Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Multiomics integration”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4Linked to original sources
PubMed2026

Integrated Genomic and Epigenomic Analysis Reveals Epigenetic Plasticity in Disease Progression and Multidrug Resistance in Multiple Myeloma.

UNLABELLED: Multiple myeloma is marked by recurrent cytogenetic abnormalities and mutations that accumulate as the disease progresses. In this study, we sought to elucidate the transitions driving tumorigenesis and therapy resistance in multiple myeloma using a unique cohort of nearly 900 patients spanning premalignant to late-stage refractory multiple myeloma, comprehensively characterized at molecular and clinical levels. Waves of epigenetic dysregulation drove these critical transitions. In this paradigm, genomic and cytogenetic events unlocked epigenetic plasticity, reshaping multiple myeloma cell biology to evade tumor microenvironment constraints and therapeutic pressures. Functional perturbation studies in an isogenic proteasome inhibitor-resistant cell line model demonstrated enhanced reliance on transcriptional cofactors, supporting a mechanistic link between chromatin plasticity and therapy adaptation. Collectively, these findings support a unifying framework in which genomic heterogeneity unlocks gene regulatory plasticity, enabling plasma cells (PC) to evade microenvironmental constraints and therapeutic pressure. These results provide a mechanistic explanation for sequential relapse without new genomic alterations and nominate epigenetic plasticity-mediated PC adaptation as a therapeutic vulnerability in the heterogeneous genetic background of multiple myeloma. SIGNIFICANCE: Assembly and analysis of a multiple myeloma cohort spanning the continuum from premalignant to late relapse that integrates bulk transcriptomics with single-cell multiomic data provides insights into disease progression and epigenetic plasticity.

Multiple Myeloma↗
PubMed2026

Molecular biomarker profiling in noninfectious uveitis: a chronological review of discovery.

PURPOSE OR REVIEW: Noninfectious uveitis (NIU) encompasses a heterogeneous group of immune-mediated intraocular inflammatory diseases whose complexity has driven systematic molecular biomarker discovery. This review presents NIU molecular biomarkers organized by biological category; autoantigens, human leukocyte antigens (HLA) and genetic markers, cellular immune subsets, cytokines, chemokines, and multiomics platforms including proteomics, microbiome metagenomics, metabolomics, and single-cell transcriptomics with each category presented in strict chronological order of landmark discovery. RECENT FINDINGS: We present a review organized along two nested timelines. Categories are presented in the order they historically emerged in the field, and within each category, landmark discoveries appear in chronological sequence. This allows the reader to trace how each biomarker category evolved: from foundational autoantigen identification in experimental uveitis models, through the genomic revolution of HLA association studies, into cellular immunophenotyping, cytokine profiling of aqueous humor, chemokine mapping of intraocular trafficking, and finally the emerging omics platforms that may potentially anchor precision medicine in NIU. Each biomarker is paired in line with its linked targeted therapeutic. SUMMARY: Biomarker research has transformed the understanding of NIU from a clinically defined syndrome into a group of molecularly distinct immune disorders. Advances spanning autoantigens, genetics, immune-cell profiling, cytokines, chemokines, and multiomics have revealed novel pathogenic mechanisms and therapeutic targets. Integration of these biomarkers with targeted therapies may accelerate the transition toward precision medicine in uveitis care.

cytokines↗
PubMed2025

Colorectal Liver Metastasis Pathomics Model: Integrating Single-Cell and Spatial Transcriptome Analysis With Pathomics for Predicting Liver Metastasis in Colorectal Cancer.

The liver is the primary target organ for hematologic metastasis of colorectal cancer (CRC), and CRC liver metastasis (CRLM) often precludes radical resection, making it the leading cause of death in patients with CRC. To improve the identification and prediction of liver metastasis risk, we identified a cell type of liver metastasis--triggering malignant cells (LMTMCs) through integrating single-cell RNA sequencing and spatial transcriptome analysis. Multiomics cell communication analysis indicated that the interaction between fibroblasts and LMTMCs through the COL1A1-CD44/SDC4 and LAMA4-CD44 signaling axes could promote CRLM. By applying the one-class logistic regression algorithm, we developed a CRLM scoring system in the bulk RNA-sequencing data according to the abundance of LMTMCs in each individual. Using the grouping labels derived from the CRLM scoring system in the bulk data and the corresponding whole-slide images without any manual annotations at the region or pixel level, processed via slide-level weakly supervised learning, a deep-learning model based on the ResNet18 architecture, called Colorectal Liver Metastasis Pathomics Model, was developed to predict the risk of liver metastasis in patients with CRC. The Colorectal Liver Metastasis Pathomics Model achieved an area under the curve of 0.84 at the internal test set of The Cancer Genome Atlas-CRC histology images. In the external independent validation sets, namely the Affiliated Hospital of Southwest Medical University and the Affiliated Traditional Chinese Medicine Hospital of Southwest Medical University cohorts, the areas under the curve were 0.89 and 0.72, respectively, indicating effective classification performances. This study provided new insights and tools for the early identification of CRLM and demonstrated the potential of combining multiomics with deep learning-based pathomics in cancer research.

Humans↗
PubMed2026

Hox/Meis-dependent gene-regulatory transition underlies cardiopharyngeal neural crest diversification.

Neural crest cells (NCCs) are multipotent migratory cells essential for cardiac development, yet the lineage trajectories and gene regulatory networks underlying their differentiation in the cardiopharyngeal region remain unclear. Here, we integrate single-cell RNA-seq, spatial transcriptomics, and multiomic analyses to construct a comprehensive map of NCC lineages in developing mouse cardiopharyngeal tissues. We identify a transition from Hox-positive pharyngeal NCCs to Hox-negative intracardiac populations associated with the outflow tract cushion, accompanied by a shift in Meis transcription factor binding and gene-regulatory network architecture. By contrast, NCCs forming the aorticopulmonary septum and great vessel smooth muscle retain distinct Hox-codes. A Meis2-Sox9-Scx gene-regulatory network defines a skeletogenic progenitor-like intermediate state that gives rise to coronary artery smooth muscle and semilunar valves. Our findings suggest that the loss of Hox-dependent regional identity enables pharyngeal NCCs to acquire new fates upon entering the cardiac cushion, providing insight into the developmental origins of coronary and valvular calcification.

Journal Article↗
PubMed2026

Strategy for Simultaneous Multiomic Survey of N-Glycomic and Extracellular Matrix Proteome by Mass Spectrometry Imaging.

Recent advances in spatially resolved molecular profiling have positioned matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) as a powerful platform for multiomic tissue analyses. However, conventional workflows that sequentially target distinct molecular classes are time- and resource-intensive, requiring repeated sequential sample preparation, imaging, and data integration. Here, we evaluate streamlined strategies for simultaneous or combined acquisition of N-glycan and collagen-derived peptide information using PNGase F and collagenase. In-solution studies demonstrate that simultaneous enzymatic digestion yields comparable peptide identifications and glycan profiles relative to traditional sequential workflows, with minimal impact on enzymatic specificity. On the basis of these findings, we developed and optimized MALDI-MSI protocols enabling either simultaneous enzyme application or sequential enzyme treatment with unified matrix deposition and single-pass imaging. While direct coapplication reduced image uniformity, a hybrid approach that used sequential enzyme deposition with combined imaging preserved spatial fidelity and spectral quality while significantly reducing processing and computational demands. Application to human tissues, including vertebral bone and ocular samples, highlights the utility of this workflow for fragile specimens and exploratory multiomic surveys. Collectively, these results establish a framework for integrated glycomic and proteomic imaging targeting the extracellular microenvironment, expanding multiomic MALDI-MSI analyses.

Spectrometry, Mass, Matrix-Assisted Laser Desorpti↗
PubMed2026

Multimodal Integration of Protein Interactomes With Genomic and Molecular Data Discovers Distinct Rheumatoid Arthritis Endotypes.

OBJECTIVE: Rheumatoid arthritis (RA) is a heterogeneous autoimmune disease characterized by clinical and molecular heterogeneity, notably in the presence of anti-cyclic citrullinated peptide (CCP) antibodies. Patients with CCP+ RA exhibit more severe disease progression and distinct treatment responses compared to patients with CCP- RA. Although previous studies have investigated cellular and molecular differences between these subtypes, their genetic differences are understudied. METHODS: We leveraged the Rheumatoid Arthritis Comparative Effectiveness Research cohort, comprising 555 patients with CCP+/rheumatoid factor (RF)+ RA and 384 patients with CCP-/RF+ RA. Using a novel framework, we integrated a network-based genome-wide association study (GWAS) with multiomic data to uncover corresponding genetic and molecular differences. RESULTS: We uncovered a significant heritability difference between these disease groups. Network-based GWAS uncovered 14 putative gene modules, including many genes outside the HLA loci, that explained genetic differences between CCP+/RF+ and CCP-/RF+ RA. Heritability partitioning and multivariate expression analyses validated four modules, highlighting novel genetic loci underlying phenotypic differences. Module functional significance was established using multiple orthogonal cohorts, underscoring their biologic relevance. CONCLUSION: Our findings demonstrate the use of network-based approaches in revealing differential genetic risk factors underlying CCP+/RF+ and CCP-/RF+ RA. Disease-associated gene modules detected in synovial tissue were also observed in peripheral blood, indicating joint-specific molecular programs are reflected systemically. This cross-tissue concordance highlights the potential for blood-based assays to capture pathogenic mechanisms active in the joints, enabling practical patient stratification. Our findings highlight why patients with CCP+/RF+ and CCP-/RF+ RA exhibit distinct clinical courses and therapeutic responses, supporting precision-guided treatment strategy development in RA.

Humans↗
PubMed2026

Integrative Pan-Cancer Characterization of lncRNA UPK1A-AS1 and Its Role in Hypoxia-Associated Sorafenib Resistance in Hepatocellular Carcinoma.

Long noncoding RNAs (lncRNAs) are emerging as critical regulators of tumor initiation and progression through transcriptional and posttranscriptional mechanisms. UPK1A antisense RNA 1 (UPK1A-AS1), a cancer-associated lncRNA, has been reported to participate in oncogenic processes; however, its overall landscape across human malignancies and its biological role in therapy resistance remain poorly understood. Given the increasing importance of identifying functional lncRNAs with prognostic and therapeutic potential, this study presents a comprehensive multiomics characterization of UPK1A-AS1 and its experimental validation in hepatocellular carcinoma (HCC). We integrated datasets from The Cancer Genome Atlas (TCGA), the Genotype-Tissue Expression Project (GTEx), the cancer immunology data engine (CIDE), and the cBioPortal for cancer genomics (cBioPortal) to systematically assess its expression pattern, genomic alterations, clinical significance, and immunological associations. Our analyses revealed that UPK1A-AS1 is significantly upregulated in multiple tumor types, with copy-number amplification as the predominant genomic alteration driving its overexpression. Elevated UPK1A-AS1 expression was correlated with advanced disease stage, poor differentiation, immune exclusion, and unfavorable prognosis, supporting its potential as a cancer type-dependent biomarker. In parallel, functional studies demonstrated that hypoxia transcriptionally induces UPK1A-AS1 in HCC, where it promotes sorafenib resistance by suppressing apoptosis. Silencing UPK1A-AS1 restored apoptotic and enhanced sorafenib efficacy both in vitro and in vivo. Collectively, our findings suggest that UPK1A-AS1 is a hypoxia-inducible oncogenic lncRNA that plays dual roles in cancer, with cancer type-dependent associations with progression and immune modulation across malignancies and mechanistically mediating hypoxia-associated drug resistance in HCC.

Humans↗
PubMed2025

Integrating multi-omics approaches in acute myeloid leukemia (AML): Advancements and clinical implications.

Acute myeloid leukemia (AML) is a highly heterogeneous and aggressive hematologic malignancy characterized by clonal proliferation of myeloid precursors. Despite significant advancements in genomic profiling and targeted therapies, patient outcomes remain suboptimal due to disease complexity, resistance mechanisms, and high relapse rates. The integration of multi-omics approaches-spanning genomics, epigenomics, transcriptomics, proteomics, and metabolomics-has revolutionized AML research, offering a comprehensive understanding of leukemogenesis, tumor heterogeneity, and therapeutic vulnerabilities. Recent studies leveraging high-throughput sequencing, mass spectrometry, and advanced computational tools have uncovered novel biomarkers, clonal evolution dynamics, and microenvironmental interactions that drive AML progression and resistance. For instance, single-cell multi-omics has revealed chemotherapy-resistant leukemic stem cell populations, while proteogenomic analyses have identified actionable targets such as MCL1 and metabolic dependencies like OXPHOS. Clinically, integrated omics platforms are refining risk stratification, minimal residual disease (MRD) monitoring, and personalized therapy selection. However, challenges such as data integration complexity, cost barriers, and ethical considerations remain. This review highlights the transformative potential of multi-omics in AML, emphasizing recent advancements in technology, biomarker discovery, and therapeutic innovation. By bridging the gap between molecular insights and clinical practice, multi-omics integration promises to redefine AML management, paving the way for precision oncology and improved patient outcomes.

Humans↗
PubMed2024

Understanding the biological processes of kidney carcinogenesis: an integrative multi-omics approach.

Biological mechanisms related to cancer development can leave distinct molecular fingerprints in tumours. By leveraging multi-omics and epidemiological information, we can unveil relationships between carcinogenesis processes that would otherwise remain hidden. Our integrative analysis of DNA methylome, transcriptome, and somatic mutation profiles of kidney tumours linked ageing, epithelial-mesenchymal transition (EMT), and xenobiotic metabolism to kidney carcinogenesis. Ageing process was represented by associations with cellular mitotic clocks such as epiTOC2, SBS1, telomere length, and PBRM1 and SETD2 mutations, which ticked faster as tumours progressed. We identified a relationship between BAP1 driver mutations and the epigenetic upregulation of EMT genes (IL20RB and WT1), correlating with increased tumour immune infiltration, advanced stage, and poorer patient survival. We also observed an interaction between epigenetic silencing of the xenobiotic metabolism gene GSTP1 and tobacco use, suggesting a link to genotoxic effects and impaired xenobiotic metabolism. Our pan-cancer analysis showed these relationships in other tumour types. Our study enhances the understanding of kidney carcinogenesis and its relation to risk factors and progression, with implications for other tumour types.

Kidney Neoplasms↗
PubMed2024

Integrative omics analysis identifies biomarkers of septic cardiomyopathy.

Septic Cardiomyopathy (SCM) is a syndrome of acute cardiac dysfunction in septic patients, unrelated to cardiac ischemia. Multiomics studies including transcriptomics and proteomics have provided new insights into the mechanisms of SCM. In here, a rat model of SCM was established by intraperitoneal injection of lipopolysaccharide (LPS). Biomarkers of SCM were characterized via a multi-omics analysis. The differentially expressed (DE) mRNAs predominantly appeared in pathways linked to the immune response, inflammatory response, and the complement and coagulation cascades, while DE proteins were mainly enriched in pathways associated with the complement and coagulation cascades. On this basis, the integrated analysis was performed between transcriptome and proteome. The potential biomarkers were further verified by RT-qPCR and WB. The current proteotranscriptomic research has furnished a valuable dataset and fresh perspectives that will enhance our comprehension of the development of SCM. This, in turn, is expected to expedite the formulation of novel approaches for the prevention and management of SCM in patients.

Cardiomyopathies↗
PubMed2026

Comprehensive analysis of metabolomics and transcriptomics of radiation-induced rectal injury.

Radiation-induced rectal injury (RRI) significantly affects the quality of life in patients with locally advanced rectal cancer (LARC) undergoing neoadjuvant chemoradiotherapy (NCRT). Non-targeted liquid chromatography-mass spectrometry metabolomics analysis and transcriptomic analysis were conducted to explore RRI characteristics. Hematoxylin-eosin and Masson staining confirmed radiation-induced injury in rectal tissue within the radiotherapy target region. Orthogonal partial least squares discriminant analysis identified 823 differentially expressed metabolites (DEMs). Transcriptomic analysis revealed 400 differentially expressed genes (DEGs). Enrichment analysis revealed that DEMs and DEGs were primarily involved in metabolic, immune, and signal transduction pathways. Integrated analysis demonstrated significant enrichment of DEMs and DEGs in the arachidonic acid metabolism pathway. Pearson's correlation and canonical correlation analyses were used to assess the association between DEMs and DEGs within this pathway. In conclusion, this study identified key biological regulatory pathways involved in RRI through a multi-omics approach, offering potential targets for its diagnosis and treatment.

Humans↗
PubMed2026

Integrating Next-Generation Sequencing into von Willebrand Disease Diagnostics: Insights from the PCM-EVW-ES Multicenter Project.

Von Willebrand disease (VWD) is the most common inherited bleeding disorder, caused by quantitative or qualitative defects in von Willebrand factor (VWF). Diagnosis is challenging and requires integrating bleeding history, VWF antigen and activity measurements, FVIII assays, and specialized phenotyping. Genetic testing is increasingly recognized as a key component. Here, we review current concepts in VWD diagnostics and highlight the Spanish Clinical and Molecular Profile of von Willebrand Disease (PCM-EVW-ES) project as a model for genomics-enabled precision medicine. PCM-EVW-ES is a multicenter initiative involving 48 hospitals, centralized phenotypic testing, and next-generation sequencing of the VWF coding region, enabling definitive classification in 730 individuals with VWD to date. Harmonized recruitment criteria and standardized workflows improve subtype assignment, uncover complex genotypes, refine genotype-phenotype correlations, and facilitate the identification of asymptomatic carriers. The PCM-EVW-ES variant spectrum highlights recurrent disease-causing variants in Spain and underscores the value of coordinated national registries for variant curation. Building on these data, we propose a diagnostic algorithm in which bleeding assessment and first-line VWF/FVIII assays, combined with, early VWF molecular testing increases diagnostic accuracy and guides targeted second-line investigations to confirm and refine VWD subtype classification. We also outline persisting challenges, including the interpretation of variants of uncertain significance and patients without identifiable pathogenic VWF variants, and future directions integrating third-generation sequencing, expanded gene panels, functional studies, and artificial-intelligence-driven multiomic approaches. Together, these advances illustrate how robust multicenter studies can bridge the gap between complex diagnostics and clinical practice in VWD.

Humans↗
PubMed2026

Integrative Long-Read Multi-Omics of a Patient With GPI Deficiency: A Molecular Case Study of a Candidate Dual-Effect GPI Variant.

The molecular determinants of phenotypic severity in red cell enzymopathies are often obscured by the disconnect between coding sequence variants and their regulatory landscapes. Here we present a single-patient molecular case study that uses an integrative multi-omic approach-combining short-read WGS, PacBio HiFi long-read sequencing, native CpG methylation profiling, and Iso-Seq full-length transcriptomics-to characterize a severe, transfusion-dependent hemolytic anaemia. We identified a compound heterozygous state in the glucose-6-phosphate isomerase (GPI) gene, with no wild-type allele present. One allele (Haplotype 1) carried a missense variant (p.His191Arg); the other (Haplotype 2) carried a distinct missense variant, c.1414C>T (p.Arg472Cys), previously reported as biochemically unstable. Long-read phasing placed the two variants in trans. Allele-resolved transcript counts showed a directionally consistent but statistically non-significant trend toward higher expression of Haplotype 2 across two Iso-Seq replicates. Notably, the c.1414C>T transition abolishes a local CpG dinucleotide; in a small number of haplotype-2 reads spanning this position, the corresponding cytosine on the wild-type/Haplotype-1 background was methylated. We did not measure GPI protein abundance, enzymatic activity, or stability in this patient, and we do not establish that methylation at this site regulates GPI transcription. On the basis of these correlative observations in a single patient, we propose-as a hypothesis for future testing-that a coding variant might simultaneously perturb protein stability and disrupt a local epigenetic mark, and we outline the experiments required to test whether such a dual effect contributes to disease. This case illustrates the value of integrative long-read multi-omics for generating mechanistic hypotheses about variants of uncertain significance, while underscoring that causal claims require dedicated functional validation.

Humans↗
PubMed2026

Comparative metabolomic and transcriptomic profiling of flavonoid diversity and antioxidant capacity in three Isatis species.

Flavonoids are key bioactive compounds in plants with significant health benefits. This study employs an integrated multi-omics approach to investigate flavonoid diversity and antioxidant capacity across three Isatis species: I. oblongata, I. tinctoria, and I. indigotica. Metabolomic profiling identified 200 flavonoids, with glycosides being the most abundant class. I. tinctoria exhibited the highest total flavonoid content and antioxidant activity, strongly correlated with the accumulation of 53 core differential flavonoid metabolites, most of which were glycosylated derivatives. Transcriptomic analysis revealed coordinated upregulation of phenylpropanoid pathway genes and specific UDP-glycosyltransferases (UGTs) in I. tinctoria, providing a genetic basis for its enhanced glycoside production. The study establishes a clear genotype-metabolite-phenotype linkage, highlighting glycosylation as a key mechanism underlying flavonoid-driven antioxidant superiority in Isatis. Although the current evidence is primarily correlative, the consistent and strong associations across independent transcriptomic, metabolomic, and antioxidant datasets provide a robust foundation for this conclusion. These findings offer new insights into the metabolic evolution and regulatory networks of flavonoids, with implications for breeding and metabolic engineering of high-value medicinal plants.

Flavonoids↗
PubMed2026

The Health Benefits of Exercise: Molecular and Cellular Mechanisms.

Exercise is a low-cost lifestyle intervention that can prevent and alleviate various diseases. It is a potent physiological stimulus that activates conserved molecular signaling pathways. Through the coordinated integration of multiple molecules, pathways, and systems, it leads to systemic health benefits. However, most studies focus on individual systems or molecular mechanisms, lacking systematic integration of the cross-system regulation induced by exercise. We summarize the molecular mechanisms of exercise in the musculoskeletal, cardiovascular, nervous systems, among others. Exercise induces the release of exerkines (e.g., irisin, interleukin-6, and brain-derived neurotrophic factor) and extracellular vesicles, which activate key signaling pathways to enhance mitochondrial function, metabolism and physiological adaptation, while suppressing inflammation and oxidative stress, thereby alleviating diseases and delaying aging through cross-system coordination. We further explore exercise-induced adaptive regulation in extreme environments, including microgravity, hyperbaria, and hypoxia, offering a multifaceted perspective on organismal health regulation. Finally, we outline the prospects and challenges of multiomics, artificial intelligence-driven precision medicine, personalized exercise prescriptions, and exercise mimetics. Overall, this review provides a more integrated perspective on the molecular basis of exercise and offers directions for future mechanistic and translational studies.

exercise↗
PubMed2026

An integrated single-cell and spatial proteotranscriptomics atlas of fibroblast-driven immunoregulation within the human adult oral cavity.

The immunoregulatory architecture of human oral tissues remains poorly defined. We present an integrated single-cell and spatial proteotranscriptomic atlas profiling >250,000 single-cell transcriptomes and >4 million spatially resolved cells across 13 niches. Using our AI-enabled AstroSuite, we defined neighborhoods and interaction modules, revealing peri-epithelial fibroblast-centered hubs enriched in effector cytokines. We harmonized fibroblast subtypes (universal, immune, peri-epithelial, peri-vascular, peri-neural, antigen-presenting cell [APC]-like, stress responsive, and myofibroblasts) with stress-responsive subtypes partitioning between mucosae and glands (type I and II). Spatial multiomics mapped ligand-receptor programs and identified mucosal stress-responsive fibroblasts as putative immunoregulatory hubs. Niche-aware integration of healthy and diseased datasets revealed fibroblast rewiring into inflammatory and reparative niches. Disease neighborhoods exhibited expansion of major histocompatibility complex (MHC)-I+, MHC-II+, and programmed cell death ligand 1 (PD-L1)+ fibroblasts and predicted spatial engagement with T cells at tertiary lymphoid structures. Together, this atlas identifies fibroblasts as central regulators of structural immunity and provides a scalable framework to target stromal-immune interactions across barrier organs.

Journal Article↗
PubMed2026

[Applications and Challenges of Deep Learning in Human Genome Research].

In recent years, the advent of high-throughput omics technologies has fueled an explosive growth in human genomic data. Uncovering the latent functions within this vast data has become a significant challenge in functional genomics research. While traditional statistical methods have proved successful for analyzing smaller-scale datasets in the past, they exhibit clear limitations in analytical efficiency and integrating multi-dimensional data, struggling to meet the escalating demands of contemporary genomic analysis. The introduction of deep learning (DL) technologies offers a novel paradigm for this field. This review systematically examines the advances in applying deep learning to human genomics research. Studies demonstrate that when ample labeled data is available, discriminative DL computational methods-such as Convolutional Neural Networks (CNNs) and Long Short-Term Memory networks (LSTMs)-achieve high accuracy and efficiency in genomic variant discovery tasks. Furthermore, generative DL methods, particularly Large Language Models (LLMs) leveraging self-supervised pre-training strategies, effectively integrate complex genomic information and exhibit superior performance in functional genomic sequence annotation and gene regulation studies. This review also explores the application of LLMs in multi-omics data integration and prediction. Looking ahead, the continued accumulation of long-read sequencing and high-dimensional data is expected to enable DL technologies to integrate increasingly complex and heterogeneous genomic information, playing an increasingly crucial role in human genomics research.

Deep Learning↗
PubMed2026

Multiomic single-nucleus profiling reveals cell-type-specific epigenetic and transcriptional dysregulation in major depressive disorder brain.

OBJECTIVE: Major depressive disorder (MDD) is a leading global cause of disability, marked by persistent mood disturbances, cognitive deficits, and changes in prefrontal cortex neural circuitry. In this study, we aimed to define cell-type-specific molecular and regulatory mechanisms underlying MDD by mapping gene-expression and chromatin-accessibility changes in the dorsolateral prefrontal cortex (PFC) (dlPFC). METHODS: Postmortem dlPFC (BA9) tissue from 7 MDD and 8 well-matched controls was analyzed using 10× Genomics snRNA-seq and paired ATAC+RNA multiome sequencing. Sequencing data were processed with Cell Ranger pipelines, nuclei were filtered for quality and doublets/debris, and datasets were integrated and clustered using Seurat/Signac packages. Differential gene expression, chromatin accessibility, and transcription factor motif activity were tested between MDD and controls within each cell type, followed by peak-to-gene linkage and Gene Ontology (GO)/Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and PsyGeNET enrichment to interpret dysregulated regulatory mechanisms. RESULTS: A total of 20 distinct clusters encompassing major neuronal and non-neuronal populations were identified. Differential analyses uncovered extensive cell type-specific changes in chromatin accessibility and gene expression, particularly within excitatory layer 5/6 and inhibitory Pvalb neurons, as well as glial and vascular populations. Functional enrichment indicated dysregulation of synaptic organization, neurotransmission, myelination, stress-response, and immune-regulatory pathways across neuronal and non-neuronal cells. Notably, glucocorticoid-responsive transcription factors NR3C1/NR3C2 exhibited conserved regulatory networks implicating stress signaling in MDD pathophysiology. CONCLUSIONS: Together, these findings provide a comprehensive single-nucleus atlas of gene regulation in the MDD PFC, highlighting coordinated dysfunction across neurons, glia, and vascular cells.

Major Depressive Disorder↗