Error analysis of the PacBio sequencing CCS reads

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Third-generation sequencing (TGS) technologies like PacBio and Oxford Nanopore offer longer reads than next-generation sequencing (NGS).
• However, TGS long reads exhibit higher error rates, necessitating error correction, such as Circular Consensus Sequencing (CCS), before assembly.
• Accurate error modeling is crucial for reliable TGS data analysis and genome assembly.

Purpose of the Study:

• To develop a probabilistic model for error occurrence in PacBio CCS reads.
• To provide nucleotide error probabilities and Phred quality scores based on the number of sub-reads.
• To characterize the error rate distribution of CCS reads concerning the pass number.

Main Methods:

• Proposed a probabilistic model for error occurrence in CCS reads.
• Derived nucleotide error probabilities and Phred quality scores as a function of sub-reads.
• Modeled the error rate distribution using binomial distribution, approximated by normal distribution for long reads.

Main Results:

• The model accurately estimates error probabilities and Phred quality scores for CCS reads.
• The error rate distribution follows a binomial distribution, which can be approximated by a normal distribution.
• Validation with Lambda, E. coli, and Alzheimer's disease datasets confirmed the model's efficacy.

Conclusions:

• The proposed probabilistic model effectively characterizes errors in PacBio CCS reads.
• This model enhances the understanding and correction of errors in long-read sequencing data.
• The findings contribute to more accurate genome assembly and analysis using TGS technologies.