Preterm birth (PTB) is a leading global cause of infant mortality. Risk factors include genetics, lifestyle choices, and infection. Understanding the mechanism of PTB could aid the development of novel approaches to prevent PTB. This study aimed to investigate the metabolic biomarkers of PTB in early pregnancy and the association of significant metabolites with participant genotypes. Maternal serum collected at 16 and 20 weeks of gestation, from women who previously experienced PTB (high-risk) and women who did not (low-risk controls), were analysed using 1H Nuclear Magnetic Resonance (NMR) metabolomics and genome-wide screening microarray. ANOVA and probabilistic neural network (PNN) modelling were performed on the spectral bins. Metabolomics genome-wide association (mGWAS) of the spectral bins and genotype data from the same participants was applied to determine potential metabolite-gene pathways. Phenylalanine, acetate and lactate metabolite differences between PTB cases and controls were obtained by ANOVA and PNN showed strong prediction at week 20 (AUC=0.89). MGWAS identified several metabolite bins with strong genetic associations. Cis-eQTL analysis highlighted TRAF1 (involved in the inflammatory pathway) local to a non-coding SNP associated with lactate at week 20 of gestation. MGWAS of a well-defined cohort of participants highlighted a lactate-TRAF1 relationship that could potentially contribute to PTB.

中文翻译：

---

使用 NMR 和 mGWAS 方法对自发性早产高危妇女的母体血清进行代谢分析。


早产（PTB）是全球婴儿死亡的主要原因。风险因素包括遗传、生活方式选择和感染。了解 PTB 的机制有助于开发预防 PTB 的新方法。本研究旨在调查妊娠早期 PTB 的代谢生物标志物以及重要代谢物与参与者基因型的关联。使用 1H 核磁共振 (NMR) 代谢组学和全基因组筛查对妊娠 16 周和 20 周时从曾经历过 PTB（高风险）和未经历过 PTB 的女性（低风险对照）中收集的母体血清进行分析微阵列。对谱箱进行方差分析和概率神经网络 (PNN) 建模。应用来自同一参与者的光谱箱和基因型数据的代谢组学全基因组关联（mGWAS）来确定潜在的代谢物基因途径。通过方差分析获得 PTB 病例和对照之间的苯丙氨酸、乙酸盐和乳酸代谢差异，PNN 在第 20 周显示出较强的预测能力（AUC=0.89）。 MGWAS 确定了几个具有很强遗传关联的代谢物箱。 Cis-eQTL 分析强调了与妊娠第 20 周乳酸相关的非编码 SNP 局部的 TRAF1（参与炎症通路）。一组明确的参与者的 MGWAS 强调了乳酸与 TRAF1 的关系，这可能会导致 PTB。