RegScaf: a regression approach to scaffolding

doi:10.1093/bioinformatics/btac174

	RegScaf: a regression approach to scaffolding
	Li, Mengtian 1,2; Li, Lei M.1,2
刊名	BIOINFORMATICS
	2022-03-25
页码	8
ISSN号	1367-4803
DOI	10.1093/bioinformatics/btac174
英文摘要	Motivation: Crucial to the correctness of a genome assembly is the accuracy of the underlying scaffolds that specify the orders and orientations of contigs together with the gap distances between contigs. The current methods construct scaffolds based on the alignments of 'linking' reads against contigs. We found that some 'optimal' alignments are mistaken due to factors such as the contig boundary effect, particularly in the presence of repeats. Occasionally, the incorrect alignments can even overwhelm the correct ones. The detection of the incorrect linking information is challenging in any existing methods. Results: In this study, we present a novel scaffolding method RegScaf. It first examines the distribution of distances between contigs from read alignment by the kernel density. When multiple modes are shown in a density, orientation-supported links are grouped into clusters, each of which defines a linking distance corresponding to a mode. The linear model parameterizes contigs by their positions on the genome; then each linking distance between a pair of contigs is taken as an observation on the difference of their positions. The parameters are estimated by minimizing a global loss function, which is a version of trimmed sum of squares. The least trimmed squares estimate has such a high breakdown value that it can automatically remove the mistaken linking distances. The results on both synthetic and real datasets demonstrate that RegScaf outperforms some popular scaffolders, especially in the accuracy of gap estimates by substantially reducing extremely abnormal errors. Its strength in resolving repeat regions is exemplified by a real case. Its adaptability to large genomes and TGS long reads is validated as well.
资助项目	National Natural Science Foundation of China[32170679] ; National Natural Science Foundation of China[11871462] ; National Natural Science Foundation of China[91530105] ; National Center for Mathematics and Interdisciplinary Sciences of the CAS ; Key Laboratory of Systems and Control of the CAS ; Strategic Priority Research Program of the Chinese Academy of Sciences[XDB13040600]
WOS研究方向	Biochemistry & Molecular Biology ; Biotechnology & Applied Microbiology ; Computer Science ; Mathematical & Computational Biology ; Mathematics
语种	英语
出版者	OXFORD UNIV PRESS
WOS记录号	WOS:000784528800001
内容类型	期刊论文
源URL	[http://ir.amss.ac.cn/handle/2S8OKBNM/60352]
专题	中国科学院数学与系统科学研究院
通讯作者	Li, Lei M.
作者单位	1.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Acad Math & Syst Sci, Natl Ctr Math & Interdisciplinary Sci, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Li, Mengtian,Li, Lei M.. RegScaf: a regression approach to scaffolding[J]. BIOINFORMATICS,2022:8.
APA	Li, Mengtian,&Li, Lei M..(2022).RegScaf: a regression approach to scaffolding.BIOINFORMATICS,8.
MLA	Li, Mengtian,et al."RegScaf: a regression approach to scaffolding".BIOINFORMATICS (2022):8.