Single-molecule sequence assembly and analysis

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

The primary focus of the section over the past year has been the completion of the human reference genome, which started with our founding of the Telomere-to-Telomere (T2T) consortium a few years ago. The goal of this project is to complete the last remaining regions of the human genome, which have gone unfinished for the 20 years following the completion of the Human Genome Project. With the recent introduction of new long-read sequencing methods, our work towards this goal has rapidly accelerated. Prior reports have noted our completion of the first human chromosome (ChrX) and autosome (Chr8). This year we published the first truly complete sequence of a human genome (ref 1). This landmark achievement was the result of a tremendous effort, led by the GIS, and including over 100 members of the T2T consortium. In addition, members of the GIS co-authored 5 companion papers that cataloged all segmental duplications in the genome (ref 2); mapped the human centromeres for the first time (ref 3); surveyed methylation and epigenetic state across the entire genome (ref 4); annotated all transposable and repetitive elements in the genome (ref 5); and evaluated new variants that can now be discovered using this new reference genome (ref 6). These publications were enabled by the sections prior development of the HiCanu software, as well as a close collaboration with the NIH Intramural Sequencing Center (NISC), which provided nanopore sequencing that was critical to the success of the project. Further supporting the T2T project, the GIS developed several new methods that were key to the success of the project. This included new assembly polishing methods, to ensure the highest degree of base and structural accuracy possible (ref 7). An additional k-mer-based technique was developed for filtering variant calls and prioritizing corrections (ref 8). To help analyze the most repetitive regions of the genome, a new long-read mapping algorithm was developed that is more accurate within the tandemly repeated regions making up a large fraction of the newly added reference sequence (ref 9). This has enabled variant calling across the entire genome. Lastly, we developed a novel heatmap-style visualization for analyzing satellite repeats, which has revealed new insights into their evolutionary mechanisms (ref 10). Going forward, the T2T consortium has now joined with the Human Pangenome Reference Consortium (HPRC) to build a number of additional complete human genome assemblies in the coming years from a broad set of samples that will fully capture the landscape of human genomic diversity. This project will require the complete and routine assembly of diploid human genomes. To enable this, the section has developed a new assembly pipeline named Verkko that is capable of automated assembly of complete diploid chromosomes using a combination of multiple sequencing technologies. Our strategy for building a new human pangenome reference was published in a perspective piece (ref 11). Our work also continues on non-human genomes, and the GIS remains an active member of the Vertebrate Genomes Project (VGP) and Earth BioGenome Project (EBP), which together aim to sequence the genomes of all eukaryotic life on earth. These projects are producing extremely valuable genomic datasets that will guide future conservation efforts and enable large-scale comparative genomics. Both projects have benefited from genome assembly, validation, and alignment software developed by the GIS in prior years. This past year, the GIS participated in the development of quality standards for the EBP (ref 12), and published a commentary piece on indigenous data sovereignty, as it relates to the EBPs biodiversity preservation efforts. The associated VGP now sits only a few tens of genomes short of its phase 1 goal of completing at least one genome from each of the approximately 270 vertebrate taxonomic orders. In addition to the 13 papers above that were formally published this year, the section has posted 7 preprints to bioRxiv that are currently undergoing peer review, including preliminary construction and analysis of the human pangenome, Verkko, and other related tools.