K2Mem

K2Mem enhances metagenomic read classification by discovering and incorporating discriminative k-mers from input sequencing data to improve taxonomic annotation of sequencing reads.

Key Features:

• Discriminative k-mer discovery: Identifies novel k-mers directly from input sequencing reads that distinguish between different taxa.
• Reference k-mer library augmentation: Augments existing reference k-mer libraries with discovered discriminative k-mers.
• Improved classification metrics: Increases recall and F-measure while maintaining high precision in metagenomic classification.
• Robustness to reference divergence: Addresses sequence variation and highly mutated genomes by incorporating sample-derived k-mers when closely related reference genomes are unavailable.
• Empirical evaluation: Performance has been evaluated under multiple conditions and compared against existing metagenomic classification tools.

Scientific Applications:

• Taxonomic annotation of metagenomic reads: Improves assignment of sequencing reads to taxa in complex samples.
• Species identification in environmental and clinical samples: Supports detection of species where reference genomes may be incomplete or divergent.
• Analysis of rapidly evolving organisms: Enhances classification of highly mutated viral genomes and other rapidly evolving taxa.
• Metagenomic sensitivity and specificity improvement: Increases sensitivity without compromising specificity in metagenomic studies.

Methodology:

Analyzes input sequencing reads to discover unique discriminative k-mers and augments the existing reference k-mer library with those k-mers.