Inferring Resource Competition in Microbial Communities from Time Series

Area of Science:

• Ecology
• Microbiology
• Computational Biology

Background:

• Microbial communities exhibit resource competition, organizing into guilds with shared resource preferences.
• Understanding individual taxa's resource needs is crucial for community structure and resource flow.
• Inferring metabolic capabilities and competition among taxa within communities remains challenging.

Purpose of the Study:

• To develop and validate superior methods for inferring resource competition structure in microbial communities.
• To address limitations of simple correlation methods in predicting ecological interactions.
• To leverage dynamic abundance data for a deeper understanding of community dynamics.

Main Methods:

• Utilized dynamic abundance measurements from microbial communities.
• Employed spectral methods, including cross-power spectral density and coherence, to analyze time-delayed effects.
• Validated methods on synthetic consumer-resource model data and real-world oceanic plankton time-series data.

Main Results:

• Simple correlations are often insufficient and misleading for predicting resource competition.
• Spectral methods accurately infer resource competition structure, accounting for time-delayed interactions.
• Applied spectral methods to oceanic plankton data revealed interaction structures among genetically similar species.

Conclusions:

• Spectral analysis of temporal data provides a robust framework for understanding microbial community competition.
• Time-delayed spectral methods offer superior insights into ecological interactions compared to static or simple correlative approaches.
• This approach can uncover hidden community structures and inter-species relationships across various timescales.