TaxiBGC

TaxiBGC identifies experimentally verified biosynthetic gene clusters (BGCs) and infers their associated secondary metabolites (SMs) from metagenomic shotgun sequencing data.

Key Features:

• Taxonomy-guided approach: Pinpoints microbial species likely to harbor specific BGCs rather than directly mapping reads to BGC genes.
• Identification of experimentally verified BGCs and SM inference: Detects experimentally characterized BGCs and predicts their SM production potential from metagenomes.
• Improved predictive performance: Benchmarking on simulated metagenomes yielded a mean F1 score of 0.56 and mean positive predictive value (PPV) of 0.80 versus 0.49 F1 and 0.41 PPV for direct read-mapping approaches.
• Application to large human metagenome collections: Applied to over 2,650 metagenomes from the Human Microbiome Project and case-control gut microbiome studies to map BGCs and SMs across body sites and disease cohorts.
• Ecological and disease associations: Associates BGCs encoding bacteriocins with distinct human body sites and diseases and identifies shared gut microbial BGC associations among multiple pathologies.

Scientific Applications:

• Chemical ecology of microbial communities: Enables exploration of BGC-encoded metabolite distributions across niches to study microbe–microbe and host–microbe chemical interactions.
• Microbiome–disease association studies: Supports association of specific BGCs and SMs with gastrointestinal diseases such as Crohn’s disease and liver cirrhosis.
• Bacteriocin ecology and stability: Facilitates investigation of bacteriocin-encoding BGC roles in microbial ecosystem stability across human body sites.
• Comparative pathology analyses: Enables identification of shared gut BGC associations across multiple pathologies for cross-disease comparisons.

Methodology:

Uses a taxonomy-guided strategy to identify species likely to harbor experimentally verified BGCs and infer SMs, benchmarks predictions on simulated metagenomes against direct read-mapping approaches, and was applied to over 2,650 Human Microbiome Project and case-control gut metagenomes.