HapCompass: a fast cycle basis algorithm for accurate haplotype assembly of sequence data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Current genome assembly methods struggle with haplotype phase ambiguity due to sequencing technology limitations.
• Accurate haplotype phasing is essential for genetic association studies and genomic imputation but is computationally challenging and expensive.
• Existing computational haplotype assembly methods are inaccurate for large datasets or use unrealistic optimizations for modern sequencing.

Purpose of the Study:

• To introduce HapCompass, a novel algorithm for accurate haplotype assembly of densely sequenced human genome data.
• To address the limitations of current haplotype assembly methods in terms of accuracy and scalability.
• To provide a freely available tool for researchers.

Main Methods:

• Developed HapCompass, a graph-based algorithm where single nucleotide polymorphisms (SNPs) are nodes and sequence reads define edges.
• Haplotype phasing is modeled as finding spanning trees in the graph.
• Employed minimum weighted edge removal and cycle basis local optimizations to resolve conflicting evidence and estimate sequencing requirements.

Main Results:

• HapCompass demonstrated significantly improved accuracy compared to Genome Analysis ToolKit and HapCut on 1000 Genomes Project and simulated data.
• The algorithm effectively handles densely sequenced human genome data.
• Performance was evaluated using metrics from genome assembly and haplotype phasing.

Conclusions:

• HapCompass offers a substantial advancement in haplotype assembly accuracy and scalability for human genomics.
• The algorithm provides a more reliable solution for critical bioinformatics applications.
• HapCompass is available for download, facilitating its adoption in research.