Abstract
Background/Objectives: Diagnosing ASD in adults presents unique challenges, and there are currently no specific biomarkers for this condition. Most existing studies on the gut microbiota in ASD are conducted in children; however, the composition of the gut microbiota in children differs significantly from that of adults. This study aimed to study the gut microbiota of young adults with high-functioning ASD. Methods: Using metagenomic sequencing, we evaluated the gut microbiota in 45 adults with high-functioning ASD and 45 matched healthy controls. Results: Adjusting for sociodemographic information, dietary habits, and clinical data, we observed a distinct microbiota profile of adults with ASD in comparison to controls, with the intensity of autistic traits strongly correlating to microbial diversity (correlation coefficient = -0.351, p-value < 0.001). Despite a similar dietary pattern, the ASD group exhibited more gastrointestinal symptoms than the healthy controls. An internally validated machine-learning predictive model that combines the Autism Spectrum Quotient questionnaire score of individuals with their microbial features could achieve an area under the receiver operating characteristic curve (AUC) of 0.955 in diagnosing ASD in adults. Conclusions: This study evaluates the gut microbiota in adult ASD and highlights its potential as a non-invasive biomarker to enhance the diagnosis of ASD in this population group.