NPBSS

NPBSS generates high-fidelity continuous long reads (CLR) that mimic raw PacBio SMRT sequencing data for use in de novo genome assembly and benchmarking.

Key Features:

• Logarithmic Normal Distribution Sampling: Samples read lengths from a logarithmic normal distribution to reflect the natural variability of PacBio CLR read lengths.
• Base Quality Value Assignment: Assigns base quality values with varying proportions across reads to emulate base-level quality variation.
• Empirical Error Modeling: Employs an empirical model to compute an overall error probability for each base within a read.
• Error Probability Calculations: Computes deletion, substitution, and insertion probabilities based on the per-base overall error probability.

Scientific Applications:

• De Novo Genome Assembly: Provides realistic CLR datasets for testing and optimizing assembly algorithms aimed at producing reference-quality genomes.
• Tool Development and Benchmarking: Enables development, testing, and comparison of bioinformatics tools under simulated PacBio SMRT sequencing conditions.

Methodology:

NPBSS samples read lengths from a logarithmic normal distribution, assigns base quality values, employs an empirical model to compute per-base overall error probabilities, calculates deletion/substitution/insertion probabilities from those error probabilities, aligns its error-rate modeling to characteristic features of PacBio SMRT CLR data, and was evaluated against PBSIM and FASTQSim.