Methylation-induced errors in Nanopore Sequencing data, showcased by an Outbreak Analysis

Our study investigated the accuracy of Oxford Nanopore Technology by resequencing a three-year-long Klebsiella pneumoniae outbreak and comparing their genomes and relationship in an outbreak context using additional Illumina sequencing data as a point of reference.

We performed whole genome sequencing using the latest chemicals (SQK-NBD114.24) from Oxford Nanopore Technologies with improved accuracy over the previous ones. The long-read data was characterized on a sequence, nucleotide, and raw signal level while evaluating alternative kits and basecalling approaches. Compared to Illumina data, base errors detected in cgMLST (see Figure) and phylogenetic analysis of Nanopore-sequenced genomes led to false exclusions of closely related strains from an outbreak. Nearby methylation sites affect the raw signal, resulting in false bases in the genomes. These errors are not limited to K. pneumoniae and were identified among six different bacterial orders (Actinomycetales, Xanthomonadales, Burkholderiales, Flavobacteriales, Pseudomonadales and Enterobacterales). Considering these findings, we explored PCR-based sequencing (SQK-RPB114.24) and a masking strategy, which successfully addressed these inaccuracies. Additionally, we provide a bioinformatic workflow to identify and mask problematic genome positions in a reference-free manner (MPOA is freely available on GitHub under the GNUv3 license: github.com/replikation/MPOA).

Our research identified limitations in using Oxford Nanopore Technologies for sequencing prokaryotic organisms when high accuracy is mandatory, particularly in outbreak investigation. Without further technological developments, our study recommends either PCR-based sequencing or using our provided bioinformatic workflow to mask problematic genome positions.