De novo human genome assemblies across populations

The human genome reference is an integral part of modern biological research. However, a single consensus genome representation is inadequate to provide a universal reference structure due to its lack of sequence variation and structural diversity. Using 10x Genomics “linked-read” technology, we performed whole genome sequencing and de novo assembly on 18 individuals across five different populations. Here, we identified 1,872 breakpoint-resolved Novel Sequence Insertions (NSIs), and accumulatively they added up to 2.1Mb of new genomic content not currently represented in the human genome reference. These NSIs adhere to an underlying population structure, as principal component analysis can faithfully stratify these individuals based on their genetic backgrounds. Furthermore, over half of these NSIs aligned to nonhuman primates, suggesting that they are ancestral to humans. While investigating their origins, we discovered that a subset of NSIs were results of Alu-mediated recombination. Additionally, some of the NSIs aligned to either entries in the expressed sequence tag database or reads from RNA-sequencing experiments, suggesting that they bear sequences that are transcribed. Our results demonstrate that as the genomic variant catalogs continue to expand, a more comprehensive set of alternative haplotypes are necessary to depict the complete spectrum of genetic diversity across populations.