Researchers have created a nearly complete human genome reference that could help laboratories assess sequencing accuracy in regions associated with disease but often excluded from current validation tools.
Laboratories commonly evaluate sequencing workflows by comparing detected variants with established "Genome in a Bottle" reference data. However, existing resources do not cover the entire genome. Repetitive, duplicated, and structurally variable regions are particularly difficult to analyze, which can limit the assessment of clinically relevant genes within these regions.
The researchers assembled the genome of HG002, a widely used reference sample, from one end of each chromosome to the other. Unlike standard references that represent a single version of each chromosome, this diploid assembly contains the maternal and paternal copies of all 46 chromosomes. This structure more closely reflects the two sets of chromosomes present in a patient’s cells.
The completed resource had no detectable errors across approximately 99 percent of the genome. The remaining gaps were limited mainly to ribosomal DNA arrays, which contain highly repetitive sequences that remain difficult to reconstruct.
Compared with an earlier small-variant benchmark, the new resource added approximately 701 million bases of autosomal sequence and 217 million bases from the X and Y chromosomes. The previous version excluded both sex chromosomes and approximately 12 percent of the autosomes.
Multiple sequencing technologies, assembly methods, manual review, and cytogenetic testing were used to construct and validate the genome. Genes and repetitive sequences were annotated separately on the maternal and paternal chromosome sets, allowing differences between the two inherited copies to be examined.
The researchers also developed Genome Quality Checker software to assess sequencing reads, phased variant results, and complete genome assemblies. When existing data were evaluated, assembling a genome directly resolved 2 to 7 percent more sequence than conventional variant-calling approaches and produced approximately one-tenth as many errors.
The resource may support more complete validation of sequencing pipelines. It includes segmental duplications, tandem repeats, copy-number changes, and medically relevant genes that can be difficult to assess using standard reference-based methods. The accompanying software can identify coverage gaps, sequencing errors, incorrect assignment of variants to parental chromosomes, and other analysis limitations.
Patient diagnoses or demonstrated improved clinical outcomes were not assessed in this study. HG002 represents a single reference sample, and low-level genetic differences may occur between batches of its cultured cells. Results can also vary according to the alignment method used, while measures for evaluating complete genome assemblies are not yet standardized for clinical laboratories.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
