May 13, 2024

Optimized ONT Library Prep Protocol for Low-Input mNGS

Oxford Nanopore Technologies (ONT) platforms offer real-time sequencing, long reads, and low capital costs that make them attractive for metagenomic next-generation sequencing (mNGS). But standard nanopore library prep protocols assume input DNA quantities that host-depleted clinical samples rarely provide. In a micro-webinar, Dr. Mengchu Wu presented an optimized ONT library preparation protocol that combines host depletion with PCR-based amplification to generate sufficient library DNA from ultra-low biomass specimens, turning the affordable Flongle flow cell into a practical mNGS tool.

Key Takeaways

The ONT Flongle flow cell theoretically sequences approximately 2.8 picograms of double-stranded DNA (dsDNA), but the Rapid Barcoding Kit requires 200 nanograms input — a 70,000-fold gap
Host-depleted samples often yield DNA below the Qubit detection limit, making direct library construction impossible with standard ONT kits
The Rapid PCR Barcoding Kit with controlled PCR cycles generates 4x more total reads than the standard Rapid Barcoding Kit from non-depleted samples
Host-depleted samples processed with the PCR protocol produced dramatically higher microbial read percentages compared to non-depleted samples
Within 3 hours of sequencing, the PCR protocol surpasses the total read output of the non-PCR protocol’s complete 24-hour run

The Input DNA Problem on Nanopore Platforms

ONT sequencing is fundamentally different from sequencing-by-synthesis platforms. It detects electrical signals as native DNA molecules pass through nanopores, requiring no amplification and enabling direct detection of base modifications. However, these advantages come with a practical constraint: the physical-chemical kinetics of nanopore sequencing require DNA concentrations far exceeding what gets sequenced.

The Flongle flow cell offers a theoretical output of 2.6 Gb, translating to approximately 650,000 reads at a mean fragment length of 4,000 base pairs. The total mass of DNA that actually gets sequenced is roughly 2.8 picograms. Yet the ONT Rapid Barcoding Kit specifies a minimum input of 200 nanograms, and even the Rapid PCR Barcoding Kit requires 1 nanogram.

When insufficient library DNA is loaded, the sequencing duty plot degrades: fewer pores are actively sequencing, and the output drops well below the flow cell’s capacity. For host-depleted clinical samples — where DNA concentration frequently falls below the Qubit dsDNA assay detection limit — this input gap makes standard library preparation impractical. Overcoming library prep failure in low-biomass and host-depleted samples requires a fundamentally different approach.

Rapid Barcoding vs. Rapid PCR Barcoding Kit

The webinar compared two ONT library preparation approaches using simulated clinical samples (human whole blood spiked with Allobacillus halotolerans and Imtechella halotolerans), processed with and without host depletion.

Non-depleted samples yielded sufficient DNA for the standard Rapid Barcoding Kit. Library construction proceeded according to the manufacturer’s protocol.

Host-depleted samples produced DNA concentrations too low to measure by Qubit. For these samples, the Rapid PCR Barcoding Kit was used, with PCR cycle numbers optimized to generate enough library while minimizing amplification bias. Two different cycle numbers were tested; the goal was to achieve sufficient amplified library with the fewest cycles possible.

Each protocol was run in technical duplicate across two different Flongle flow cell batches, providing four independent sequencing runs for comparison.

Sequencing Efficiency and Read Output Compared

The duty plot comparison revealed a counterintuitive result. Despite loading sufficient library DNA per protocol specifications, the Rapid Barcoding Kit (non-depleted samples) showed suboptimal sequencing efficiency, with fewer pores actively sequencing. The Rapid PCR Barcoding Kit (host-depleted samples) achieved a visibly better duty plot, with more pores engaged in active sequencing.

The quantitative results were striking:

Rapid PCR Barcoding Kit (host-depleted): approximately 270,000 total reads
Rapid Barcoding Kit (non-depleted): approximately 70,000 total reads

The non-PCR protocol produced longer median read lengths, consistent with the absence of amplification. The PCR protocol generated reads with a median length of approximately 4,000 base pairs, reflecting the PCR product size distribution. Sequencing quality scores were comparable between the two approaches.

Genome coverage analysis further highlighted the advantage: the PCR protocol from host-depleted samples generated deep, uniform coverage across both spiked bacterial genomes. The non-PCR protocol from non-depleted samples showed coverage gaps where no reads mapped, insufficient for confident pathogen identification or genome assembly.

Host Depletion Transforms the Read Composition

The composition analysis revealed the combined impact of host depletion and PCR amplification. In non-depleted samples, human reads dominated the output, leaving minimal signal for the spiked bacteria. In host-depleted samples processed with the PCR protocol, human reads decreased dramatically while A. halotolerans and I. halotolerans reads increased by orders of magnitude.

The total microbial read counts from the host-depleted PCR protocol dwarfed those from the non-depleted standard protocol. The difference was large enough to enable genome assembly and antimicrobial resistance gene prediction — analyses that are ongoing.

Total Turnaround Time From Sample to Sequencing

The PCR amplification step adds approximately 3 hours (200 minutes) to library construction compared to the direct Rapid Barcoding Kit protocol. The complete timeline for the host-depleted PCR workflow:

Devin Host Depletion Filter processing and DNA extraction: approximately 2 hours
Rapid PCR Barcoding Kit library preparation: approximately 3.5 hours
Total sample-to-sequencing: approximately 6 hours within a working day

Despite the later start, the PCR protocol’s higher sequencing efficiency means it generates more absolute QC-passed reads than the non-PCR protocol within 3 hours of sequencing. By 6 hours, both protocols reach approximately 50% of their total run output; by 12 hours, approximately 80%. The final 12 hours of a 24-hour run contribute only the remaining 20% of reads.

This sequencing kinetics data suggests that laboratories may not need to run the full 24 hours. For applications where 50-80% of maximum output is sufficient, same-day results from sample to preliminary data are achievable.

How the Devin Filter Enables ONT Metagenomic Sequencing

The Devin Microbial Enrichment Kit paired with the ONT Rapid PCR Barcoding Kit creates a practical workflow for nanopore-based mNGS. Host depletion solves the read composition problem; PCR amplification solves the input DNA problem. Together, they transform the low-cost Flongle flow cell from a platform limited by host DNA interference into one capable of generating sufficient microbial reads for pathogen identification.

Micronbrane Medical’s LR PaRTI-Seq Analysis, recently launched for long-read sequencing data, uses minimap2 alignment against a curated database of over 1,400 pathogens. The tool is free for researchers with an access code and complements any existing analysis pipeline.

About the Speaker

Dr. Mengchu Wu is the co-founder, CEO, and Chairwoman of Micronbrane Medical. She previously held executive positions at HGT and Vela Diagnostics and earned her PhD in biomedical sciences at the University of Massachusetts Chan Medical School. Since co-discovering a novel host depletion method, Dr. Wu has been focused on advancing metagenomics.