Characterization, correction and de novo assembly of an Oxford Nanopore genomic dataset from Agrobacterium tumefaciens.

Stéphane Deschamps,Thiruvarangan Ramaraj,Ajith Anand,Kevin Fengler,Connor Cameron,Todd J Jones,Gregory May,Victor Llaca,Joann Mudge,Kevin Hayes

doi:10.1038/srep28625

Stéphane Deschamps, Thiruvarangan Ramaraj + Show 8 more

Open Access

https://doi.org/10.1038/srep28625

Copy DOI

Abstract

The MinION is a portable single-molecule DNA sequencing instrument that was released by Oxford Nanopore Technologies in 2014, producing long sequencing reads by measuring changes in ionic flow when single-stranded DNA molecules translocate through the pores. While MinION long reads have an error rate substantially higher than the ones produced by short-read sequencing technologies, they can generate de novo assemblies of microbial genomes, after an initial correction step that includes alignment of Illumina sequencing data or detection of overlaps between Oxford Nanopore reads to improve accuracy. In this study, MinION reads were generated from the multi-chromosome genome of Agrobacterium tumefaciens strain LBA4404. Errors in the consensus two-directional (sense and antisense) “2D” sequences were first characterized by way of comparison with an internal reference assembly. Both Illumina-based correction and self-correction were performed and the resulting corrected reads assembled into high-quality hybrid and non-hybrid assemblies. Corrected read datasets and assemblies were subsequently compared. The results shown here indicate that both hybrid and non-hybrid methods can be used to assemble Oxford Nanopore reads into informative multi-chromosome assemblies, each with slightly different outcomes in terms of contiguity and accuracy.

Highlights

Second-generation DNA sequencing technologies have revolutionized the field of genomics[1,2] and given researchers access to unprecedented amounts of sequencing data
While early reports[16] indicated a high error rate for MinION reads (~35%), their accuracy has improved over time as Oxford Nanopore has released new versions of its base calling software and nanopore chemistry
The sequencing accuracy of individual MinION 2D reads generated from the strain LBA4404 of the plant pathogen Agrobacterium tumefaciens was evaluated, following the alignment of the 2D reads to an internal Illumina and Pacific Biosciences (PacBio)-based reference assembly

Summary

Introduction

Second-generation DNA sequencing technologies have revolutionized the field of genomics[1,2] and given researchers access to unprecedented amounts of sequencing data. Loman et al.[20] successfully demonstrated that Oxford Nanopore data alone (~30X genome coverage) could be used, after read correction using multiple alignments between overlapping reads to generate a consensus sequence, to create a single-contig de novo assembly of the E. coli genome, with a percentage identity of 98.4% compared to the finished reference assembly. This assembly was performed using a series of software tools developed for extracting FASTA sequences from raw signals (“poretools”), correcting overlapping reads (“nanocorrect”) and polishing the assembly (“nanopolish”). The results shown here provide more information on the Oxford Nanopore sequencing technology and confirm that PBcR and canu can effectively correct MinION 2D reads and assemble them into a multi-chromosomal microbial genome, with a small number of contigs and very high sequence accuracy

Methods

Results

Conclusion