#AI reads Urine# Urine Metabolomic Profiling in Autism Spectrum Disorder Diagnosis
Published 28 May, 2025
This article describes a study on the diagnosis of autism spectrum disorder. The study collected first-morning urine samples from 52 children, including 32 children with autism and 20 neurotypical control children. Liquid chromatography-mass spectrometry was used to detect 293 metabolites in the samples, which were divided into 189 endogenous metabolites and 104 exogenous metabolites. Various machine learning classifiers such as random forest and logistic regression were applied to analyze the data, and 10-fold cross-validation was used to distinguish between the autism group and the control group. The results showed that when the random forest classifier used all 293 metabolites, the accuracy rate reached 85%, and the area under the curve was 0.9. When only endogenous metabolites were used, the accuracy rate was 85%, and the area under the curve was 0.86. However, when only exogenous metabolites were used, the accuracy rate was lower, at 71%, and the area under the curve was 0.72. The study indicates that urine metabolomic analysis, particularly of endogenous metabolites, has the potential to serve as an auxiliary diagnostic tool for autism. The high accuracy of machine learning classifiers demonstrates the feasibility of developing auxiliary diagnostic methods based on metabolite profiles, although further research is needed to link these profiles to specific behavioral characteristics and autism subtypes.
Metabolites. 2025 May 16;15(5):332. doi: 10.3390/metabo15050332.
Youhe Gao
Statement: During the preparation of this work the author(s) used Doubao / AI reading for summarizing the content. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.
For earlier AI Reads Urine articles：