Enrichment of long DNA fragments from mixed samples for Nanopore sequencing

Whole-genome sequencing of pathogenic organisms directly from clinical samples combines detection and genotyping in one step. This can speed up diagnosis, especially for slow-growing organisms like Mycobacterium tuberculosis (Mtb), which need considerable time to grow in subculture, and can provide vital information for effective personalised treatment. Within the PATHSEEK project, we have developed a bait-capture approach to selectively enrich DNA/RNA from specific bacterial and viral pathogens present in clinical samples. Here, we present a variation of the method that allows enrichment of large fragments of target DNA for sequencing on an Oxford Nanopore MinIONTM sequencer. We enriched and sequenced cDNA from Influenza A (FluA), genomic DNA (gDNA) from human cytomegalovirus (CMV) and from two strains of Mtb, and present an evaluation of the method together with analysis of the sequencing results from a MinIONTM and an Illumina MiSeq sequencer. While unenriched FluA and CMV samples had no reads matching the target organism due to the high background of DNA from host cell lines, enriched samples had 56.7% and 90.9% on-target reads respectively for the best quality Nanopore reads.