Integrating widely targeted lipidomics and transcriptomics unravels aberrant lipid metabolism and identifies potential biomarkers of food allergies in rats

Oral food challenges (OFCs) are currently the gold standard for determining the clinical reactivity of food allergy (FA) but are time-consuming, expensive, and risky. To screen novel peripheral biomarkers of FA and characterize the aberrant lipid metabolism in serum, 24 rats were divided into four groups: peanut, milk, and shrimp allergy (PA, MA, and SA, respectively) and control groups, with six rats in each group, and used for widely targeted lipidomics and transcriptomics analysis. Widely targeted lipidomics revealed 144, 162, and 206 differentially accumulated lipids in PA, MA, and SA groups, respectively. We integrated widely targeted lipidomics and transcriptomics and identified abnormal lipid metabolism correlated with widespread differential accumulation of diverse lipids (including triacylglycerol, diacylglycerol, sphingolipid, and glycerophospholipid) in PA, MA, and SA. Simplified random forest classifier was constructed through five repetitions of 10-fold cross-validation to distinguish allergy from control. A subset of 15 lipids as potential biomarkers allowed for more reliable and more accurate prediction of FA. Independent replication validated the reproducibility of potential biomarkers. Our results revealed the major abnormalities in lipid metabolism and suggested the potential role of lipids as novel molecular signatures for FA.