Related Experiment Video
Updated: Jan 4, 2026
11:26
Sequencing of mRNA from Whole Blood using Nanopore Sequencing
Published on: June 3, 2019
14.6K
De novo Nanopore read quality improvement using deep learning.
Nathan LaPierre1, Rob Egan2, Wei Wang3
1Department of Computer Science, University of California, Los Angeles, Los Angeles, CA, 90095, USA.
BMC Bioinformatics
|November 8, 2019
Summary
MiniScrub, a novel Convolutional Neural Network (CNN) method, effectively identifies and removes low-quality segments from Oxford Nanopore sequencing reads. This improves de novo genome assembly accuracy by reducing mis-assemblies and indel errors, particularly in metagenomic datasets.
Area of Science:
- Genomics
- Bioinformatics
- Computational Biology
Background:
- Long-read sequencing technologies like Oxford Nanopore simplify genome assembly and structural variation identification.
- Oxford Nanopore reads exhibit high error rates, often clustered in low-quality segments.
- Existing error correction methods struggle with sensitivity, leading to mis-assemblies.
Purpose of the Study:
- To develop a robust method for identifying and removing low-quality segments from Nanopore reads.
- To minimize the impact of read errors on downstream genome assembly processes.
- To enhance the accuracy of de novo genome assembly and structural variation analysis.
Main Methods:
- Developed MiniScrub, a Convolutional Neural Network (CNN) based approach.
- Generated read-to-read overlaps using MiniMap2.
- Encoded overlaps into images for CNN model training to predict low-quality segments.
Main Results:
- MiniScrub effectively identifies and removes low-quality Nanopore read segments.
- Demonstrated robust improvement in read quality and error correction, especially for metagenomic data.
- De novo genome assembly using scrubbed reads resulted in significantly fewer mis-assemblies and large indel errors.
Conclusions:
- MiniScrub robustly enhances Oxford Nanopore read quality, benefiting downstream applications like de novo assembly.
- The tool is particularly effective in metagenomic settings.
- MiniScrub is proposed as a preprocessing tool for Nanopore reads and is available as open-source software.
