The abstract is read at the publisher; the summary is JClub's.
GRASER phases graph-aligned short reads by using haplotype-path evidence directly in a pangenome graph, assigning 22.6–24.5% of simulated fragments with >96% correct placement.
- Why it matters: Linear-reference surjection can reintroduce reference bias, limiting haplotype inference when reads span non-reference graph sequence.
- What they did: GRASER builds per-individual haplotype paths via graph genotyping plus statistical phasing, then phases reads using edges unique to one haplotype path from graph alignments.
- The result: Compared with a WhatsHap-based linear-reference workflow, GRASER assigned more fragments in every tested configuration and, in whole-genome bisulfite sequencing, recovered allele-specific methylation at imprinted loci and haplotype-resolved methylation within a structural variant.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.