TinkerHap-a novel read-based phasing algorithm with integrated multimethod support for enhanced accuracy.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate genetic phasing is crucial for understanding genetic variation and identifying disease-causing variants.
• Traditional phasing methods struggle with rare variants and external data dependencies.

Purpose of the Study:

• To develop a novel phasing algorithm, TinkerHap, that overcomes limitations of existing methods.
• To integrate read-based phasing with external phased data for improved accuracy and contiguity.

Main Methods:

• Developed TinkerHap, a hybrid phasing algorithm combining a pairwise distance-based unsupervised classification read-phaser with external phased data.
• Evaluated TinkerHap using 1,040 UK Biobank parent-offspring trios (short reads) and a GIAB Ashkenazi trio (long reads).

Main Results:

• TinkerHap's read-based phaser alone surpassed other algorithms in accuracy (95.1% short reads, 97.5% long reads).
• The hybrid approach achieved 96.3% accuracy for short reads, phasing 99.5% of heterozygous sites.
• Extended haplotype block sizes (median 79,449 bp for long reads) and improved accuracy for SNPs and indels.

Conclusions:

• TinkerHap offers a powerful and versatile tool for genomic analysis.
• Its robust read-based algorithm and hybrid integration enhance phasing accuracy, contiguity, and comprehensiveness.
• Enables more effective genomic studies across various sequencing platforms and variant types.