Authors: Karen H. Y. Wong, Walfred Ma, Chun-Yu Wei, Erh-Chan Yeh, Wan-Jia Lin, Elin H. F. Wang, Jen-Ping Su, Feng-Jen Hsieh, Hsiao-Jung Kao, Hsiao-Huei Chen, Stephen K. Chow, Eleanor Young, Catherine Chu, Annie Poon, Chi-Fan Yang, Dar-Shong Lin, Yu-Feng Hu, Jer-Yuarn Wu, Ni-Chung Lee, Wuh-Liang Hwu, Dario Boffelli, David Martin, Ming Xiao, Pui-Yan Kwok
Published: 2020-10-30
DOI: 10.1038/s41467-020-19311-w
Source: Full article
AbstractThe current human reference genome is predominantly derived from a single individual and it does not adequately reflect human genetic diversity. Here, we analyze 338 high-quality human assemblies of genetically divergent human populations to identify missing sequences in the human reference genome with breakpoint resolution. We identify 127,727 recurrent non-reference unique insertions spanning 18,048,877 bp, some of which disrupt exons and known regulatory elements. To improve genome annotations, we linearly integrate these sequences into the chromosomal assemblies and construct a Human Diversity Reference. Leveraging this reference, an average of 402,573 previously unmapped reads can be recovered for a given genome sequenced to ~40X coverage. Transcriptomic diversity among these non-reference sequences can also be directly assessed. We successfully map tens of thousands of previously discarded RNA-Seq reads to this reference and identify transcription evidence in 4781 gene loci, underlining the importance of these non-reference sequences in functional genomics. Our extensive datasets are important advances toward a comprehensive reference representation of global human genetic diversity.