SEARCH · Search PubMed

Results for “transcriptomic”

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.

SourceResearch connectionFrom yearThrough year

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

At least 325 records · Page 18Linked to original sources

PubMed2026

Primary Tumor Epigenetic and Transcriptomic Alterations Associated with Nodal Burden and Metastatic Risk in ER+/HER2- Breast Cancer.

De-escalation of axillary surgery has resulted in the loss of pathologic nodal information, yet the extent of lymph node involvement remains an important determinant of treatment decisions in estrogen receptor-positive (ER+)/HER2- disease. We examined whether primary tumors differed molecularly according to the extent of this regional dissemination. Genome-wide DNA methylation profiling of primary ER+/HER2- tumors from 47 patients with pN1 (n = 29) vs. >pN1 (n = 18) disease showed differences concentrated at promoters of developmental and cell-adhesion genes. By integrating methylomes with transcriptomes from the TCGA-BRCA cohort (n = 148) and clinical outcomes from KM Plotter (RFS, n = 1154; OS, n = 442; DMFS, n = 423), we identified four genes (ARL10, RIC3, CXCL14, KCNH2) showing concordant molecular and clinical associations, from which we derived the Lymph-node Involvement Outcome Numerator (LION) score. Lower LION scores were observed in metastatic lesions from the AURORA US cohort (n = 45). In SCAN-B (n = 3969), lower scores were associated with shorter distant recurrence-free intervals (HR = 0.38; 95% CI 0.23-0.62); this association persisted after adjustment for age, nodal and tumor category but was lost after adjustment for histological grade (HR = 0.83; 95% CI 0.48-1.44), indicating that the score and grade capture overlapping biology. These findings suggest that primary tumors already display coordinated epigenetic and transcriptional alterations associated with the extent of metastatic dissemination.

Humans↗

PubMed2026

A Practical Workflow for Spatial Transcriptomics Data Analysis: From Data Acquisition to Advanced Analyses.

Spatial transcriptomics (ST) profiles genome-wide gene expression while preserving the two-dimensional spatial context of mRNA molecules within tissue sections, enabling studies of tissue architecture and microenvironment-associated biology. However, ST analysis remains challenging because data import, quality control, integration, deconvolution, spatial statistics, and visualization often require multiple software environments and reproducible parameter choices. This protocol presents a practical computational workflow for public ST datasets in R, beginning with data acquisition and software setup and proceeding through Seurat-based data loading, quality control, normalization, multi-sample integration, clustering, and spatially variable gene analysis. The workflow then applies complementary deconvolution strategies, including reference-guided SPOTlight analysis and unsupervised STdeconvolve topic modeling, followed by Giotto-based spatial cell-cell communication analysis and interactive region-of-interest (ROI) selection using a custom Python Dash application. By emphasizing script-based execution, explicit parameter rationales, expected outputs, and troubleshooting checkpoints, the protocol provides an adaptable framework for standard array-based ST datasets and related platforms after dataset- and platform-specific parameter evaluation.

Spatial Transcriptomics↗

PubMed2026

Multistage Genetic, Transcriptomic, and Single-Cell Evidence Prioritizes MAP1LC3A among Ferroptosis-Related Genes in Glioblastoma.

Glioblastoma (GBM) remains a highly aggressive malignancy, and the contribution of ferroptosis-related genes to disease susceptibility remains incompletely understood. A genetically anchored, multistage framework was applied to prioritize ferroptosis-related genes associated with GBM. Among 483 genes curated from FerrDb V2, 315 had candidate cis-expression quantitative trait loci (cis-eQTLs) in eQTLGen, 250 retained at least three independent instruments after linkage disequilibrium clumping, and 226 yielded valid inverse-variance weighted (IVW) Mendelian randomization estimates using a GBM genome-wide association study comprising 6,183 cases and 18,169 controls. Thirty-four genes met the exploratory discovery criteria of P < 0.05 and a Benjamini-Hochberg false discovery rate (BH-FDR) < 0.20, with directionally concordant Bayesian weighted Mendelian randomization (BWMR) estimates. Replication-stage Mendelian randomization using GTEx V10 whole-blood cis-eQTLs supported four genes: ATG7, RPTOR, MAP1LC3A, and CHMP6. Evaluation across three independent tumor-control transcriptomic cohorts demonstrated that MAP1LC3A was consistently downregulated in tumor tissue and showed a significant random-effects pooled estimate (log₂ fold change, -1.273; 95% confidence interval, -1.625 to -0.920; false discovery rate = 0.016), whereas the other three genes lacked comparable cross-cohort statistical support. Single-cell virtual knockout analysis was subsequently performed in a patient-balanced subset of 2,400 malignant cells selected from 4,916 eligible cells across 20 adult IDH-wild-type GBM tumors. Across five independently seeded runs, 3, 15, 4, and 7 robust downstream genes were identified for ATG7, RPTOR, MAP1LC3A, and CHMP6, respectively. The resulting consensus sets comprised 17 unique genes, with RND3 shared across all four targets. Gene Ontology analysis indicated enrichment of cell-adhesion and cell-surface processes, whereas no KEGG or Reactome pathways remained significant after multiple-testing correction. Collectively, these findings prioritize MAP1LC3A for future experimental investigation while distinguishing genetic association, tumor-expression concordance, and computational perturbation from definitive evidence of causality or mechanism.

Humans↗

PubMed2026

Distinct periarticular muscle transcriptomes: inflammation in rheumatoid arthritis versus metabolic dysregulation in osteoarthritis.

OBJECTIVES: Periarticular skeletal muscle abnormalities are recognised in rheumatoid arthritis (RA) and osteoarthritis (OA), but their divergent molecular pathologies are poorly defined. This study aimed to elucidate and directly compare the transcriptomic profiles of periarticular muscle in patients with RA and OA. METHODS: We performed bulk RNA sequencing of periarticular skeletal muscle samples collected during total joint arthroplasty from RA (n=6) and OA (n=4) patients. Differential gene expression analysis, weighted gene co-expression network analysis (WGCNA), pathway enrichment, and gene set variation analyses were conducted to identify disease-specific molecular features and their clinical associations. RESULTS: The two conditions showed fundamentally distinct profiles. RA muscle exhibited a pronounced inflammatory signature, characterised by upregulation of cytokine-responsive genes including FOS, EGR1, and CXCL2, and enrichment of tumour necrosis factor-α and interleukin-6 (IL-6)/JAK-STAT3 signalling. In contrast, OA muscle was characterised by metabolic dysregulation, with upregulation of genes linked to adipogenesis (PCK1, SFRP4) and significant enrichment of epithelial-to-mesenchymal transition (EMT) signalling. These divergent profiles were further supported by WGCNA, which identified distinct modules reflecting heightened innate immune and complement activation in RA, and disrupted metabolic processes in OA. Notably, in RA, the IL-2-STAT5 signalling pathway was unique among those tested in showing a strong positive correlation with DAS28-ESR (r=0.94, p=0.019). CONCLUSIONS: This study reveals distinct molecular pathologies in the periarticular muscle of RA and OA. RA muscle shows an intense inflammatory profile potentially linked to cachexia, whereas OA muscle displays features of metabolic disease and pro-fibrotic remodelling.

Humans↗

PubMed2026

Phenotypic and transcriptomic characterization of biallelic RNU2-2 developmental and epileptic encephalopathy.

OBJECTIVE: A significant proportion of individuals with suspected genetic developmental and epileptic encephalopathies (DEEs) remain unsolved following whole genome sequencing (WGS). Here we describe biallelic RNU2-2 variants causing a recently reported, severe, recessive DEE. METHODS: We screened individuals who have received WGS analyses at the Genomic Medicine Centre Karolinska for Rare Diseases for biallelic RNU2-2 variants. Deep phenotyping was performed through reviewing entire medical histories and phenotypic traits were transcribed to their corresponding Human Phenotype Ontology (HPO) term. HPO terms were used to generate pairwise phenotypic similarity scores and assess for significantly shared phenotype enrichment in the RNU2-2 sub-cohort. RNA sequencing analyses were performed in fibroblast and blood tissues to compare splicing events between RNU2-2 individuals and two independent control groups. RESULTS: We identified 14 individuals from nine families with 12 ultra-rare biallelic RNU2-2 variants clustering in the conserved 5' domains. Genotype data from 13 of 14 individuals has been reported previously as part of a larger cohort. All individuals presented with a highly concordant, severe DEE, characterized by severe to profound intellectual disability, inability to walk or communicate, hyperkinesia, and refractory seizures. Infantile spasms and tonic seizures were the predominant seizure types and a Lennox-Gastaut syndrome-like phenotype was common. These individuals had a significantly similar phenotypic signature when compared with 703 individuals with complex pediatric epilepsies (two-sided Monte Carlo permutation test, p = .005). RNA sequencing analyses showed aberrant splicing, with the most pronounced effects in fibroblast tissues in mutually exclusive exon and alternate 3' splice-site events, which were not detectable in blood. SIGNIFICANCE: We present deep phenotyping data and transcriptomic analyses that provide support for rare, 5' clustering biallelic RNU2-2 variants causing this novel, severe DEE. We propose an RNA sequencing methodology on fibroblast tissue for future validation of RNU2-2 variants.

autosomal recessive disease↗

PubMed2026

Synthesis of Padina boergesenii-Derived Zinc Oxide Nanoparticles and their Therapeutic Potential Against Oral Squamous Cell Carcinoma: A Transcriptomic and in Vitro Evaluation.

Cancer remains a major health challenge, with oral squamous cell carcinoma (OSCC) being an high aggressive subtype of head and neck squamous cell carcinoma that lacks effective therapeutic options. Current study integrates the synthesis of zinc oxide nanoparticles (ZnO-NPs) from the marine brown algae Padina boergesenii with the OSCC gene expression profile to evaluate their potential therapeutic effects against OSCC. Herein, the ZnO-NPs from Padina boergesenii were prepared through the green synthesis method. The obtained ZnO-NPs were characterized through spectroscopic methods, the UV spectrophotometer shows maximum absorbance at 372 nm, FT-IR presents Zn-O functional band at 516 cm- 1, HR-TEM confirms average particle size of 55.70 nm and the Zetasizer shows zeta potential of + 12.9 mV, indicating colloidal stability. The cytotoxicity assay with ZnO-NPs against oral cancer cell lines exhibited a reduction in cell viability at IC₅₀ value of 20 µg/mL. Meanwhile, the transcriptome analysis of OSCC highlights that MYC, STAT3, BRCA1, and AKT1 are the OSCC therapeutic targets involved in proliferation, immune evasion, genomic instability, and cancer signalling pathways. Further, qRT-PCR based gene expression analysis demonstrates significant down-regulation of these targets upon ZnO-NPs treatment in KB cell lines. Overall, this study emphasizes the anticancer potential of Padina boergesenii-derived ZnO-NPs that could effectively modulate the therapeutic targets and may benefit the treatment of OSCC cancer.

Cytotoxicity↗

PubMed2026

Uncovering hub genes and key pathways responsive to drought stress in rice via meta-analysis of transcriptomic data.

Drought stress presents a formidable threat to global rice cultivation, triggering complex molecular responses that impact plant growth and productivity. To decipher the underlying gene expression dynamics, we performed a comprehensive meta-analysis of transcriptomic datasets derived from drought-tolerant rice genotypes. Via microarray data from three independent studies, we identified a set of consistently expressed differentially expressed genes (DEGs) under drought conditions. Integration of functional annotation tools, including GO and KEGG pathway enrichment, revealed key biological processes and signaling cascades involved in stress mitigation, such as ABA signaling, protein folding, and photosynthesis suppression. Protein-protein interaction (PPI) network construction, followed by hub gene identification via maximal clique centrality (MCC), highlighted pivotal regulators including LEA proteins, dehydrins, HSP70, and several transcription factors. Machine learning approaches further prioritize potential biomarkers, with Random Forest models achieving high classification accuracy and pinpointing key predictive genes. Chromosomal localization analysis provided spatial insights into the distribution of these hub genes, whose expression patterns were further compared against qRT-PCR data from previously published studies. This integrative approach identifies candidate genomic markers and mechanistic insights that may support future breeding strategies for drought-tolerant rice, pending experimental validation.

Cytoscape↗

PubMed2026

Deciphering novel targets in salivary gland pleomorphic adenoma by integrating plasma proteomics and parotid transcriptomics analyses.

BACKGROUND/PURPOSE: Pleomorphic adenoma (PA) is the most common salivary gland benign tumor, with its molecular drivers elusive due to a lack of experimental models. This study aimed to decipher novel targets in PA by systematically integrating plasma protein quantitative trait loci (pQTL)-based Mendelian randomization (MR) with multi-omics profiling of parotid gland tissues. MATERIALS AND METHODS: We performed two-sample MR using 5450 plasma pQTLs and genome-wide association study summary for benign or broader salivary gland diseases from FinnGen consortium. Bulk RNA-sequencing (RNA-seq) and single-cell RNA-seq (scRNA-seq) comparing PA to normal tissue were used for transcriptomic validation. Immunohistochemistry (IHC) was applied for protein-level validation in human PA, adenoid cystic carcinoma (ACC), and murine inflammatory lesions. RESULTS: MR identified 12 plasma proteins associated with benign salivary gland tumor risk. Transmembrane serine protease 6 (TMPRSS6) was the only protein significantly risk-increasing for both benign and broader salivary gland diseases. Strikingly, mitogen-activated protein kinase kinase 4 (MAP2K4) showed opposite MR effects between benign and all-lesion outcomes. Bulk RNA-seq showed limited concordance with MR findings, while scRNA-seq revealed a unique plastic epithelium and partially validated candidates at cellular resolution. Critically, IHC confirmed MAP2K4 protein overexpression specifically in human PA, but not in ACC or inflammatory lesions, while TMPRSS6 was downregulated in established pathologies despite its genetic risk association. CONCLUSION: By integrating plasma proteome-based causal inference with parotid tissue multi-omics, this study unveils MAP2K4 as a potential PA-specific driver. This integrative framework provides novel, context-specific targets for further functional investigation in salivary gland tumorigenesis.

Gene expression profiling↗

PubMed2026

Genomic and transcriptomic quality control for an autologous iPSC-derived cell therapy for Parkinson's disease.

Toward development of an autologous, induced pluripotent stem cell (iPSC)-based cell therapy for Parkinson's disease (PD), we demonstrate successful, reproducible genomic and transcriptomic qualification of patient-derived dopaminergic neuron precursor cells (DANPCs) across multiple donors. Our analysis includes whole-genome sequencing data from fibroblasts, iPSCs, and DANPCs and the development of NeuriTest, an RNAseq-based bioinformatic analysis of DANPCs designed to predict cell quality based on empirical animal data. Autologous cell therapies are immune matched to the patient, potentially augmenting durability of benefit compared to allogeneic cells while negating the need for immunosuppression and accompanying side effects. Patient-specific iPSCs are an autologous cell source that can be differentiated to dopaminergic neurons, the cell type lost in PD. We report here our preclinical manufacturing strategy and results demonstrating efficacy in a PD rodent model and safety in a 9-month GLP toxicology study.

Parkinson’s disease↗

PubMed2026

Dynamic evolution of chaperone-mediated autophagy is associated with tumor microenvironment remodeling and prognostic stratification in lung adenocarcinoma: insights from single-cell transcriptomics, ensemble machine learning, and experimental validation.

BACKGROUND: Lung adenocarcinoma (LUAD) shows prognostic heterogeneity, and tumor-node-metastasis (TNM) staging is limited for individualized management. Chaperone-mediated autophagy (CMA) maintains proteostasis, but its role during adenocarcinoma in situ (AIS)-minimally invasive adenocarcinoma (MIA)-invasive adenocarcinoma (IAC) progression remains unclear. METHODS: Single-cell RNA sequencing (scRNA-seq) data from GSE189357 and bulk transcriptomes from The Cancer Genome Atlas (TCGA)-LUAD and Gene Expression Omnibus (GEO) cohorts were integrated. CMA activity, cell-cell communication, weighted gene co-expression network analysis (WGCNA), tumor-normal differential expression, machine-learning survival modeling, tumor microenvironment (TME) features, drug sensitivity, and EPC1 function were analyzed. RESULTS: CMA-high tumor epithelial cells increased from AIS (58.1%) to MIA (65.7%) but declined in IAC (44.4%; p < 0.001). CMA-low cells preferentially received fibroblast-derived extracellular matrix cues. A CMA-negatively correlated module identified 69 core genes. Random survival forest (RSF) performed best among 117 machine-learning combinations (mean concordance index > 0.873). High-risk patients had worse survival across cohorts, and the risk score was independently associated with overall survival (hazard ratio = 16.013, 95% confidence interval: 9.579-26.768, p < 0.001). High-risk tumors showed proliferative activation and M0 macrophage enrichment, whereas low-risk tumors showed stronger immune-related signaling. EPC1 overexpression suppressed malignant phenotypes in A549 cells. CONCLUSION: CMA dynamics are associated with stromal and immune remodeling during LUAD progression. A CMA-based model provides robust prognostic stratification and may offer a basis for future TME-guided studies.

Chaperone-mediated autophagy↗

PubMed2026

Genetic dissection of cardiac iron regulation using transcriptome network analysis and systems genetics in BXD mice.

Cardiac iron homeostasis is essential for myocardial energy metabolism and contractile function, yet the genetic and molecular mechanisms governing iron levels within the heart remain poorly understood. We used a systems genetics approach to dissect the transcriptional regulation of cardiac iron homeostasis. Myocardial iron level varies substantially across BXD strains (40-112 μg/g) and is under heritable genetic control (H2 = 0.38). Elevated cardiac iron is associated with reduced ventricular mass, increased ventricular ectopy, and prolonged atrioventricular conduction in the BXD population. Weighted gene co-expression network analysis of the BXD heart transcriptome identified a co-expression module that was significantly and negatively correlated with cardiac iron levels in both young and old BXD mice and enriched for pathways related to metabolic regulation, cyclic AMP (cAMP) signaling, circadian entrainment, and cardiovascular physiology. The module showed substantial overlap with a curated cardiac iron gene set, and cross-species enrichment analysis confirmed its conservation in human cardiomyopathy differentially expressed genes (enrichment ratio = 1.49; false discovery rate [FDR] = 0.0342). Quantitative trait locus (QTL) mapping of the first principal component of the overlapping module iron genes (n = 38), corroborated by individual gene mapping, identified trans-eQTL hotspots on multiple chromosomes, implicating Fcho2, Gcc2, and Rmdn1 as candidate upstream regulators operating through sequential steps of intracellular iron trafficking. Together, these findings establish a systems-level map of cardiac iron gene regulation, identify candidate genetic regulators, and provide a molecular framework linking disruption of iron-related transcriptional networks to structural and electrical cardiac dysfunction with implications for iron-related heart diseases.

BXD mouse population↗

PubMed2026

A single-nucleus transcriptomic atlas of human inner ear development.

Hearing and balance rely on coordinated activity of multiple inner ear cell types, yet the mechanisms governing their development and specification in humans remain unclear. Consequently, this limits our understanding of how disease genes affect cell type formation and function, limiting the development of targeted treatments, including gene therapies. Here we present the Human Inner Ear Development snRNA-seq Atlas (HIEDRA), a single-nucleus transcriptomic atlas of the human inner ear spanning the first and second trimesters. HIEDRA maps sensory and nonsensory epithelia, neurons and mesenchyme-associated populations, including undercharacterized secretory cells required for ion homeostasis. We identify selective vulnerability in sensory and secretory lineages to disease-associated genes, infer regulatory networks and show that Hedgehog signaling suppression is required for secretory cell specification. We validate this mechanism in human inner ear organoids, expanding the model to include all major cell types. Altogether, these findings provide insights into human inner ear cell type specification, improve in vitro models and establish HIEDRA as a resource for investigating human inner ear development.

Journal Article↗

PubMed2025

Genome topology analysis and transcriptomics of human osteoclasts reveals enhancer-promoter interactions at loci for bone traits and diseases.

Genome-wide association studies (GWAS) relevant to osteoporosis have identified hundreds of loci; however, understanding how these variants influence the phenotype is complicated because most reside in non-coding DNA sequence that serves as transcriptional enhancers and repressors. To advance knowledge on these regulatory elements in osteoclasts (OCs), we performed Micro-C analysis, which informs on the genome topology of these cells and integrated the results with transcriptome and GWAS data to further define loci linked to BMD. Using blood cells isolated from 4 healthy participants aged 31-61 yr, we cultured OC in vitro and generated a Micro-C chromatin conformation capture dataset. We characterized chromatin loops (CLs) in OC from among more than 69 million chromatin interactions identified in the genome. Of the CL identified in OC, >16 000 were unique compared to precursor cells. When sentinel single nucleotide polymorphisms from osteoporosis and bone-related GWAS and those in linkage disequilibrium at r 2 > 0.6 were mapped to CL for OC, 12 588 of these variants were observed within chromatin contact regions. Notable in differential gene ontology enrichment analyses of the topology data for OC and precursors were pathways regulating pluripotency of stem cells, Wnt signaling, nucleotide-binding oligomerization domain (NOD)-like receptor signaling and chemokine signaling. These data, in combination with other 3D genome architecture and epigenetic data (eg, histone modifications and chromatin accessibility), will be useful in modeling to predict genome-wide, which enhancers regulate which genes in OC. This data will therefore also be informative for resolving GWAS hits. In conclusion, we have generated a high-resolution genome topology dataset for human OC and have used this to identify CLs relevant to studies of the genetics of osteoporosis. This data will serve as a powerful resource to inform future functional studies of OC biology.

BMD↗

PubMed2026

Genomic and Transcriptomic Correlates of Deep PSA Response in Patients with Metastatic Androgen Pathway Modulation-Sensitive Prostate Cancer.

BACKGROUND: Despite advances in metastatic androgen pathway modulation-sensitive prostate cancer (mAPMS) treatment, outcomes remain heterogeneous. Achieving a post-treatment undetectable prostate specific antigen (PSA) is a strong prognostic marker. We aimed to identify genomic and transcriptomic determinants of PSA response in a real-world clinical-genomic cohort. PATIENTS AND METHODS: Patients with mAPMS who underwent DNA (Tempus xT) and, in a subset, RNA (Tempus xR) sequencing were identified from the Tempus Lens database. Inclusion required stage IV disease within 90 days of sample collection and samples obtained within 12 months before or 3 months after treatment initiation. Patients with PSA at 6 months (n = 525) were classified as PSA-low (<0.1 ng/mL, n = 240) or PSA-high (≥0.1 ng/mL, n = 285). Overall survival (OS) was assessed by 6-month landmark analysis with delayed-entry adjustment. Logistic and Cox models were adjusted for clinical variables. Sensitivity analyses used a relative definition of > 95% PSA decline from baseline. RESULTS: Baseline PSA was lower in PSA-low versus PSA-high patients (24 vs 36 ng/mL, p = 0.01). SPOP (17% vs 11%) and ZFHX3 (2.5% vs 6%) alterations differed between groups, but neither persisted after adjustment. Using the relative definition, ZMYM3 and JAK1 alterations were independently associated with failure to achieve a deep PSA response. Expression of PSMA, TROP2, B7-H3, and STEAP1 did not differ between groups. PSA-low status was independently associated with improved OS, as was deep relative response. CONCLUSION: Deep PSA response at 6 months correlates with improved OS in mAPMS. Integrating molecular markers with PSA response may inform treatment intensification or de-escalation strategies.

Biomarkers↗

PubMed2026

Tertiary lymphoid structure transcriptomic signatures show limited and cohort-dependent value for predicting axillary nodal involvement in oestrogen receptor-positive luminal breast cancer.

Tertiary lymphoid structures (TLS) are associated with prognosis in solid tumours. Their value for predicting axillary nodal involvement in oestrogen receptor-positive luminal breast cancer remains uncertain. Three published TLS signatures were scored by single-sample gene set enrichment analysis in oestrogen receptor-positive luminal tumours. The Cancer Genome Atlas Breast Invasive Carcinoma cohort (TCGA-BRCA) included 632 cases, of which 379 met strict consensus. METABRIC included 1086 cases, of which 663 met strict consensus. Logistic models adjusted for age and pathological tumour stage. Strict consensus, majority vote, and continuous scores were compared. Performance assessment included bootstrapped changes in area under the receiver-operating-characteristic curve, Brier scores, calibration, and decision-curve analysis. Survival was evaluated in METABRIC and explored in TCGA-BRCA. Strict-consensus TLS status was not associated with nodal positivity in TCGA-BRCA (adjusted odds ratio: 0.95, 95% confidence interval: 0.62-1.45, P = 0.822). METABRIC was similar (odds ratio: 0.76, 95% confidence interval: 0.55-1.06, P = 0.105). Full-cohort METABRIC analyses detected small majority-vote and continuous-score associations, absent in TCGA-BRCA. Across specifications, bootstrapped changes in area under the receiver-operating-characteristic curve ranged from 0.0002 to 0.0089, with minimal Brier-score improvement and no stable decision-curve benefit. In METABRIC, the univariable overall survival association attenuated after age adjustment (hazard ratio: 1.33-1.10). TCGA-BRCA survival analyses were nonsignificant. TLS transcriptomic signals showed small, cohort-dependent associations with nodal status but no reproducible or clinically meaningful incremental predictive value. These data do not support replacing sentinel lymph node biopsy with a TLS signature in oestrogen receptor-positive luminal breast cancer.

breast cancer↗

PubMed2023

Single-cell transcriptomic atlas of Alzheimer's disease middle temporal gyrus reveals region, cell type and sex specificity of gene expression with novel genetic risk for MERTK in female.

Alzheimer's disease, the most common age-related neurodegenerative disease, is closely associated with both amyloid-ß plaque and neuroinflammation. Two thirds of Alzheimer's disease patients are females and they have a higher disease risk. Moreover, women with Alzheimer's disease have more extensive brain histological changes than men along with more severe cognitive symptoms and neurodegeneration. To identify how sex difference induces structural brain changes, we performed unbiased massively parallel single nucleus RNA sequencing on Alzheimer's disease and control brains focusing on the middle temporal gyrus, a brain region strongly affected by the disease but not previously studied with these methods. We identified a subpopulation of selectively vulnerable layer 2/3 excitatory neurons that that were RORB-negative and CDH9-expressing. This vulnerability differs from that reported for other brain regions, but there was no detectable difference between male and female patterns in middle temporal gyrus samples. Disease-associated, but sex-independent, reactive astrocyte signatures were also present. In clear contrast, the microglia signatures of diseased brains differed between males and females. Combining single cell transcriptomic data with results from genome-wide association studies (GWAS), we identified MERTK genetic variation as a risk factor for Alzheimer's disease selectively in females. Taken together, our single cell dataset revealed a unique cellular-level view of sex-specific transcriptional changes in Alzheimer's disease, illuminating GWAS identification of sex-specific Alzheimer's risk genes. These data serve as a rich resource for interrogation of the molecular and cellular basis of Alzheimer's disease.

Journal Article↗

PubMed2023

Inferring Metabolic States from Single Cell Transcriptomic Data via Geometric Deep Learning.

The ability to measure gene expression at single-cell resolution has elevated our understanding of how biological features emerge from complex and interdependent networks at molecular, cellular, and tissue scales. As technologies have evolved that complement scRNAseq measurements with things like single-cell proteomic, epigenomic, and genomic information, it becomes increasingly apparent how much biology exists as a product of multimodal regulation. Biological processes such as transcription, translation, and post-translational or epigenetic modification impose both energetic and specific molecular demands on a cell and are therefore implicitly constrained by the metabolic state of the cell. While metabolomics is crucial for defining a holistic model of any biological process, the chemical heterogeneity of the metabolome makes it particularly difficult to measure, and technologies capable of doing this at single-cell resolution are far behind other multiomics modalities. To address these challenges, we present GEFMAP (Gene Expression-based Flux Mapping and Metabolic Pathway Prediction), a method based on geometric deep learning for predicting flux through reactions in a global metabolic network using transcriptomics data, which we ultimately apply to scRNAseq. GEFMAP leverages the natural graph structure of metabolic networks to learn both a biological objective for each cell and estimate a mass-balanced relative flux rate for each reaction in each cell using novel deep learning models.

Preprint↗

PubMed2024

Transcriptome-Wide Root Causal Inference.

Root causal genes correspond to the first gene expression levels perturbed during pathogenesis by genetic or non-genetic factors. Targeting root causal genes has the potential to alleviate disease entirely by eliminating pathology near its onset. No existing algorithm discovers root causal genes from observational data alone. We therefore propose the Transcriptome-Wide Root Causal Inference (TWRCI) algorithm that identifies root causal genes and their causal graph using a combination of genetic variant and unperturbed bulk RNA sequencing data. TWRCI uses a novel competitive regression procedure to annotate cis and trans-genetic variants to the gene expression levels they directly cause. The algorithm simultaneously recovers a causal ordering of the expression levels to pinpoint the underlying causal graph and estimate root causal effects. TWRCI outperforms alternative approaches across a diverse group of metrics by directly targeting root causal genes while accounting for distal relations, linkage disequilibrium, patient heterogeneity and widespread pleiotropy. We demonstrate the algorithm by uncovering the root causal mechanisms of two complex diseases, which we confirm by replication using independent genome-wide summary statistics.

Journal Article↗