Metakallisto

Metakallisto performs pseudoalignment-based mapping and expectation-maximization–driven read assignment to quantify strains and microbial genomes in metagenomic sequencing data.

Key Features:

• Pseudoalignment Technique: Implements pseudoalignment (derived from RNA-Seq transcript quantification) to map metagenomic reads to reference sequences without full base-to-base alignment.
• Expectation-Maximization (EM) Algorithm: Applies an EM algorithm to iteratively estimate read assignment probabilities and resolve ambiguous mappings among similar strains.
• Strain-Level Quantification: Estimates abundances of individual microbial genomes and strains from metagenomic samples.
• Rapid Processing: Reduces computational time relative to traditional alignment-based methods, enabling large-scale strain-level analyses.

Scientific Applications:

• Ecological studies: Quantifies strain composition and dynamics in environmental microbial communities.
• Clinical microbiology: Profiles strain-level abundances relevant to disease mechanisms and clinical samples.
• Microbial diversity and evolution: Enables analysis of strain-level diversity, interactions, and evolutionary processes within complex communities.

Methodology:

Metagenomic reads are mapped to a reference database using pseudoalignment, followed by iterative refinement of read assignment probabilities via the EM algorithm.