A Sequence-Based Novel Approach for Quality Evaluation of Third-Generation Sequencing Reads

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Third-generation sequencing (TGS) technologies offer long reads for advanced bioinformatics analyses.
• However, TGS reads have high error rates, challenging accurate genome assembly and alignment.
• Efficient methods are needed to select high-quality reads for improved downstream processing.

Purpose of the Study:

• To develop a novel tool, the Read Quality Evaluation and Selection Tool (REQUEST), for assessing third-generation long-read quality.
• To improve genome assembly and long-read alignment accuracy by prioritizing high-quality reads.

Main Methods:

• REQUEST characterizes reads by nucleotide combinations to generate training data for quality scoring.
• A linear regression model is employed to score the quality of individual reads.
• The method was validated on three diverse species datasets.

Main Results:

• Top-scored reads selected by REQUEST demonstrated higher alignment accuracies compared to all reads.
• Contig assembly using REQUEST-selected reads outperformed conventional methods that utilize all reads.
• REQUEST effectively distinguishes high-quality from low-quality reads without requiring reference genomes.

Conclusions:

• REQUEST provides a valuable, reference-free method for evaluating third-generation sequencing read quality.
• Prioritizing high-quality reads with REQUEST enhances the accuracy of genome assembly and alignment.
• This tool offers a promising alternative to alignment-based quality assessment methods.