Efficient indexing and querying of annotations in a pangenome graph

Area of Science:

• Genomics
• Bioinformatics

Background:

• Genomic annotations are crucial for understanding biological function.
• Current annotation formats (VCF, BED) are limited for pangenomic references.
• Efficient tools for pangenomic annotation are needed.

Purpose of the Study:

• To develop and implement efficient tools for indexing and querying the Graph Alignment Format (GAF) for pangenomic annotations.
• To extend existing bioinformatics tools (vg, HTSlib) for enhanced GAF support.
• To enable projection and analysis of diverse genomic annotations within pangenome graphs.

Main Methods:

• Extensions to vg and HTSlib for sorting, indexing, and querying GAF files.
• Compression of GAF files using BGZIP and indexing with HTSlib/tabix.
• Updated vg annotate tool to generate GAF files from linear sequence annotations (BED, GFF3).

Main Results:

• GAF files are shown to be twice as fast to sort and half the size of binary GAM format.
• Efficient querying of annotations overlapping subgraphs within pangenome graphs is achieved.
• Successful projection of human pangenome annotations, GWAS variants, and eQTLs into a pangenome graph.

Conclusions:

• The developed extensions provide efficient and accessible methods for pangenomic annotation using GAF.
• These tools facilitate the integration and analysis of diverse genomic data within complex pangenome structures.
• The approach enables visualization and querying of rich annotations in pangenome graphs, advancing genomic research.