High throughput recurrent pregnancy loss screening: urine metabolic fingerprints via LDI-MS and machine learning

Abstract

Infertility is a significant challenge faced by many families worldwide, with recurrent pregnancy loss (RPL) being a prevalent cause of infertility among women. This condition causes immense emotional and physical distress for affected individuals and their families. In this study, we present a rapid, efficient, and high-throughput analytical method using PS@Fe₃O₄-NH₂ magnetic beads as a matrix for the detection of urinary metabolite fingerprints in RPL patients via laser desorption/ionization mass spectrometry (LDI-MS) combined with machine learning (ML). This approach offers rich metabolic information from urine samples, through subsequent analysis we identify 17 metabolites that significantly differ between RPL patients and healthy controls (HC). The application of mass spectrometry features in conjunction with ML enabled effective screening of RPL patients and the identification of dysregulated metabolic pathways. This method presents a promising, non-invasive, and rapid screening approach for early detection of RPL, facilitating timely intervention and contributing to women's health.