Towards omics-based predictions of planktonic functional composition from environmental data.

Area of Science:

• Marine microbiology
• Computational biology
• Oceanography

Background:

• Marine microbes are essential for climate regulation, biogeochemical cycles, and marine food webs.
• Vast amounts of planktonic community data necessitate advanced data-driven methods for functional prediction.

Purpose of the Study:

• To develop and apply a network-based approach to analyze marine metagenome-assembled genomes.
• To quantify and predict ecosystemic functions of marine microbes using machine learning and environmental context.

Main Methods:

• Reanalyzed 885 marine metagenome-assembled genomes using a network-based approach.
• Detected 233,756 protein functional clusters, with 15% unannotated.
• Employed machine learning to investigate cluster distributions and predict abundances based on environmental factors.

Main Results:

• Identified biogeographical provinces as key predictors of protein functional cluster abundance.
• Found that 14,585 clusters, including 1347 unannotated ones, are predictable from environmental context.
• Highlighted the Mediterranean Sea as an outlier in protein functional cluster composition.

Conclusions:

• The developed approach enables quantitative predictions of marine microbial functional composition from environmental data.
• This methodology is applicable to any sequence dataset, advancing our understanding of marine ecosystem functions.
• Identified predictable microbial functions and highlighted unique oceanic regions like the Mediterranean Sea.