IDDF2024-ABS-0406 Dissecting gut-brain axis: multi-kingdom and functional gut microbiota markers for autism spectrum disorder

<h3>Background</h3> Autism spectrum disorder (ASD) is a neurodevelopmental disorder commonly complicated by gastrointestinal symptoms. Associations between the gut microbiome and ASD have been investigated, although most studies have focused on the bacterial component of the microbiome. Whether gut archaea, fungi, and viruses, or the function of the gut microbiome, is altered in ASD is unclear. We aimed to identify robust multi-kingdom ASD-microbiome associations and identify their potential role in the gut-brain axis. <h3>Methods</h3> We performed metagenomic sequencing on faecal samples from 1,627 children (aged 1-13 years, 24.4% female) with or without ASD, with extensive phenotype data. The multi-kingdom gut microbiome was profiled by Kraken 2. Microbial pathways and genes were profiled by HUMAnN 3. A fully matched sub-cohort was constructed for machine learning, of which the performance was further validated in independent in-house and publicly available cohorts. <h3>Results</h3> Children with ASD showed a decrease in the diversity of archaea, bacteria, and viruses compared to children considered neurotypical. Alterations in bacteria species in children with ASD were driven by the depletion of Streptococcus thermophilus, and short-chain fatty acids-producing bacteria, such as Bacteroides sp. PHL2737 and Lawsonibacter asaccharolyticus. Neurotransmitters-related biosynthesis pathways of ubiquinol-7 and thiamine diphosphate were reduced in children with ASD, compared with children considered neurotypical. Machine learning using single kingdom panels showed an area under the curve (AUC) of 0.68 to 0.87 in differentiating children with ASD from those who are neurotypical. A panel of 31 multi-kingdom and functional markers showed a superior diagnostic accuracy with an AUC of 0.91, with comparable performance for males and females. Our model maintained AUCs of around 0.90 in two independent validation cohorts of different ages and an AUC of 0.78 in public datasets across Asia, Europe, and America. The accuracy of the model was predominantly driven by the biosynthesis pathways of ubiquinol-7 or thiamine diphosphate, which were less abundant in children with ASD, suggesting a relationship between microbial metabolism and ASD through the gut-brain axis. <h3>Conclusions</h3> Our findings highlight the involvement of multi-kingdom microbes in the gut-brain axis related to ASD.