Ocean-resistome

Ocean-resistome analyzes the abundance, distribution, and dynamics of antibiotic resistance genes (ARGs) in marine metagenomic samples from the TARA Oceans project using machine learning to detect and quantify ARGs.

Key Features:

• Metagenomic dataset: Uses 293 TARA Oceans metagenomic samples to survey ARGs across diverse marine biomes.
• Machine learning integration: Applies advanced machine learning methodologies to detect and quantify ARGs in environmental open reading frames (ORFs).
• ARG catalog: Identified 99,205 ORFs as potential ARGs, categorized into 560 ARG families conferring resistance to 26 antibiotic classes.
• Mobile genetic elements: Identified 24,567 ORFs within plasmidial sequences and 4,804 contigs containing multiple ARGs, including contigs with up to five different ARGs.
• Horizontal gene transfer: Highlights potential horizontal gene transfer between clinical and natural environments via plasmid-like sequences that harbor multiple ARGs.
• Geographical and environmental correlation: Shows ARG distribution varies across biomes with certain classes enriched in coastal regions and MCR-1 family ARGs detected at high abundance in Polar biomes.
• Clinical relevance: Detects clinically relevant ARGs including 15 MCR-1 family ARGs associated with the genus Psychrobacter, an opportunistic human pathogen.

Scientific Applications:

• Global ARG surveillance: Enables characterization of the global distribution and evolution of ARGs in marine environments.
• Environmental association studies: Provides insights into how environmental factors and biomes influence ARG prevalence and composition.
• HGT pathway identification: Supports identification of potential pathways for ARG transfer between natural ecosystems and clinical settings, particularly via plasmid-mediated transfer.
• Microbial ecology and public health assessment: Informs microbial ecology studies and assessments of public health risk from marine ARG reservoirs.
• Mitigation strategy support: Informs development of strategies to mitigate spread of resistance genes.

Methodology:

Applies advanced machine learning methodologies to detect and quantify ARGs in environmental ORFs and includes extensive manual curation of metagenomic data.