Frequency-based haplotype reconstruction from deep sequencing data of bacterial populations

Area of Science:

• Population Genetics
• Genomic Sequencing
• Bioinformatics

Background:

• Clonal populations accumulate mutations, leading to diverse haplotypes.
• Reconstructing haplotypes from deep sequencing data is challenging, particularly in low-mutation clonal systems due to ambiguous phasing from limited segregating sites.

Purpose of the Study:

• To develop a novel haplotype reconstruction method, EVORhA, that addresses limitations of existing approaches in low-mutation rate clonal populations.
• To improve the accuracy of genome-wide haplotype reconstruction by integrating haplotype frequency estimations with sequence overlap information.

Main Methods:

• EVORhA combines haplotype phasing information from read overlaps with frequency estimations of inferred local haplotypes.
• The method's performance was evaluated using simulated data with varying read lengths and mutation rates.
• The applicability of EVORhA was demonstrated on real-world data from evolved bacterial populations and clinical samples.

Main Results:

• EVORhA reliably reconstructs genome-wide haplotypes even when read lengths are short or mutation rates are low, outperforming state-of-the-art methods under these restrictive conditions.
• The method accurately determined the population composition of evolved bacterial populations.
• EVORhA successfully decomposed mixed bacterial infections from clinical samples.

Conclusions:

• EVORhA offers a robust solution for haplotype reconstruction in clonal populations, especially those with low mutation frequencies.
• The integration of frequency information significantly enhances the accuracy of genome-wide haplotype phasing.
• EVORhA has practical applications in microbial population genetics and clinical diagnostics for analyzing complex bacterial communities.