Related Experiment Video
Updated: Jan 17, 2026
05:45
Validating Whole Genome Nanopore Sequencing, using Usutu Virus as an Example
Published on: March 11, 2020
9.3K
A comprehensive benchmarking of adaptive sampling tools for nanopore sequencing.
Lang Yang1, Yanfeng Lin2, Peihan Li1
1Chinese PLA Center for Disease Control and Prevention, Beijing, China.
Genome Biology
|September 17, 2025
Summary
This study benchmarks nanopore adaptive sampling tools, finding nucleotide-alignment strategies like MinKNOW offer broad utility. Deep learning methods show high accuracy for efficient target enrichment and host DNA depletion.
Area of Science:
- Genomics
- Bioinformatics
Background:
- Adaptive sampling is an emerging technology for real-time nanopore sequencing.
- It enriches target DNA reads while depleting unwanted sequences.
- Various tools exist for read rejection, but optimal performance evaluation is lacking.
Purpose of the Study:
- To evaluate six popular nanopore adaptive sampling tools.
- To identify optimal strategies for specific tasks like gene enrichment and host DNA depletion.
- To compare nucleotide-alignment-based and deep learning approaches.
Main Methods:
- Tested tools on intraspecies gene enrichment, interspecies enrichment, and host DNA depletion.
- Evaluated performance using Guppy base calling and minimap2 read alignment.
- Compared conventional signal-based methods with deep learning on raw signals.
Main Results:
- All tested tools increased target coverage depth (1.50- to 4.86-fold).
- Guppy and minimap2 combination proved optimal for read classification accuracy.
- MinKNOW, Readfish, and BOSS-RUNS showed excellent performance with this strategy.
- Deep learning methods achieved high accuracy and speed in host depletion.
Conclusions:
- Nucleotide-alignment-based approaches are versatile for various targets.
- MinKNOW is a recommended tool for most adaptive sampling scenarios.
- Deep learning on raw signals offers superior classification efficiency and accuracy, meriting further development.
