Haplotype-resolved de novo assembly using phased assembly graphs with hifiasm.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Haplotype-resolved de novo assembly is essential for comprehensive genome variation analysis.
• Current assemblers struggle with accurately separating heterozygous alleles, leading to incomplete phased assemblies.

Purpose of the Study:

• To develop a novel de novo assembler, hifiasm, capable of high-quality haplotype-resolved genome assembly.
• To leverage long, high-fidelity reads for faithful representation of haplotype information in assembly graphs.

Main Methods:

• Developed hifiasm, a de novo assembler designed to preserve contiguity of all haplotypes.
• Implemented a graph trio binning algorithm that improves upon standard trio binning methods.
• Utilized long, high-fidelity sequencing reads for assembly.

Main Results:

• Hifiasm successfully represents haplotype information within a phased assembly graph.
• The graph trio binning algorithm significantly enhances trio binning capabilities.
• Evaluated on diverse datasets, including a hexaploid redwood genome, hifiasm produced superior assemblies compared to existing tools.
• Hifiasm consistently outperformed other methods in haplotype-resolved assembly.

Conclusions:

• Hifiasm offers a significant advancement in de novo genome assembly, particularly for haplotype resolution.
• The assembler's ability to preserve all haplotype contiguities enables more accurate and complete phased genome assemblies.
• Hifiasm is a powerful tool for genomic research requiring high-quality, haplotype-resolved assemblies.