Investigation of chimeric reads using the MinION

Abstract

Following a nanopore sequencing run of PCR products of three amplicons less than 1kb, an abundance of reads failed quality control due to template/complement mismatch. A BLAST search demonstrated that some of the failed reads mapped to two different genes -- an unexpected observation, given that PCR was carried out separately for each amplicon. A further investigation was carried out specifically to search for chimeric reads, using separate barcodes for each amplicon and trying two different ligation methods prior to sample loading. Despite the separation of ligation products, chimeric reads formed from different amplicons were still observed in the base-called sequence. The long-read nature of nanopore sequencing presents an effective tool for the discovery and filtering of chimeric reads. We have found that at least 1.7% of reads prepared using the Nanopore LSK002 2D Ligation Kit include post-amplification chimeric elements. This finding has potential implications for other amplicon sequencing technologies, as the process is unlikely to be specific to the sample preparation used for nanopore sequencing.

Long-read only assembly of Drechmeria coniospora genomes reveals widespread chromosome plasticity and illustrates the limitations of current nanopore methods

“…They reflected a short (<50 bp) discrepancy between the individual reads and the final sequence. There was no indication of any sequence bias at the break points of the remaining chimeric reads supporting the notion that these reads arise from too rapid reloading of the sequencing pore [17].…”

Long-read only assembly of Drechmeria coniospora genomes reveals widespread chromosome plasticity and illustrates the limitations of current nanopore methods

“…They reflected a short (<50 bp) discrepancy between the individual reads and the final sequence. There was no indication of any sequence bias at the break points of the remaining chimeric reads, supporting the notion that these reads arise from excessively rapid reloading of the sequencing pore [ 17 ].…”

JAFFAL: Detecting fusion genes with long read transcriptome sequencing

“…We observed a similar difference in the number of low confidence calls from JAFFAL in the direct RNA (5 fusions reported) compared to the cDNA protocol (94 fusions reported) (Table 1). We hypothesise this is due to chimeric molecule creation during cDNA library preparation [34,35]. The ranking of these fusions as low confidence is consistent with this hypothesis because a hallmark of these events are breakpoints occurring within exons, rather than at exon boundaries and allows them to be separated from true fusions by JAFFAL’s ranking.…”