P-smoother: efficient PBWT smoothing of large haplotype panels

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Large haplotype panels are crucial for genetic studies.
• Exact haplotype matching is hindered by recent mutations and genotyping errors.
• Existing algorithms passively tolerate mismatches, limiting accuracy.

Purpose of the Study:

• To develop an efficient algorithm for correcting mismatches in large haplotype panels.
• To enhance the accuracy of identifying shared haplotypes and identical-by-descent (IBD) segments.
• To enable state-of-the-art performance in multiway IBD segment detection.

Main Methods:

• Proposed P-smoother, a PBWT-based smoothing algorithm.
• Implemented bidirectional PBWT scanning to correct mismatching alleles using the IBD prior.
• Evaluated performance on simulated data and UK Biobank data.

Main Results:

• P-smoother reliably corrected 85% of errors in a simulated panel with a 0.2% error rate.
• Smoothed panels enabled PBWT algorithms to identify more pairwise IBD segments.
• The PS-cluster algorithm achieved state-of-the-art performance in identifying multiway IBD segments.

Conclusions:

• P-smoother effectively corrects errors in haplotype panels, improving IBD detection.
• PS-cluster demonstrates state-of-the-art efficiency and accuracy for multiway IBD analysis.
• P-smoother enables new possibilities for error-tolerant algorithms in biobank-scale genomics.