Computational Molecular Biology 2025, Vol.15, No.3, 122-130 http://bioscipublisher.com/index.php/cmb 127 5.2 Implementation of integration pipeline and visualization tools We ran a custom pipeline on the DMD case: first, we normalized gene expression and identified differentially expressed genes (DEGs) between DMD and control muscle. Variants in the DMD gene and other muscle-related genes were annotated. Using our multi-omics integration framework, we constructed a network linking mutated genes to downstream expression changes via known muscle pathway interactions (Lu et al., 2019). We also applied a Bayesian factor model to jointly analyze the genomic and transcriptomic data, revealing latent factors correlated with disease severity. Results are visualized in the web portal: for instance, a genome browser highlights the large deletions in DMD for each patient, while an interactive heatmap shows upregulated muscle regeneration genes in DMD samples (Schneegans et al., 2023). We also implemented Circos plots to display connections between genetic loci and transcriptional changes. These tools allow a user to explore how a variant (e.g., an exon deletion in DMD) propagates to altered gene networks and pathways. 5.3 Insights gained: molecular biomarkers and potential therapeutic targets The integrated analysis yielded new insights into DMD pathology. Consistent with prior studies, we observed upregulation of genes involved in muscle regeneration (e.g., MYOG, PAX7) and fibrosis (e.g., collagen genes) in DMD patients. Our model identified fibro-adipogenic progenitor (FAP) cells as key regulators: their aberrant signaling (through PDGF and TGF-β pathways) likely drives excess extracellular matrix deposition (Figure 2). Importantly, by cross-referencing DEGs with drug databases, we propose candidate molecular targets. For example, the gene SPP1 (osteopontin) was highly overexpressed and is known to modulate muscle inflammation; it emerges as a potential drug target or biomarker (Vera et al., 2022). Also, our protein network analysis suggests that upregulated myogenesis genes (e.g., MYH8, ACTA1) could serve as blood biomarkers for disease progression. These findings align with and extend published DMD research. This case study demonstrates the platform’s ability to generate clinically relevant hypotheses from integrated omics data. 6 Conclusion This study constructed an integrated multi-omics database for rare diseases and conducted case analyses of Duchenne muscular dystrophy and others based on this platform. We have integrated multi-level data such as genomics, transcriptomics, proteomics, metabolomics and phenotypes in accordance with a unified architecture, and developed standardized data models and interoperability frameworks. Through strict quality control and standard annotation, the database ensures the compatibility of data from different sources. Users can access it through a friendly Web interface and API for interactive analysis and visualization. This provides rare disease researchers with a pioneering tool, which is different from the previous single data resource and realizes the organic integration of knowledge and data. We have integrated multiple algorithms such as network analysis, Bayesian models, and machine learning into the platform to mine the biological significance of multi-omics data. These methods complement each other and reveal the key pathways, molecular modules and modifying factors of rare diseases from different perspectives. We utilize methods such as path enrichment and knowledge graphs to transform complex results into interpretable knowledge. This analytical system can be extended and applied to a variety of rare diseases, accelerating the process from data to discovery. This study demonstrates that by integrating multi-omics big data, biomarkers and drug targets of rare diseases can be systematically identified, thereby guiding patient stratification and individualized intervention. This has overturned the traditional model that relies on experience and fragmented research, pioneering a data-driven and AI-assisted research paradigm for rare diseases. Our platform has demonstrated potential in areas such as auxiliary diagnosis, drug reuse, and prognosis prediction, and will drive the development of precision medicine for rare diseases. Although there are still challenges ahead, we firmly believe that as long as we adhere to the concepts of open cooperation, technological innovation and patient-centeredness, the platform will surely continue to grow and thrive. From a current academic research tool, it has grown into one of the key cornerstones supporting precision medicine for rare diseases and the development of new drugs. We look forward to seeing the actual changes it has
RkJQdWJsaXNoZXIy MjQ4ODYzNA==