A SNP discovery method to assess variant allele probability from next-generation resequencing data

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate genetic variant identification is crucial for large-scale genomic projects like the 1000 Genomes Project.
• Distinguishing true single nucleotide polymorphisms (SNPs) from sequencing errors is a major challenge in next-generation sequencing (NGS) data analysis.
• Understanding base call error probabilities is essential for reliable SNP discovery.

Purpose of the Study:

• To develop a computational tool, Atlas-SNP2, for accurate SNP detection from NGS data.
• To account for systematic sequencing errors influenced by context-related variables.
• To estimate the posterior error probability for substitutions to differentiate true SNPs from errors.

Main Methods:

• Developed Atlas-SNP2, a computational tool utilizing a logistic regression model trained on datasets.
• Incorporated context-related variables to detect and account for systematic sequencing errors.
• Employed a Bayesian formula to estimate posterior error probabilities, integrating prior error rates and SNP rates.

Main Results:

• Atlas-SNP2 effectively distinguishes true SNPs from sequencing errors.
• Achieved a false-positive rate below 10%.
• Demonstrated a false-negative rate of approximately 5% or lower.

Conclusions:

• Atlas-SNP2 provides a robust method for accurate SNP identification in NGS data.
• The tool's ability to account for systematic errors improves variant calling reliability.
• The low false-positive and false-negative rates make Atlas-SNP2 valuable for genomic research.