Topic modeling for multi-omic integration in the human gut microbiome and implications for Autism

Area of Science:

• Microbiome research
• Computational biology
• Neuroscience

Background:

• Healthy gut microbiomes are crucial for human health, but microbial processes are often poorly understood due to limitations in profiling techniques.
• Multi-omic analysis offers a way to define generalizable microbial processes, particularly for complex conditions like Autism.
• Integrating heterogeneous data from multiple profiling methods presents challenges, but Latent Dirichlet Allocation (LDA) can overcome these.

Purpose of the Study:

• To apply Latent Dirichlet Allocation (LDA) to multi-omic microbial data from children with and without Autism.
• To identify microbial processes (topics) and associated metabolites that differ between these groups.
• To define generalizable microbial processes across different profiling methods.

Main Methods:

• Applied Latent Dirichlet Allocation (LDA) to multi-omic microbial data (16S rRNA amplicon, shotgun metagenomic, shotgun metatranscriptomic, and untargeted metabolomic profiling) from 81 children.
• Identified topics representing microbial processes and subset samples by topic distribution.
• Analyzed differences in metabolites, specifically neurotransmitter precursors and fatty acid derivatives, between children with and without Autism.

Main Results:

• Identified diet-associated microbial processes (topics) within the gut microbiome.
• Discovered significant differences in specific metabolites (neurotransmitter precursors, fatty acid derivatives) between children with and without Autism.
• Identified "cross-omic topics" representing generalizable microbial processes observable across different profiling methods.

Conclusions:

• Latent Dirichlet Allocation (LDA) is effective for integrating multi-omic microbial data to identify diet-associated microbial processes.
• Specific microbial processes and metabolites in the gut microbiome are associated with Autism in children.
• The study provides a framework for understanding complex host-microbiome interactions in Autism using computational approaches.