Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• The standard human reference genome is limited for genotyping due to its representation of few individuals.
• Accurate genotyping requires comprehensive genomic data, including variants and haplotypes.

Purpose of the Study:

• To introduce HISAT2, a novel alignment method for DNA and RNA sequences.
• To develop an expanded human reference genome model incorporating population-level genetic variation.
• To improve the accuracy and detail of variant analyses and genotyping applications.

Main Methods:

• Developed HISAT2, utilizing a graph Ferragina Manzini index for sequence alignment.
• Integrated over 14.5 million genomic variants and haplotypes into the HISAT2 data structure.
• Benchmarked HISAT2 against existing methods using simulated and real datasets.

Main Results:

• HISAT2 provides more detailed and accurate variant analyses compared to other methods.
• The HISAT-genotype software, leveraging HISAT2, demonstrates superior performance in HLA typing and DNA fingerprinting.
• HISAT-genotype matches or exceeds the performance of laboratory-based assays.

Conclusions:

• Representing a population of genomes with a fast, memory-efficient algorithm significantly enhances variant analysis.
• HISAT2 and HISAT-genotype offer advanced computational tools for haplotype-resolved genomic analyses.
• The developed method provides a powerful alternative to traditional genotyping and laboratory assays.