Machine learning and causal inference applied to the gut metagenome-metabolome axis reveals a link between neonatal jaundice and autism spectrum disorder.

Both neonatal jaundice and ASD were associated with increased diversity of bile acid metabolism genes in the gut.

• Three cohorts were analyzed: NJ cohort 1 (68 neonates with NJ and 68 healthy controls), NJ cohort 2 (56 infants with NJ and 14 HCs), and ASD cohort (43 children with ASD and 31 typically developing children).
• Shotgun metagenomics was used to quantify bile acid metabolism genes in NJ cohort 2 and the ASD cohort.
• Increased diversity of bile acid metabolism genes was identified as having biomarker potential for both conditions.
• 16S rRNA sequencing and liquid chromatography with tandem mass spectrometry metabolomics were also performed on NJ cohort 1 and the ASD cohort.

The gut DNA virome was identified as a potential biomarker for both neonatal jaundice and ASD.

• Fecal samples were collected aseptically and analyzed for the gut DNA virome across cohorts.
• Causal mediation analysis showed that the gut DNA virome influences bile acid metabolism genes in both NJ and ASD conditions.
• The gut DNA virome was characterized as part of multi-omics integration using causal mediation and machine learning causal inference approaches.

Gut human betaherpesviruses were found to contribute to neonatal jaundice, mediated by gut bile acid-metabolizing bacteria.

• Machine learning-based causal modeling was used to identify this relationship.
• The effect of human betaherpesviruses on NJ was mediated by gut bile acid-metabolizing bacteria.
• This finding was derived from analysis of NJ cohort 2 using shotgun metagenomics.

Gut human mastadenoviruses were found to contribute to ASD, mediated by gut bile acid-metabolizing bacteria.

• Machine learning-based causal modeling identified human mastadenoviruses as contributing to ASD.
• The effect was mediated by gut bile acid-metabolizing bacteria.
• This finding was derived from analysis of the ASD cohort (43 children with ASD and 31 typically developing children) using shotgun metagenomics.

Causal mediation analysis demonstrated that the gut DNA virome influences bile acid metabolism genes in both neonatal jaundice and ASD.

• Multi-omics data were integrated using causal mediation analysis and machine learning causal inference.
• The causal pathway identified was: gut DNA virome → bile acid-metabolizing bacteria → bile acid metabolism gene diversity.
• This causal structure was observed consistently across both NJ and ASD conditions.

Dysbiosis of the gut DNA virome and bile acid-metabolizing bacteria may mechanistically explain the link between neonatal jaundice and ASD.

• Human epidemiological studies have established an association between perinatal pathogenic infections and ASD.
• The study found shared perturbations in virome composition and bile acid metabolism between NJ and ASD cohorts.
• The authors propose that regulating the bile acid-gut microbiota axis may be crucial for developing novel preventive and therapeutic strategies for ASD.
• Both conditions were linked to altered gut bile acid metabolism as measured by metabolomics and metagenomics.