首页 | 本学科首页   官方微博 | 高级检索  
   检索      


Jabba: hybrid error correction for long sequencing reads
Authors:Giles Miclotte  Mahdi Heydari  Piet Demeester  Stephane Rombauts  Yves Van de Peer  Jan Fostier
Institution:1.Department of Information Technology,Ghent University - iMinds,Ghent,Belgium;2.Department of Plant Systems Biology,VIB,Ghent,Belgium;3.Department of Biotechnology and Bioinformatics,Ghent University,Ghent,Belgium;4.Bioinformatics Institute Ghent,Ghent,Belgium;5.Department of Genetics, Genome Research Institute,University of Pretoria,Pretoria,South Africa
Abstract:

Background

Third generation sequencing platforms produce longer reads with higher error rates than second generation technologies. While the improved read length can provide useful information for downstream analysis, underlying algorithms are challenged by the high error rate. Error correction methods in which accurate short reads are used to correct noisy long reads appear to be attractive to generate high-quality long reads. Methods that align short reads to long reads do not optimally use the information contained in the second generation data, and suffer from large runtimes. Recently, a new hybrid error correcting method has been proposed, where the second generation data is first assembled into a de Bruijn graph, on which the long reads are then aligned.

Results

In this context we present Jabba, a hybrid method to correct long third generation reads by mapping them on a corrected de Bruijn graph that was constructed from second generation data. Unique to our method is the use of a pseudo alignment approach with a seed-and-extend methodology, using maximal exact matches (MEMs) as seeds. In addition to benchmark results, certain theoretical results concerning the possibilities and limitations of the use of MEMs in the context of third generation reads are presented.

Conclusion

Jabba produces highly reliable corrected reads: almost all corrected reads align to the reference, and these alignments have a very high identity. Many of the aligned reads are error-free. Additionally, Jabba corrects reads using a very low amount of CPU time. From this we conclude that pseudo alignment with MEMs is a fast and reliable method to map long highly erroneous sequences on a de Bruijn graph.
Keywords:
本文献已被 SpringerLink 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号