Estimating Haplotype Structure and Frequencies: A Bayesian Approach to Unknown Design in Pooled Genomic Data

Area of Science:

• Genomics
• Computational Biology
• Statistical Genetics

Background:

• Haplotype structure and frequencies are vital for understanding genome composition.
• Existing methods often focus on single individuals or provide point estimates for pooled samples.
• Reconstructing haplotypes from pooled sequencing data presents unique challenges.

Purpose of the Study:

• To develop a Bayesian hierarchical model for estimating haplotype structure and frequencies from pooled sequencing data.
• To address the challenge of identifiability caused by permutation ambiguity in pooled samples.
• To quantify the uncertainty associated with haplotype reconstruction.

Main Methods:

• Proposed a Bayesian hierarchical model incorporating an order-preserving shrinkage prior.
• Introduced a blocked Gibbs sampler for efficient inference under permutation constraints.
• Validated the method using simulation studies and real-world pooled sequencing data.

Main Results:

• The Bayesian model successfully reconstructs haplotype structure and frequencies from pooled samples.
• The proposed prior and sampler effectively handle permutation identifiability issues.
• The method quantifies uncertainty in haplotype estimations, providing a more comprehensive analysis.

Conclusions:

• The developed Bayesian approach offers a robust and informative method for haplotype reconstruction from pooled sequencing data.
• This technique enhances genomic analysis by providing reliable haplotype estimates with uncertainty quantification.
• The model demonstrates strong performance in complex, high-dimensional genomic datasets.