De novo assembly of the NCTC 11168 Campylobacter jejuni subsp.jejuni genome A Comparison of sequence data generated from a single strain using Oxford Nanopore MinION, PacBio, Illumina and Sanger Sequencing technologies.

Systematic comparisons of sequencing data quality between different sequencing technologies usually focus on a limited number of those currently available and there is a need for a side by side comparison of their relative advantages when sequencing an identical target. The quality of data generated by MinION, Illumina and PacBio sequencing technologies was compared with the reference strain sequenced using Sanger sequencing technology. In addition an evaluation was carried out to determine whether multiplex sequenced data quality are similar within a single MinION flowcell and the possibility that high MinION read depth improves sequence accuracy was investigated. A single DNA preparation was extracted from a Campylobacter jejuni subsp.jejuni (NCTC 11168) isolate and sequenced using MinION and Illumina HiSeq 2500 sequencing technologies. DNA was sequenced in triplex on a single MinION flow cell. The read sets were trimmed and quality assessed. The MinION read sets were assembled using Canu and Miniasm and Illumina reads were assembled using SPAdes. Hybrid assembly (Illumina + MinION) was carried out using SPAdes and Unicycler. The resulting hybrid assemblies were then refined with Illumina reads using Pilon, whereas MinION contigs generated by Canu were polished using Nanopolish. Using the Sanger sequence as the gold standard the analysis showed that for MinION-only assembly tools Canu + Nanopolish performed the best and for hybrid (Illumina/MinION) assembly tools SPAdes + Pilon performed the best consistently. PacBio and hybrid (SPAdes + Pilon) assemblies showed the highest sequence quality (low number of mismatches and indels and generated one complete contig). In comparison to annotated Sanger sequence, PacBio and hybrid (SPAdes + Pilon) assemblies identified 98.92% and 99.11% of genes with 100% identity, respectively. Whereas, MinION-only (Canu + Nanopolish) and Illumina-only (SPAdes) identified 58.99% and 98.28% of genes, respectively. The quality of three MinION replicates that were sequenced on a single MinION flowcell using the Oxford Nanopore Technologies (ONT) barcoding kit was sufficiently stable that consistent results were obtained for all replicates. Reducing the MinION data depth of coverage from 2093 (read lengths N50:7652) to 231 (read lengths N50:7683) did not reduce hybrid sequence accuracy but reduced the MinION-only sequence accuracy. In conclusion, the quality of data generated from the hybrid assembly approaches that of PacBio data. MinION technology can now sequence 12 Campylobacter jejuni samples on the same flowcell and generate an accurate complete genome using a hybrid assembly. This is potentially beneficial to generate high quality complete genomes de novo in a short turnaround time at low cost.