Randomized Lasso Links Microbial Taxa with Aquatic Functional Groups Inferred from Flow Cytometry

Area of Science:

• Microbial Ecology
• Aquatic Microbiology
• Bioinformatics

Background:

• High-nucleic-acid (HNA) and low-nucleic-acid (LNA) bacteria are distinct functional groups in aquatic systems identified by flow cytometry (FCM).
• Previous research suggests HNA cell density correlates with heterotrophic production, while LNA does not, but the specific taxa involved remain unclear.

Purpose of the Study:

• To identify specific bacterial taxa associated with HNA and LNA groups in freshwater lakes.
• To understand the relationship between bacterial community structure and heterotrophic productivity using an integrated approach.

Main Methods:

• Applied machine learning variable selection to integrate flow cytometry (FCM) and 16S rRNA gene sequencing data from 14 freshwater lakes.
• Correlated bacterial heterotrophic production with HNA and LNA cell abundances.
• Identified and analyzed operational taxonomic units (OTUs) associated with HNA and LNA groups.

Main Results:

• Strong association found between bacterial heterotrophic production and HNA cell abundances (R² = 0.65), confirming HNA as key contributors to carbon cycling.
• Machine learning models successfully predicted HNA and LNA abundances using taxonomic data across all levels.
• Identified OTUs were largely lake-specific (89.5%–99.2%), with a subset showing potential plasticity between HNA and LNA groups (12.5%–33.3%).

Conclusions:

• HNA and LNA bacteria represent functionally distinct groups, with HNA bacteria disproportionately influencing carbon cycling in freshwater ecosystems.
• This study provides a framework for identifying system-specific ecological indicators of heterotrophic productivity.
• The findings highlight the potential for rare or phenotypically plastic taxa to significantly contribute to ecosystem functioning.