Automated assembly of high-quality diploid human reference genomes

Abstract

The current human reference genome, GRCh38, represents over 20 years of effort to generate a high-quality assembly, which has greatly benefited society^{1, 2}. However, it still has many gaps and errors, and does not represent a biological human genome since it is a blend of multiple individuals^{3, 4}. Recently, a high-quality telomere-to-telomere reference genome, CHM13, was generated with the latest long-read technologies, but it was derived from a hydatidiform mole cell line with a duplicate genome, and is thus nearly homozygous⁵. To address these limitations, the Human Pangenome Reference Consortium (HPRC) recently formed with the goal of creating a collection of high-quality, cost-effective, diploid genome assemblies for a pangenome reference that represents human genetic diversity⁶. Here, in our first scientific report, we determined which combination of current genome sequencing and automated assembly approaches yields the most complete, accurate, and cost-effective diploid genome assemblies with minimal manual curation. Approaches that used highly accurate long reads and parent-child data to sort haplotypes during assembly outperformed those that did not. Developing a combination of all the top performing methods, we generated our first high- quality diploid reference assembly, containing only ∼4 gaps (range 0-12) per chromosome, most within + 1% of CHM13’s length. Nearly 1/4th of protein coding genes have synonymous amino acid changes between haplotypes, and centromeric regions showed the highest density of variation. Our findings serve as a foundation for assembling near-complete diploid human genomes at the scale required for constructing a human pangenome reference that captures all genetic variation from single nucleotides to large structural rearrangements.