Haplotype inference from short sequence reads using a population genealogical history model

Area of Science:

• Population genomics
• Bioinformatics
• Computational biology

Background:

• High-throughput sequencing generates vast amounts of short-read data, transforming biological research.
• Inferring haplotypes from population sequencing data is crucial for understanding genetic variation.
• Existing methods for haplotype inference face challenges with short-read data and recombination.

Purpose of the Study:

• To develop a novel computational method for haplotype inference directly from short-read sequencing data.
• To address the challenges posed by short reads and recombination in population genomics.
• To provide a robust and efficient approach for reconstructing haplotypes.

Main Methods:

• Formulation of the computational problem of haplotype inference with short reads.
• Development of a probabilistic model for short reads to infer haplotypes directly, bypassing genotype calling.
• Application of integer linear programming for optimal haplotype inference without recombination and a heuristic method for larger datasets with recombination.

Main Results:

• The proposed method successfully infers haplotypes by modeling their local genealogical history as a perfect phylogeny.
• Integer linear programming efficiently finds optimal haplotypes for smaller datasets without recombination.
• A heuristic method demonstrates effectiveness for larger datasets and accommodates recombination, showing competitive performance.

Conclusions:

• The developed method offers a direct and effective approach to haplotype inference from high-throughput sequencing data.
• The strategy of modeling genealogical history as a perfect phylogeny provides a powerful framework for population genomics.
• This research contributes a valuable tool for analyzing population genetic variation using modern sequencing technologies.