Libra

Libra performs large-scale all-vs-all metagenome comparisons using k-mer analysis to quantify microbial community biodiversity and functional relationships from shotgun metagenomic reads.

Key Features:

• Scalable Architecture: Uses a Hadoop distributed framework to process massive metagenomic datasets.
• All-vs-All Comparison: Performs comprehensive pairwise comparisons between metagenomes based on k-mer content to support clustering and comparative analyses.
• Cosine Similarity Metric: Computes Cosine Similarity between k-mer abundance profiles, accounting for sequence composition and abundance and normalizing for sequencing depth.
• Full-read k-mer Quantification: Calculates k-mer abundances from the entirety of sequencing reads while avoiding read count normalization or presence/absence reduction.

Scientific Applications:

• De novo comparative metagenomics: Comparing metagenomes based on total genetic content without relying on reference databases.
• Linking microbial signatures to processes: Associating metagenomic k-mer signatures with microbial biodiversity and functional dynamics.
• High-resolution shotgun metagenomics: Enabling large-scale, high-resolution analysis of shotgun metagenomic datasets across cohorts or environmental surveys.

Methodology:

Calculates k-mer abundances from all reads, performs all-vs-all pairwise comparisons using Cosine Similarity with sequencing-depth normalization, and leverages a Hadoop distributed computing framework for scalability.