Introduction: Alzheimer’s disease (AD) and glioblastoma (GBM) are severe neurological disorders that pose significant global healthcare challenges. Despite extensive research, the molecular mechanisms, particularly those involving mitochondrial dysfunction, remain poorly understood. A major limitation in current studies is the lack of cell-specific markers that effectively represent mitochondrial dynamics in AD and GBM.Methods: In this study, we analyzed single-cell transcriptomic data using 10 machine learning algorithms to identify mitochondria-associated cell-specific markers. We validated these markers through the integration of gene expression and methylation data across diverse cell types. Our dataset comprised single-nucleus RNA sequencing (snRNA-seq) from AD patients, single-cell RNA sequencing (scRNA-seq) from GBM patients, and additional DNA methylation and transcriptomic data from the ROSMAP, ADNI, TCGA, and CGGA cohorts.Results: Our analysis identified four significant cross-disease mitochondrial markers: EFHD1, SASH1, FAM110B, and SLC25A18. These markers showed both shared and unique expression profiles in AD and GBM, suggesting a common mitochondrial mechanism contributing to both diseases. Additionally, oligodendrocytes and their interactions with astrocytes were implicated in disease progression, particularly through the APP signaling pathway. Key hub genes, such as HS6ST3 and TUBB2B, were identified across different cellular subpopulations, highlighting a cell-specific co-expression network linked to mitochondrial function.
27 October 2025 in CNS Neuroscience & Therapeutics
Aims: Alzheimer's disease (AD) is a devastating neurodegenerative disorder where mitochondrial dysfunction is increasingly recognized as pivotal, yet its comprehensive molecular underpinnings remain incompletely understood. This study aimed to systematically identify and validate mitochondria‐related biomarkers associated with AD risk and brain resilience, thus elucidating the molecular mechanisms underpinning mitochondrial dysfunction in AD. Methods: We innovatively integrated a multi‐omics approach, encompassing genomics, DNA methylation, RNA‐sequencing, and miRNA profiles from the ROSMAP and ADNI cohorts (sample sizes ranging from 638 to 2090 per omic layer). Additionally, we applied 10 distinct machine learning methods to robustly identify and validate critical mitochondrial biomarkers relevant to AD progression. Subsequent validation was performed using a two‐tiered approach: an in vivo AD mouse model to establish phenotypic relevance and an in vitro H 2 O 2 ‐induced oxidative stress model in HT22 cells to provide direct mechanistic validation. Results: Our computational analyses identified key biomarkers such as hsa‐miR‐129‐5p and SLC6A12 as pivotal regulators and highlighted the importance of the tricarboxylic acid (TCA) cycle. Experimentally, our AD mouse model exhibited significant cognitive deficits and brain remodeling, linked to a specific transcriptomic signature. Our in vitro model functionally recapitulated mitochondrial dysfunction and oxidative stress. Crucially, a cross‐model analysis revealed a core signature of seven genes (including APOE , CDKN1A , and CLOCK ) consistently dysregulated in both the cognitively impaired mouse brain and in neuronal cells subjected to direct oxidative insult. This provides powerful functional evidence linking our computationally derived targets, such as mitochondrial‐epistatic genes ( CLOCK ), to AD‐relevant pathology. Conclusion: These functionally validated findings provide deeper insights into the complex mitochondrial regulatory mechanisms involved in AD pathogenesis, offering robust biomarkers and novel potential avenues for developing targeted therapeutic strategies to address this challenging neurodegenerative disease.
Topologically associating domains (TADs) are fundamental units of three-dimensional genome organization, and their boundaries play important roles in gene regulation and genomic stability. However, accurate computational prediction of TAD boundaries remains challenging because of the complex interplay between DNA sequence and epigenomic regulatory signals. Here we present CT-TADB, a hybrid CNN-Transformer framework that integrates DNA sequence with histone modification and CTCF binding signals to predict TAD boundaries without requiring Hi-C data. Trained on six human cell lines, CT-TADB achieved AUC values of 0.932-0.950, demonstrating competitive or superior performance relative to existing multi-modal methods. The model maintained stable performance on strictly independent and dataset-specific boundary subsets, and exhibited robust cross-cell-type transferability and cross-species generalization between human and mouse (AUC > 0.80 for human-to-mouse transfer). Feature attribution analysis identified CTCF as the dominant boundary-associated factor, supported by active chromatin marks, and quantitative attention analysis revealed significant long-range CTCF-mediated dependencies (p < 0.001). CT-TADB further identified a clinically validated PITX2-associated TAD boundary linked to cardiac arrhythmia and captured condition-specific boundary dynamics. By leveraging routinely available epigenomic data, CT-TADB provides a practical and complementary framework for investigating chromatin architecture across diverse biological contexts.
Chromatin interactions regulate gene expression and genome organization, but computational prediction across cell types remains challenging. We developed UniChrom, a deep learning framework integrating DNA sequences and epigenomic features through attention-based neural networks to predict chromatin interactions. Evaluation across human lymphoblastoid, leukemia, and fibroblast cell lines demonstrates superior performance compared to existing methods, with fivefold cross-validation and Wilcoxon tests confirming statistical significance ( p < 0.05). Distance-stratified analysis reveals robust performance across all genomic scales, including long-range interactions exceeding 1.77 megabases (AUC: 0.976). Independent validation on endothelial cells confirms cross-lineage generalization (AUC: 0.962). Bootstrapping analysis with 1,000 iterations validates performance stability with tight 95% confidence intervals. DeepSHAP interpretability identifies CTCF and cohesin components as dominant features alongside cell-type-specific histone modifications, while DeepLIFT reveals functional regulatory motifs at interaction anchors. UniChrom provides a statistically validated framework for investigating genome architecture across cellular contexts with potential applications in understanding gene regulation in development and disease.