Scaffolding of long read assemblies using long range contact information

BMC Genomics

Methodology Article

Derek Bickhart¹  Jay Ghurye²  Mihai Pop²  Sergey Koren³  Chen-Shan Chin⁴ 

[1] Cell Wall Biology and Utilization Research, US Dairy Forage Research Center, 53706 Madison, Wisconsin, USA;Department of Computer Science, University of Maryland, 20742 College Park, Maryland, USA;Genome Informatics Section, Computational and Statistical Genomics Branch, National Human Genome Research Institute, National Institutes of Health, 21702 Bethesda, Maryland, USA;Pacific Biosciences, 94205 Menlo Park, California, USA;

关键词: Assembly;    Scaffolding;    Hi-C;    Long reads;   

DOI  :  10.1186/s12864-017-3879-z

 received in 2017-01-28, accepted in 2017-06-20,  发布年份 2017

来源: Springer

【 摘 要 】

BackgroundLong read technologies have revolutionized de novo genome assembly by generating contigs orders of magnitude longer than that of short read assemblies. Although assembly contiguity has increased, it usually does not reconstruct a full chromosome or an arm of the chromosome, resulting in an unfinished chromosome level assembly. To increase the contiguity of the assembly to the chromosome level, different strategies are used which exploit long range contact information between chromosomes in the genome.MethodsWe develop a scalable and computationally efficient scaffolding method that can boost the assembly contiguity to a large extent using genome-wide chromatin interaction data such as Hi-C.Resultswe demonstrate an algorithm that uses Hi-C data for longer-range scaffolding of de novo long read genome assemblies. We tested our methods on the human and goat genome assemblies. We compare our scaffolds with the scaffolds generated by LACHESIS based on various metrics.ConclusionOur new algorithm SALSA produces more accurate scaffolds compared to the existing state of the art method LACHESIS.

【 授权许可】

CC BY   
© The Author(s) 2017