Hsu Paul Wei-Che, Yeh Chi-Hsiao, Lo Chi-Jen, Tsai Tsung-Hsien, Chan Yun-Hsuan, Chou Yi-Ju, Yang Ning-I, Cheng Mei-Ling, Sheu Wayne Huey-Herng, Lai Chi-Chun, Sytwu Huey-Kang, Tsai Ting-Fen
Institute of Molecular and Genomic Medicine, National Health Research Institutes, Miaoli, Taiwan.
Department of Thoracic and Cardiovascular Surgery, Chang Gung Memorial Hospital, Linkou, Taoyuan, Taiwan.
Biomark Res. 2025 Aug 20;13(1):107. doi: 10.1186/s40364-025-00821-y.
Coronary artery disease (CAD) remains a leading cause of mortality in developed nations. While previous genome-wide association studies have identified single-nucleotide polymorphisms (SNPs) linked to CAD, their impact on disease progression requires trans-omics validation.
This study merges whole genome SNP analysis and metabolomic profiling to distinguish CAD patients from high-risk and healthy individuals. A cross-sectional study was conducted, enrolling participants from the Northeastern Taiwan Community Medicine Research Cohort, which spans the period between August 2013 and November 2020. A total of 781 participants were included in the study and categorized into three groups: control (n = 271), high-risk (n = 363), and CAD (n = 147) groups, following a stratification protocol. The study integrated K-clustering of metabolomics and SNP datasets. Subsequently, a machine-learning (ML)-assisted prediction model was developed specifically for CAD identification.
Four significant findings emerged. Firstly, plasma levels of phospholipids decline from healthy controls to high-risk individuals and then decline further among CAD patients. This indicates that plasma phospholipids have potential as biomarkers and implies that they have a role in CAD progression. Secondly, five genes are linked to lipidomic alterations via their top-ranking among CAD-associated SNPs. Thirdly, a specific LPCAT1 haplotype is associated with CAD using a trans-omics approach. Lastly, an ML-assisted trans-omics prediction model for CAD was developed, which achieves an area under the curve of 0.917, with LPCAT1 among the 16 top-ranked predictive features.
This study highlights the usefulness of a multi-omics signature when discriminating CAD patients and suggests that abnormalities in phospholipid metabolism are influenced by LPCAT1 genetic variants. Our findings underscore the potential of multi-omics approaches to our understanding and identification of critical factors in CAD development.
ClinicalTrials.gov Identifier: NCT04839796; Aug 2013.
冠状动脉疾病（CAD）仍是发达国家的主要死因。虽然先前的全基因组关联研究已经确定了与CAD相关的单核苷酸多态性（SNP），但其对疾病进展的影响需要跨组学验证。
本研究将全基因组SNP分析与代谢组学分析相结合，以区分CAD患者与高危个体和健康个体。进行了一项横断面研究，招募了来自台湾东北部社区医学研究队列的参与者，该队列涵盖2013年8月至2020年11月期间。按照分层方案，共有781名参与者纳入研究并分为三组：对照组（n = 271）、高危组（n = 363）和CAD组（n = 147）。该研究整合了代谢组学和SNP数据集的K聚类。随后，专门开发了一种机器学习（ML）辅助预测模型用于CAD识别。
出现了四项重要发现。首先，血浆磷脂水平从健康对照组到高危个体逐渐下降，在CAD患者中进一步下降。这表明血浆磷脂有作为生物标志物的潜力，并暗示它们在CAD进展中起作用。其次，五个基因在与CAD相关的SNP中排名靠前，与脂质组改变有关。第三，使用跨组学方法，一种特定的LPCAT1单倍型与CAD相关。最后，开发了一种用于CAD的ML辅助跨组学预测模型，其曲线下面积为0.917，LPCAT1是16个排名靠前的预测特征之一。
本研究强调了多组学特征在区分CAD患者时的有用性，并表明磷脂代谢异常受LPCAT1基因变异影响。我们的研究结果强调了多组学方法在我们理解和识别CAD发展关键因素方面的潜力。
ClinicalTrials.gov标识符：NCT04839796；2013年8月。
