An empirical Bayes mixture model for SNP detection in pooled sequencing data.

Area of Science:

• Genomics
• Bioinformatics
• Statistical Genetics

Background:

• Detecting single-nucleotide polymorphisms (SNPs) in pooled sequencing data presents challenges due to sampling variations.
• Accurate estimation of sequencing errors is crucial for differentiating true SNP signals from errors.

Purpose of the Study:

• To propose an empirical Bayes mixture (EBM) model for robust SNP detection and allele frequency estimation in pooled sequencing data.
• To enhance the sensitivity and reliability of SNP detection by accounting for pooled sequencing data characteristics.

Main Methods:

• Developed an empirical Bayes mixture (EBM) model tailored for pooled sequencing data.
• The model learns error distributions by integrating information across pools and genomic positions.
• Implemented methods for flexible and robust estimation and control of the local false discovery rate for large-scale inference.

Main Results:

• The EBM model effectively learns error distributions, improving SNP detection sensitivity.
• Demonstrated robust performance in both simulation studies and real-world data applications.
• Provides a flexible framework for allele frequency estimation and SNP discovery.

Conclusions:

• The proposed EBM model offers a significant advancement for SNP detection in pooled sequencing data.
• This method improves accuracy and sensitivity by effectively modeling sequencing errors.
• The implementation is publicly available for broader research use.