A complete diploid human genome benchmark for personalized genomics.

A telomere-to-telomere human genome benchmark achieves 99.4% accuracy, adding 701.4 Mb of previously missing autosomal and sex chromosome sequences.

• Why it matters: Current genome sequencing methods rely on reference mapping, which introduces biases and omits complex regions, limiting the precision of personalized medicine.
• What they did: The team created a comprehensive diploid genome benchmark using new methods to measure accuracy, annotating nearly 20,000 genes per haplotype and including challenging regions.
• The result: De novo assembly surpasses traditional variant calling by resolving 2–7% more sequence and greatly improves accuracy, broadening the potential for fully personalized genomic analysis.