Hierarchically Labeled Database Indexing Allows Scalable Characterization of Microbiomes.

Area of Science:

• Microbiology
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing generates vast microbial data, necessitating efficient analysis tools.
• Accurate read classification is challenging due to shared matches among reference sequences.

Purpose of the Study:

• To develop a scalable and efficient method for microbial read classification.
• To address the challenge of multi-matching in sequence read analysis.
• To enable detailed interpretation of microbial community composition and function.

Main Methods:

• Developed PRROMenade, a classifier enhancing the generalized Burrows-Wheeler transform with a labeling step.
• Utilized annotation trees to assign reads to the lowest taxonomic unit.
• Simulated data with 5% sequence differences to assess accuracy.

Main Results:

• PRROMenade achieved a low error rate (1.5%) for functional classification in simulations.
• The classifier efficiently handles variable-size sequence classification.
• Applied to metatranscriptomic data, PRROMenade identified diet-induced functional pathways in the human gut microbiome.

Conclusions:

• PRROMenade offers a scalable and accurate solution for microbial read classification.
• The method effectively resolves multi-matching issues in sequence analysis.
• PRROMenade aids in understanding microbial community function, particularly in response to environmental factors like diet.