PhenoGMM

PhenoGMM models microbial phenotypic distributions with Gaussian mixture models to integrate flow cytometry and 16S rRNA gene amplicon sequencing data for predictive characterization of microbial diversity.

Key Features:

• Model-Based Fingerprinting Approach: Implements Gaussian mixture models (GMMs) for model-based cytometric fingerprinting to characterize microbial communities by phenotypic properties measured by flow cytometry.
• Rapid Estimation of Microbial Diversity: Enables rapid estimation and prediction of taxonomic diversity by correlating cytometric fingerprints with 16S rRNA gene amplicon sequencing using supervised machine learning.
• Integration of Multi-Modal Data: Integrates flow cytometry and genetic (16S rRNA gene amplicon sequencing) data to link cytometric phenotypes with taxonomic composition.
• Extensive Validation Across Ecosystems: Has been evaluated on multiple datasets from diverse ecosystems and benchmarked against generic binning fingerprinting approaches.
• Facilitation of Microbial Diversity Studies: Provides a quantitative workflow for analyzing flow cytometry data to study microbial community structure and dynamics across ecosystems.

Scientific Applications:

• Ecological Research: Characterizes microbial diversity and community structure in ecological studies using integrated cytometric and 16S data.
• Environmental Microbiology: Predicts taxonomic diversity and links phenotypic profiles to microbial taxa in environmental microbiology investigations.
• Biogeochemical Cycle Studies: Supports studies of microbial roles in biogeochemical cycles by enabling rapid phenotypic-to-taxonomic inference.
• Ecosystem Health Monitoring: Facilitates monitoring of ecosystem health through quantitative assessment of microbial community changes from flow cytometry and sequencing correlations.
• Microbial Community Dynamics: Enables analysis of temporal and spatial dynamics of microbial communities by combining phenotypic fingerprints with 16S-based taxonomic information.

Methodology:

Applies Gaussian mixture models for model-based fingerprinting of flow cytometry data, correlates cytometric fingerprints with 16S rRNA gene amplicon sequencing, uses supervised machine learning for prediction, and compares performance against generic binning fingerprinting approaches using multiple datasets.