A complete diploid human genome benchmark for personalized genomics.

Telomere-to-telomere benchmark of the diploid HG002 genome, annotated on both haplotypes, allowing accuracy of reads, phased variant call sets and assemblies to be measured against a diploid reference rather than a single reference genome.

Human genome sequencing usually relies on mapping reads to a reference genome, which introduces technical biases and excludes duplicated and structurally polymorphic regions. The authors present a telomere-to-telomere benchmark with near-perfect accuracy across 99.4% of the diploid HG002 genome. It adds 701.4 Mb of autosomal sequence and both sex chromosomes (216.8 Mb), absent from prior benchmarks, with genes and repeats annotated on both haplotypes, including 19,956 protein-coding genes on the maternal haplotype and 19,190 on the paternal haplotype. New methods were developed to measure the accuracy of reads, phased variant call sets and assemblies against a diploid reference. Genome-wide, de novo assembly resolves 2% to 7% more sequence and outperforms variant calling accuracy by an order of magnitude.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

In practical terms, this is the validation reference that was missing: a laboratory can now measure its pipeline's performance in regions so far excluded from any sensitivity figure, notably segmental duplications and the sex chromosomes. The caveat is the single sample, HG002: a benchmark is not a cohort, and it says nothing about performance on other genetic backgrounds or on DNA of variable clinical quality. The order-of-magnitude advantage of de novo assembly is the real signal, but moving from a reference-aligned pipeline to an assembly-based one means rebuilding the whole analytical chain, something no laboratory will undertake on the strength of a benchmark alone.

Analysis by Dr Thibaut Benquey