Parsa Eskandar

The paper introduces GBZ-base and GAF-base, indexed pangenome file formats for efficient subgraph and alignment extraction.

460 haplotypes from 230 genomes, selected to maximize common-variant coverage, capture 99.43% of All of Us common small variants and halve structurally unreliable regions and base errors versus HPRC1.

The paper introduces a lossless indexing framework for pangenome graphs using tag arrays, enabling efficient haplotype-aware querying and coordinate translation across haplotypes.

The paper analyzes chromosome 22q11.2 haplotypes to show how structural variations in specific regions correlate with microdeletion and inversion risks across populations.

The paper uses long-read sequencing to identify pathogenic variants in autism cases, finding three pathogenic mutations and nine candidate structural variants missed by short-read sequencing.

This paper introduces pangenome-aware DeepVariant, a variant caller that improves accuracy by using a pangenome reference, outperforming linear-reference-based DeepVariant by up to 25.5%.

The paper introduces a personalized pangenome approach to improve short-read mapping and variant calling by reducing reference bias and errors.