Analyzing genome coverage profiles with applications to quality control in metagenomics

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genome coverage, a measure of sequencing reads mapped to a genome, indicates experimental irregularities.
• Current computational genomics algorithms underutilize coverage profile information, leading to unaccounted biases from fragmented or erroneous reference genomes.
• Accessible coverage profile analysis can enhance sequencing experiment quality and quantitative analyses.

Purpose of the Study:

• To introduce a novel framework for fitting probability distribution mixtures to genome coverage profiles.
• To develop new analysis strategies for assessing reference genome validity using (meta-) genomic read data.
• To improve the quality and accuracy of genomic and metagenomic analyses by exploiting full coverage profile information.

Main Methods:

• Fitting mixtures of probability distributions to genome coverage profiles using the Expectation-Maximization algorithm.
• Incorporating distributions tailored to common sequencing artifacts.
• Developing and applying new analysis strategies focused on metagenomics and reference genome validation.

Main Results:

• The framework successfully fits mixture models to genome coverage profiles, accounting for common artifacts.
• New analysis strategies were developed and evaluated on simulated and real-world metagenomic data.
• The study computed the validity of 75 microbial genomes in a large-scale metagenomic study, highlighting the impact of reference genome quality.

Conclusions:

• The choice and quality of reference genomes are critical for the accuracy of metagenomic analyses.
• Validation of genome coverage profiles is essential to prevent erroneous conclusions in genomic studies.
• The developed framework provides a powerful tool for assessing genome coverage and improving data quality.