Oxford Nanopore's long-read sequencing produces data that short-read platforms cannot: structural variants, complete plasmid assemblies, phased haplotypes, resolution of repetitive genomic regions. In antimicrobial resistance research, it is increasingly the only way to understand how resistance genes are organised on mobile elements. This matters because co-location on the same plasmid is precisely what allows multiple resistance traits to spread together.

The catch is sample preparation. Long-read sequencing requires DNA that is intact and high molecular weight. The shearing that occurs during conventional extraction (centrifugation steps, column binding, multiple transfers, vigorous pipetting) degrades fragment length and reduces the number of reads above the length thresholds that make long-read data informative. Extraction method matters more for long-read sequencing than for any other downstream application.

The Connelly et al. 2025 study of pathogenic E. coli from diseased pigs in England is a useful illustration. For standard Illumina short-read sequencing of the full 208-isolate dataset, the authors used an automated magnetic bead extraction system. For the subset of four isolates taken to Oxford Nanopore long-read sequencing, specifically the isolates where they needed to resolve complete plasmid structures, the PDQeX was the extraction method.

That is not a coincidence. Enzymatic extraction generates high-molecular-weight DNA because it lyses cells without the mechanical disruption and harsh chemistry that truncate fragments. For long-read workflows specifically, the extraction step is not interchangeable with whatever happens to be on the bench.

If your lab is moving toward long-read sequencing, the extraction method is worth reconsidering before the first run.