Skin lipid chemistry influences host-microbiome-pathogen interactions in snake fungal disease (ophidiomycosis)

Within host-microbiome-pathogen systems, the host chemical microenvironment is often overlooked despite its inherent role in host physiology. We used a multifaceted experimental approach encompassing culture-dependent and -independent methods, genomic and metagenomic analyses, and deep neural network modeling to assess the impact of host skin chemistry and the bacterial microbiome on the growth of Ophidiomyces ophidiicola (ophidiomycosis, snake fungal disease). Results suggest that host skin lipid chemistry (e.g., oleic acid, squalene) and bacteria isolated from wild snake skins (e.g., Chryseobacterium sp. and Stenotrophomonas maltophilia) suppress O. ophidiicola growth. Notably, the O. ophidiicola genome contains biosynthetic gene clusters (BGCs) that encode metabolites that may suppress host lipid production, facilitating fungal pathogenicity. A contrastive deep neural network produced a near-perfect alignment of snake skin lipid and microbiome profiles for both individual snakes and disease states. BGCs from bacterial genomes isolated from snake skin overlap with metagenome profiles from wild snakes and are associated with disease state. We highlight antifungal activity found in the diverse lipid milieu of snake skin and bacterial-fungal interactions that structure the skin microbiome. Our results illustrate a strong relationship among a fungal pathogen, the microbiome, and host skin lipid chemistry that may underpin disease susceptibility. Deep learning models accurately align multimodal data including snake skin lipid chemistry and microbiome-pathogen interactions in the snake fungal disease system.