The abstract is read at the publisher; the summary is JClub's.
DiGi reconstructs paternal and maternal copies of 90% of key HLA genes absent from a pangenome reference, achieving sequence identity above 99.9% in target genomes.
- Why it matters: Pangenome genotyping can detect more variation than single-reference approaches, but diploid inference is hard because jointly finding two mosaic paths causes an exponential search-space blowup.
- What they did: DiGi builds a haplotype trellis graph that limits switching between reference haplotypes to predefined positions, then solves an integer programming problem using reads to infer a pair of haplotype paths with parameters set from read length and coverage.
- The result: On the HG002 Genome in a Bottle benchmark, DiGi attains favorable F1 scores for SNVs, indels, and structural variants in the MHC region and stays accurate across a wide range of sequencing coverages.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.