HTTMM

HTTMM estimates taxon abundances from metagenomic short reads by modeling genome-specific and homologous reads on hierarchical taxonomy trees to improve microbial community composition inference.

Key Features:

• Hierarchical Structure Utilization: Represents genome-specific reads at leaf nodes and homologous reads at intermediate nodes by integrating the hierarchical taxonomy tree structure.
• Expectation-Maximization Algorithm: Employs an expectation-maximization algorithm to iteratively solve the mixture model and optimize abundance estimates.
• Superior Performance: Demonstrates improved accuracy relative to flat mixture models and lowest common ancestry-based methods on simulated and real-world data.
• Revealing Homologous Genomes: Identifies homologous genomes within microbial communities that are not resolved by flat or LCA-based approaches.

Scientific Applications:

• Metagenomic Abundance Estimation: Estimates relative abundances of taxa from sequencing short reads in metagenomic samples.
• Microbiome Diversity and Composition Analysis: Characterizes microbial community composition across ecosystems and animal tissues.
• Environmental Influence Studies: Supports investigations into how environmental factors influence microbial composition and function.

Methodology:

Inputs are sequencing short reads from metagenomic samples; the computational framework is a hierarchical taxonomy tree-based mixture model that integrates taxonomy tree structure; model parameters are estimated using an expectation-maximization algorithm to iteratively refine abundance estimates.