Optimized filtering reduces the error rate in detecting genomic variants by short-read sequencing.

Area of Science:

• Genomics
• Bioinformatics

Background:

• Distinguishing true single-nucleotide variants (SNVs) from sequencing errors in whole-genome data is a significant challenge.
• Existing methods struggle with high error rates from short-read sequencing technologies.

Purpose of the Study:

• To develop and validate a set of filters and a software tool to reduce error rates in whole-genome sequencing.
• To improve the accuracy and reliability of variant detection, particularly for single-nucleotide variants.

Main Methods:

• Utilized whole-genome sequences from monozygotic twins to differentiate true variants from errors.
• Optimized thresholds for 12 individual filters and evaluated 1,048 filter combinations for sensitivity and specificity.
• Developed a freely accessible software tool implementing the optimized filters.

Main Results:

• The cumulative application of effective filters reduced the error rate by 290-fold.
• Successfully identified genetic differences between monozygotic twins.
• Demonstrated reliable identification of somatic mutations in tumor genomes and variants in the NA19240 HapMap genome.

Conclusions:

• The developed filtering strategy and software tool significantly enhance variant detection accuracy in whole-genome sequencing.
• This approach facilitates the identification of true genetic variations, including subtle differences between individuals and somatic mutations.