Shi Junzhu, Zeng Shanshui, Zhang Yonggang, Zuo Zhihua, Tan Xiaoyu
Department of Clinical Laboratory, Shenzhen Longhua District Central Hospital, Shenzhen 518110, China.
Department of Clinical Laboratory, Guangzhou Women and Children's Medical Center, Guangzhou 510623, China.
Clin Chim Acta. 2023 Jan 15;539:206-214. doi: 10.1016/j.cca.2022.12.012. Epub 2022 Dec 22.
There are no approaches for the early detection of pre-eclampsia (PE). Using parallel reaction monitoring proteomics, we investigated 79 maternal serum protein changes before PE onset and its predictive capability.
We conducted a nested case-control study with 60 PE patients and 60 normotensive pregnant women matched for age and gestational week. These differentially expressed proteins were quantified using the data-dependent acquisition (DDA) combined parallel response monitoring (PRM) approach, and a PE prediction model was developed using the least absolute shrinkage and selection operator (LASSO) regression. We further examined the link between these biomarkers and PE using bioinformatics. ELISA assay was used to investigate the expression and clinical significance of the critical variables.
Among the 79 analyzed proteins, we identified 11 serum proteins with significantly abnormal expression. Fibrinogen beta chain (FGB) was likely connected with the progression of PE due to the positive correlation between their levels and those of hypertension and proteinuria. In addition, an early prediction model for PE with an AUC of 92% was developed using LASSO regression.
Our research employs predictive algorithms and screens for relevant variables, which could result in a potential approach to enhancing PE prediction.
目前尚无子痫前期(PE)的早期检测方法。我们采用平行反应监测蛋白质组学技术,研究了79种母体血清蛋白在PE发病前的变化及其预测能力。
我们进行了一项巢式病例对照研究,纳入60例PE患者和60例年龄及孕周匹配的血压正常孕妇。使用数据依赖采集(DDA)联合平行反应监测(PRM)方法对这些差异表达蛋白进行定量,并使用最小绝对收缩和选择算子(LASSO)回归建立PE预测模型。我们进一步利用生物信息学研究这些生物标志物与PE之间的联系。采用ELISA检测法研究关键变量的表达及临床意义。
在分析的79种蛋白质中,我们鉴定出11种血清蛋白表达显著异常。纤维蛋白原β链(FGB)的水平与高血压和蛋白尿水平呈正相关,可能与PE的进展有关。此外,使用LASSO回归建立了AUC为92%的PE早期预测模型。
我们的研究采用预测算法并筛选相关变量,这可能为增强PE预测提供一种潜在方法。