Subirana-Granés Marc, Zhang Haoyu, Gupta Prashant, Pividori Milton
Department of Biomedical Informatics, University of Colorado Anschutz Medical Campus, Aurora, CO, USA.
Department of Biomedical Informatics, University of Colorado School of Medicine, Aurora, CO 80045, USA; Colorado Center for Personalized Medicine, University of Colorado Anschutz Medical Campus, Aurora, CO, USA.
bioRxiv. 2025 Jun 12:2025.06.05.658129. doi: 10.1101/2025.06.05.658129.
Down syndrome (DS) is caused by trisomy of chromosome 21 and is associated with diverse clinical manifestations, yet the molecular pathways linking chromosome-21 dosage effects to DS comorbidities remain poorly defined. Here we address this gap by applying a network-based, integrative framework that combines whole-blood transcriptomic data with gene-trait associations to uncover mechanistic insights into DS-associated conditions. First, we performed matrix factorization using PLIER on Human Trisome Project (HTP) RNA-Seq profiles from 304 trisomy-21 (T21) and 95 euploid (D21) individuals, deriving 156 biologically interpretable gene modules. We then identified 92 modules whose activity differed significantly between T21 and D21 and annotated these with prior-knowledge and KEGG pathways. To connect modules to clinical traits, we integrated PrediXcan-derived TWAS results from the UK Biobank, revealing 25 T21-specific modules with significant gene-trait associations (FDR < 0.1), including modules linked to cardiovascular, hematological, immune, metabolic, and neurological phenotypes relevant to DS. Using HTP clinical records as a replication cohort, 13 of these modules reliably predicted comorbidity status (AUC > 0.65, mAPS > 0.65). Most notably module 37, an interferon-stimulated gene network, whose elevated expression robustly distinguished DS individuals with pulmonary hypertension (AUC = 0.76, mAPS = 0.73). Overall, our study demonstrates that integrating blood-derived gene modules with population-scale genetic data uncovers coherent molecular signatures underlying DS comorbidities, identifies candidate biomarkers and therapeutic targets (e.g., , , ), and highlights the power of combining transcriptomic and genetic evidence to elucidate complex disease mechanisms.
唐氏综合征（DS）由21号染色体三体性引起，与多种临床表现相关，但将21号染色体剂量效应与DS合并症联系起来的分子途径仍不清楚。在这里，我们通过应用基于网络的综合框架来填补这一空白，该框架将全血转录组数据与基因-性状关联相结合，以揭示DS相关病症的机制性见解。首先，我们使用PLIER对来自304名21号染色体三体（T21）和95名整倍体（D21）个体的人类三体项目（HTP）RNA测序图谱进行矩阵分解，得出156个具有生物学可解释性的基因模块。然后，我们确定了92个在T21和D21之间活性有显著差异的模块，并用先验知识和KEGG途径对这些模块进行注释。为了将模块与临床性状联系起来，我们整合了来自英国生物银行的PrediXcan衍生的全转录组关联研究（TWAS）结果，揭示了25个具有显著基因-性状关联（FDR<0.1）的T21特异性模块，包括与DS相关的心血管、血液、免疫、代谢和神经表型相关的模块。使用HTP临床记录作为复制队列，其中13个模块可靠地预测了合并症状态（AUC>0.65，mAPS>0.65）。最值得注意的是模块37，一个干扰素刺激基因网络，其表达升高能有力地区分患有肺动脉高压的DS个体（AUC=0.76，mAPS=0.73）。总体而言，我们的研究表明，将血液来源的基因模块与群体规模的遗传数据相结合，揭示了DS合并症背后连贯的分子特征，识别了候选生物标志物和治疗靶点（例如， ， ， ），并突出了结合转录组学和遗传证据来阐明复杂疾病机制的力量。
