Genomic GC bias correction improves species abundance estimation from metagenomic data

Area of Science:

• Bioinformatics
• Microbiology
• Genomics

Background:

• Metagenomic sequencing is vital for understanding microbial communities and their role in diseases like colorectal cancer.
• GC-content-dependent biases can affect the accuracy of quantitative microbiome analyses.
• Existing methods struggle with accurate species abundance quantification due to these biases.

Purpose of the Study:

• To develop a computational method, GuaCAMOLE, for detecting and removing GC bias from metagenomic sequencing data.
• To provide unbiased species abundances for improved microbiome research.
• To enhance the accuracy and comparability of species abundance data across studies.

Main Methods:

• GuaCAMOLE algorithm compares individual species within a single sample to estimate GC-content-specific sequencing efficiency.
• It does not require calibration experiments or multiple samples.
• The method outputs corrected, unbiased species abundances.

Main Results:

• Analysis of 3435 gut microbiomes revealed significant variation in GC bias across studies.
• A common bias against GC-poor species was observed in existing methods.
• GuaCAMOLE corrected the abundance of GC-poor species, like F. nucleatum, by up to twofold.

Conclusions:

• GuaCAMOLE effectively detects and removes GC bias in metagenomic data.
• The method improves the quantitative accuracy of microbial species abundances.
• This contributes to a more robust understanding of microbiomes in health and disease.