A New Fast Phasing Method Based On Haplotype Subtraction.

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Errors were identified in existing phased 1000 Genomes Project data.
• Existing genetic phasing methods may rely on inference of new alleles, potentially introducing inaccuracies.
• A need exists for a more robust and accurate genetic phasing approach.

Purpose of the Study:

• To develop a novel genetic phasing method that relies solely on molecules and genotype frequency.
• To address and correct errors found in the 1000 Genomes Project data.
• To create a faster and more accurate algorithm for genetic phasing.

Main Methods:

• Developed an algorithm that clusters identical genotypes by frequency.
• Utilized homozygotes to define alleles and identify new putative alleles.
• Iteratively confirmed putative alleles using independent genotypes.
• Validated the approach using single-molecule sequencing data.

Main Results:

• The novel method identified 145 alleles across eight loci.
• An average of 98.2% of 1000 Genomes Project individuals were explained at these loci.
• The new method demonstrated comparable accuracy to PHASE and SHAPEIT2.
• The algorithm was significantly faster, being 14.6-fold faster than PHASE and 10.8-fold faster than SHAPEIT2.

Conclusions:

• The developed phasing approach is accurate and efficient.
• This method provides a reliable alternative for genetic phasing, particularly for large-scale projects.
• The approach successfully addresses limitations in previous genetic data phasing.