Suhua Chang

Attention deficit hyperactivity disorder (ADHD) is a common neurodevelopmental disorder modulated by genetic and environmental factors. Approximately 20-55% of individuals with ADHD experience concurrent sleep disturbances, yet the dynamic relationship between the two and their shared genetic mechanisms remain unclear. This study analyzed 11,288 participants (52.1% male) from the Adolescent Brain Cognitive Development Study using the Sleep Disturbance Scale for Children. We found a robust positive correlation between ADHD and sleep traits, though ADHD medication weakened this relationship. Latent growth curve modeling revealed that ADHD symptoms significantly impacted sleep traits with a lag effect. Polygenic risk score analyses showed significant but very small associations (explained variance ~0.28-0.30%). Pleiotropic analyses identified two loci associated with ADHD and excessive somnolence (implicating 30 genes), and cross-cohort comparisons with UK Biobank insomnia data uncovered 12 additional pleiotropic loci (98 genes) enriched in brain tissue. Functional enrichment analysis highlighted excitatory and inhibitory neuronal subtypes in visual and frontal cortices, suggesting roles in synaptic transmission and neural circuit development. Protein-protein interaction network analysis identified 10 hub genes involved in cell growth regulation and neurodevelopmental processes. This study demonstrates robust ADHD-sleep associations, with medication moderating this link, and provides preliminary insights into their shared genetic architecture.

Autism spectrum disorder (ASD) is a highly heterogeneous neurodevelopmental condition with a complex genetic architecture. While over a thousand risk genes have been cataloged, a fundamental challenge remains how this vast genetic landscape translates into diverse clinical manifestations. To address this, we propose a "many-to-few" framework, shifting from traditional "many-to-one" convergence model toward an intrinsic organizational architecture where disparate ASD risk genes funnel into discrete molecular dimensions. By leveraging similarity network fusion (SNF) to integrate bulk and single-nucleus RNA sequencing data, we decomposed the 311 ASD reliable risk genes into three stable, spatiotemporally distinct molecular subtypes. These subtypes represent coordinated expression programs that integrate functionally diverse genes, such as those involved in synaptic signaling, mRNA stabilization, and histone modification. Mapping de novo variants from the SPARK cohort onto these subtypes enabled stratification of probands into three genetically defined subgroups (S1-S3) with divergent clinical profiles. Transcriptomic decomposition partitioned the 311 reliable risk genes into three stable molecular subtypes, Synaptic Signaling (C1), mRNA Stabilization (C2), and Histone Modification (C3), exhibiting divergent spatiotemporal trajectories and cell-type-specific enrichment. Patient subgroups stratified based on these molecular subtypes displayed significant differences in adaptive functioning, core symptoms, and psychiatric comorbidities, while a reference group (S4) lacking de novo variants of these highly reliable ASD risk genes exhibited the most preserved functions. Furthermore, diverging rare and common genetic liability profiles across subgroups, particularly between S1 and S3, provide empirical support for a molecular subtype-based liability threshold model. Our study establishes a biologically informed framework that links intrinsic molecular subtypes to multidimensional phenotypic constellation, advancing mechanistic insight and offering translational potential for precision stratification and intervention.

Autism spectrum disorder (ASD) is highly heterogeneous. Molecular subtyping of ASD based on gene expression data will help address the biological heterogeneity of ASD. We analyze RNA-seq data from 1711 ASD samples, employing unsupervised non-negative matrix factorization (NMF) to identify subtypes, and characterize them with five categories: overall ASD symptoms, social-communication deficits, restricted-repetitive-stereotyped behaviors, cognitive-adaptive function and behavior problems. We also explore subtype-specific differentially expressed genes, functional pathways, and spatiotemporal/cell-type characteristics. NMF analysis identifies three ASD subtypes. The three subtypes exhibit different phenotypic patterns: Cluster 1 has severe restricted and repetitive behaviors, Cluster 3 mainly has social communication impairments, Cluster 2 presents milder symptoms and better cognitive function. Differential expressed genes analysis links Cluster 1 and Cluster3 to nervous system dysfunctions and morphogenesis of a branching structure, while Cluster 2 to immune system processes. The clusters also exhibit distinct gene expression patterns across brain regions, developmental stages, and cell types. Validation in two independent datasets confirms the reproducibility of the distinct clusters. This molecular subtyping reveals biologically and clinically distinct subtypes of ASD, helping to understand the atypical mechanisms underlying ASD heterogeneity. It also contributes to the development of personalized therapeutic strategies for distinct subtypes.

Neuropsychiatric disorders exhibit complex polygenic architectures, yet the cell-type-specific mechanisms underlying most risk loci remain unclear. Here, we integrate single-cell expression quantitative trait locus (sc-eQTL) data from brain and blood tissues with genome-wide association studies (GWAS) of six neuropsychiatric disorders (schizophrenia (SCZ), Parkinson’s disease (PD), bipolar disorder (BP), major depressive disorder (MDD), attention-deficit/hyperactivity disorder (ADHD), and autism spectrum disorder (ASD)) to systematically identify putative causal genes at cellular resolution. Employing summary-data-based Mendelian randomization (SMR) across diverse neuronal and immune cell types, we discovered 345 cell-type-specific risk genes for various diseases, including both replicated candidates (such as MAPT in astrocytes for SCZ and PD and FLOT1 in excitatory neurons and inhibitory neurons for SCZ, BP and MDD) and novel associations (such as APTX in microglia for SCZ). Cross-disorder analyses revealed shared pathways in synaptic function and immune regulation. In contrast, disease-specific and tissue-specific patterns were observed across different disorders. Strikingly, we found that brain-derived risk genes exhibited significantly higher cell-type specificity than those identified in blood, underscoring the more focused cellular context of genetic effects in the central nervous system. Our findings suggest that neuropsychiatric disorders arise from a combination of neuronal dysfunction and immune system dysregulation. The study demonstrates how cell-type-specific mapping uncovers etiological mechanisms obscured in bulk-tissue analyses, proving novel information for clarifying the biological mechanism of gene expression implicated in the development of the six neuropsychiatric disorders.

Attention deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder characterized by genetic predisposition and alterations in brain structural connectivity. While existing studies have established associations between genetic variants and neuroanatomical features, the specific relationships in ADHD remained poorly understood. To address this gap, we developed adversarial deep canonical correlation analysis models (A-DCCA) to disentangle ADHD-specific and non-specific “gene-white matter” association patterns. Utilizing diffusion tensor imaging and genotype data from six-hundred ADHD and typically developed children in a Chiese cohort, the current study revealed ADHD-specific correlations between the right cerebral peduncle, right posterior limb of the internal capsule, and genes regulating neural apoptotic processes (CAMK1D, METTL15, and MAP2K4). In contrast, associations involving the left cerebral peduncle, left posterior limb of the internal capsule, right superior longitudinal fasciculus, and right posterior thalamic radiation with genes related to early neural development (FYN, PHF2, ZSCAN31, and CD82) presented associations shared by ADHD and non-ADHD groups. Incorporating interpretable deep learning models, the current study unveiled white matter regions vulnerable to genetic influences in ADHD-specific and non-specific ways, shedding light on the understanding of biological substrates of ADHD.

Rapid industrialization has introduced a range of chemicals into the environment, posing significant risks to fetal and child brain development. Using the Comparative Toxicogenomics Database (CTD), we constructed chemical exposome frameworks for seven neurodevelopmental disorders (NDDs) and identified chemical pollutants of epidemiological concern, including air pollutants (n = 8), toxic elements (n = 14), pesticides and related compounds (n = 18), synthetic organic chemicals (n = 16), and solvents (n = 5). Gene set enrichment analysis validated and revealed significant toxicogenomic associations between these chemical pollutants and NDDs, including autism spectrum disorder (ASD) (12 pollutants, proportional reporting ratio (PRR) 3.56–7.21) and intellectual disability (ID) (9 pollutants, PRR 3.13–5.59). Functional annotation of pollutant-specific gene sets highlighted shared biological processes, such as metabolic processes (e.g., xenobiotic metabolic process, xenobiotic catabolic process, and cytochrome P450 pathway) for ASD and cognitive processes (e.g., cognition, social behavior, and synapse assembly) for ID (Bonferroni-corrected p-values < 0.05). Time trajectory analysis of developmental transcriptomic data from the BrainSpan database for ASD (275 genes) and ID (93 genes) revealed three distinct expression patterns of chemical-pollutant-associated genes—higher prenatal, postnatal, and perinatal expression—indicating common and divergent underlying mechanisms across critical windows of chemical pollutant exposure.

Although the efficacy of pharmacy in the treatment of attention deficit/hyperactivity disorder (ADHD) has been well established, the lack of predictors of treatment response poses great challenges for personalized treatment. The current study employed a comprehensive approach, combining genome-wide association analyses (GWAS) and deep learning (DL) methods, to elucidate the genetic underpinnings of pharmacological treatment response in ADHD. Based on genotype data of medication-naïve patients with ADHD who received pharmacological treatments for 12 weeks, the current study performed GWAS using the percentage changes in ADHD-RS score as phenotype. Then, DL models were constructed to predict percentage changes in symptom scores using genetic variants selected based on four different genome-wide P thresholds (E-02, E-03, E-04, E-05) as inputs. The current GWAS results identified two significant loci (rs10880574, P = 2.39E-09; rs2000900, P = 3.31E-09) which implicated two genes, TMEM117 and MYO5B, that were primarily associated with both brain- and gut-related disorders. The convolutional neural network (CNN) model, using variants with genome-wide P values less than E-02 (5516 SNPs), demonstrated the best performance with mean squared error (MSE) equals 0.012 (Accuracy = 0.83; Sensitivity = 0.90; Specificity = 0.75) in the validation dataset, 0.081 in an independent test dataset (Acc = 0.61, Sensitivity = 0.81; Specificity = 0.26). Notably, the variant that contributed most to the CNN model was NKAIN2, an ADHD-related gene, which is also associated with metabolic processes. To conclude, the integration of GWAS and DL methods revealed new genes contribute to ADHD pharmacological treatment responses, and underscored the interplay between neural systems and metabolic processes, potentially providing critical insights into precision treatment. Furthermore, our CNN model exhibited good performance in an independent dataset, encouraged future studies and implied potential clinical applications.