GraphChainer

GraphChainer aligns erroneous long reads to string-labeled acyclic variation graphs to improve read mapping accuracy and coverage for pangenomic analyses.

Key Features:

• Co-linear Chaining Algorithm: GraphChainer co-linearly chains multiple seed occurrences within string-labeled acyclic variation graphs to link seeds into contiguous alignments.
• Comparative Alignment Performance: On real PacBio reads from human chromosomes 1, 22, and the entire human pangenome, GraphChainer aligned 12%–17% more reads and 21%–28% more total read length than GraphAligner.
• High Alignment Coverage: Achieves alignment coverage between 95% and 99% for both simulated and real datasets.
• Competitive Accuracy: Demonstrated substantially higher accuracy compared to minigraph and minichain, which achieved less than 60% accuracy in comparable settings.

Scientific Applications:

• Pangenomics: Improves alignment of long reads to variation graphs to facilitate more accurate variant calling and genome assembly across population-scale pangenomes.
• Genomic Research: Enables exploration of complex genomic regions using PacBio long reads with higher-confidence mappings for applications in personalized medicine and evolutionary biology.

Methodology:

Performs co-linear chaining of seeds by identifying candidate read occurrences through extending the best seeds across the variation graph and linking those occurrences in linear order within the string-labeled acyclic graph.