatry 2015; 5: e687.\u003c/li\u003e\n\u003cli\u003eYing K, Wang C, Liu S, Kuang Y, Tao Q, Hu X. Diverse Ras-related GTPase DIRAS2, downregulated by PSMD2 in a proteasome-mediated way, inhibits colorectal cancer proliferation by blocking NF-\u0026kappa;B signaling. Int J Biol Sci 2022; 18: 1039\u0026ndash;1050.\u003c/li\u003e\n\u003cli\u003eHe HY, Ahsan A, Bera R, McLain N, Faulkner R, Ramachandran KV, et al. Neuronal membrane proteasomes regulate neuronal circuit activity in vivo and are required for learning-induced behavioral plasticity. Proc Natl Acad Sci USA 2023; 120: e2216537120.\u003c/li\u003e\n\u003cli\u003eBreuss MW, Leca I, Gstrein T, Hansen AH, Keays DA. Tubulins and brain development; The origins of functional specification. Mol Cell Neurosci 2017; 84: 58\u0026ndash;67.\u003c/li\u003e\n\u003cli\u003eRustici G, Kolesnikov N, Brandizi M, Burdett T, Dylag M, Emam I, et al. ArrayExpress update\u0026ndash;trends in database growth and links to data analysis tools. Nucleic Acids Res 2013; 41: D987-990.\u003c/li\u003e\n\u003cli\u003eVan Schoor E, Vandenbulcke M, Bercier V, Vandenberghe R, van der Zee J, Van Broeckhoven C, et al. Frontotemporal lobar degeneration case with an N-terminal TUBA4A mutation exhibits reduced TUBA4A levels in the brain and TDP-43 pathology. Biomolecules 2022; 12: 440.\u003c/li\u003e\n\u003cli\u003eMourier A, Ruzzenente B, Brandt T, K\u0026uuml;hlbrandt W, Larsson NG. Loss of LRPPRC causes ATP synthase deficiency. Hum Mol Genet 2014; 23: 2580\u0026ndash;2592.\u003c/li\u003e\n\u003cli\u003eCuillerier A, Honarmand S, Cadete VJJ, Ruiz M, Forest A, Desch\u0026ecirc;nes S, et al. Loss of hepatic LRPPRC alters mitochondrial bioenergetics, regulation of permeability transition and trans-membrane ROS diffusion. Hum Mol Genet 2017; 26: 3186\u0026ndash;3201.\u003c/li\u003e\n\u003cli\u003eZhang G, Li S, Cheng KW, Chou TF. AAA ATPases as therapeutic targets: Structure, functions, and small-molecule inhibitors. Eur J Med Chem 2021; 219: 113446.\u003c/li\u003e\n\u003cli\u003eMelliou S, Sangster KT, Kao J, Zarrei M, Lam KHB, Howe J, et al. Regionally defined proteomic profiles of human cerebral tissue and organoids reveal conserved molecular modules of neurodevelopment. Cell Rep 2022; 39: 110846.\u003c/li\u003e\n\u003cli\u003eWei N, Deng XW. The COP9 signalosome. Annu Rev Cell Dev Biol 2003; 19: 261\u0026ndash;286.\u003c/li\u003e\n\u003cli\u003eZheng Q, Huang T, Zhang L, Zhou Y, Luo H, Xu H, et al. Dysregulation of ubiquitin-proteasome system in neurodegenerative diseases. Front Aging Neurosci 2016; 8: 303.\u003c/li\u003e\n\u003cli\u003eLago SG, Bahn S. The druggable schizophrenia genome: from repurposing opportunities to unexplored drug targets. npj Genom Med 2022; 7: 25.\u003c/li\u003e\n\u003cli\u003eWainberg M, Sinnott-Armstrong N, Mancuso N, Barbeira AN, Knowles DA, Golan D, et al. Opportunities and challenges for transcriptome-wide association studies. Nat Genet 2019; 51: 592\u0026ndash;599.\u003c/li\u003e\n\u003cli\u003eLien Y-J, Hsiao P-C, Liu C-M, Faraone SV, Tsuang MT, Hwu H-G, et al. A genome-wide linkage scan for distinct subsets of schizophrenia characterized by age at onset and neurocognitive deficits. PLoS ONE 2011; 6: e24103.\u003c/li\u003e\n\u003cli\u003eBortolato B, Miskowiak KW, K\u0026ouml;hler CA, Vieta E, Carvalho AF. Cognitive dysfunction in bipolar disorder and schizophrenia: a systematic review of meta-analyses. Neuropsychiatr Dis Treat 2015; 11: 3111\u0026ndash;3125.\u003c/li\u003e\n\u003cli\u003eDeng M, Liu Z, Zhang W, Wu Z, Cao H, Yang J, et al. Associations between polygenic risk, negative symptoms, and functional connectome topology during a working memory task in early-onset schizophrenia. Schizophrenia 2022; 8: 54.\u003c/li\u003e\n\u003cli\u003eSada-Fuente E, Aranda S, Papiol S, Heilbronner U, Molt\u0026oacute; MD, Aguilar EJ, et al. Common genetic variants contribute to heritability of age at onset of schizophrenia. Transl Psychiatry 2023; 13: 201.\u003c/li\u003e\n\u003cli\u003eFabbri C, Leggio GM, Drago F, Serretti A. Imputed expression of schizophrenia-associated genes and cognitive measures in patients with schizophrenia. Mol Genet Genomic Med 2022; 10: e1942.\u003c/li\u003e\n\u003cli\u003eGordon A, Forsingdal A, Klewe IV, Nielsen J, Didriksen M, Werge T, et al. Transcriptomic networks implicate neuronal energetic abnormalities in three mouse models harboring autism and schizophrenia-associated mutations. Mol Psychiatry 2021; 26: 1520\u0026ndash;1534.\u003c/li\u003e\n\u003cli\u003eHu X, Qiao D, Kim W, Moll M, Balte PP, Lange LA, et al. Polygenic transcriptome risk scores for COPD and lung function improve cross-ethnic portability of prediction in the NHLBI TOPMed program. Am J Hum Genet 2022; 109: 857\u0026ndash;870.\u003c/li\u003e\n\u003cli\u003eLiang Y, Pividori M, Manichaikul A, Palmer AA, Cox NJ, Wheeler HE, et al. Polygenic transcriptome risk scores (PTRS) can improve portability of polygenic risk scores across ancestries. Genome Biol 2022; 23: 23.\u003c/li\u003e\n\u003cli\u003ePain O, Glanville KP, Hagenaars S, Selzam S, F\u0026uuml;rtjes A, Coleman JRI, et al. Imputed gene expression risk scores: a functionally informed component of polygenic risk. Hum Mol Genet 2021; 30: 727\u0026ndash;738.\u003c/li\u003e\n\u003cli\u003eHe T, Chan KCC. Measuring boundedness for protein complex identification in PPI networks. IEEE/ACM Trans Comput Biol Bioinform 2019; 16: 967\u0026ndash;979.\u003c/li\u003e\n\u003cli\u003eDear R, Wagstyl K, Seidlitz J, Markello RD, Arnatkevičiūtė A, Anderson KM, et al. Cortical gene expression architecture links healthy neurodevelopment to the imaging, transcriptomics and genetics of autism and schizophrenia. Nat Neurosci 2024; 27: 1075\u0026ndash;1086.\u003c/li\u003e\n\u003cli\u003eLiu SH, Du Y, Chen L, Cheng Y. Glial cell abnormalities in major psychiatric diseases: A systematic review of postmortem brain studies. Mol Neurobiol 2022; 59: 1665\u0026ndash;1692.\u003c/li\u003e\n\u003cli\u003eChen S, Tan Y, Tian L. Immunophenotypes in psychosis: is it a premature inflamm-aging disorder? Mol Psychiatry 2024; 29: 2834\u0026ndash;2848.\u003c/li\u003e\n\u003cli\u003eRuzicka WB, Mohammadi S, Fullard JF, Davila-Velderrain J, Subburaju S, Tso DR, et al. Single-cell multi-cohort dissection of the schizophrenia transcriptome. Science 2024; 384: eadg5136.\u003c/li\u003e\n\u003cli\u003eFanous AH, Kendler KS. Genetic heterogeneity, modifier genes, and quantitative phenotypes in psychiatric illness: searching for a framework. Mol Psychiatry 2005; 10: 6\u0026ndash;13.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Schizophrenia, Polygenic risk score, Gene expression risk scores, Early onset, Coexpression analyses, Network analysis","lastPublishedDoi":"10.21203/rs.3.rs-5833160/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5833160/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBACKGROUND\u003c/h2\u003e \u003cp\u003eDespite a high heritability of schizophrenia (SZ), only limited variance was attributed to gene loci or the polygenic risk score in genome-wide association studies (GWAS). Early-onset SZ, a more homogeneous SZ subtype, may aid in bridging the genotype-phenotype gap and the identification of its hub genes is critical for early intervention in clinical practice. We aimed to examine the gene expression risk score (GeRS) in patients from both multiplex and simplex families to identify hub genes for early-onset SZ, and perform enrichment analysis to understand the biological functions of the hub genes.\u003c/p\u003e\u003ch2\u003eMETHODS\u003c/h2\u003e \u003cp\u003eBased on the GWAS genotype data from patients with SZ in multiplex families (223 early-onset and 372 late-onset) and those from simplex families (matched for sex and onset age), GeRSs for SZ (SZ-GeRSs) were estimated using the SNP-expression prediction model derived from existing brain tissues of patients with psychiatric disorders. Module-based SZ-GeRS was summed over genes from empirically derived gene clusters, network analysis was conducted to identify hub genes, and enrichment analysis was used for functional mapping.\u003c/p\u003e\u003ch2\u003eRESULTS\u003c/h2\u003e \u003cp\u003eAmong the 13 modules from existing coexpression analyses of \u003cem\u003epostmortem\u003c/em\u003e brains of patients with psychiatric disorders, the meta-analysis revealed that associations with early-onset SZ existed for the GeRS of module 10 in subset, M10\u003csub\u003esub\u003c/sub\u003e-GeRS (adjusted odds ratio [aOR]\u0026thinsp;=\u0026thinsp;1.38, 95% CI\u0026thinsp;=\u0026thinsp;1.22\u0026ndash;1.57), and six hub genes, M10\u003csub\u003ehub\u003c/sub\u003e-GeRS (aOR\u0026thinsp;=\u0026thinsp;1.22, 95% CI\u0026thinsp;=\u0026thinsp;1.07\u0026ndash;1.39), after adjustment for covariates. Functional mapping of the genes revealed their enrichment in excitatory neurons and immune-regulatory pathways.\u003c/p\u003e\u003ch2\u003eCONCLUSIONS\u003c/h2\u003e \u003cp\u003eGeRS for SZ helps identify six hub genes for early-onset schizophrenia, and the enrichment analysis sheds light on their possible roles in the pathophysiology. These findings will enhance the understanding of SZ etiology and may contribute to early screening and personalized prevention efforts.\u003c/p\u003e","manuscriptTitle":"Identification of Hub Genes Involved in Early-onset Schizophrenia: From Genetic Susceptibility to Predicted Regulated Gene Expression","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-01-20 15:52:27","doi":"10.21203/rs.3.rs-5833160/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"bf92d1a6-52e6-47a9-a150-b85d7d944e5b","owner":[],"postedDate":"January 20th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-01-22T09:39:16+00:00","versionOfRecord":[],"versionCreatedAt":"2025-01-20 15:52:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5833160","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5833160","identity":"rs-5833160","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}