PaSS

PaSS simulates sequencing reads from Pacific Biosciences (PacBio) third-generation platforms by reproducing read-length distributions and context-specific error patterns to support development and evaluation of bioinformatics methods.

Key Features:

• Context-specific error model: Employs a context-specific sequencing error model that reproduces the distribution of error rates observed in PacBio datasets.
• Read-length distribution modeling: Models PacBio-specific long-read characteristics and read-length distributions distinct from second-generation (NGS) technologies.
• Data-driven learning: Learns sequence patterns directly from existing PacBio sequencing data to capture platform-specific characteristics.
• Realistic simulated reads: Generates simulated reads that closely mimic experimental PacBio sequencing data for evaluation purposes.
• Comparative performance: Demonstrated performance improvements over PBSIM, LongISLND, and NPBSS across various evaluation metrics.

Scientific Applications:

• Algorithm benchmarking: Benchmarking and validation of algorithms and software tools for PacBio data analysis.
• Genome assembly testing: Pre-testing and evaluation of genome assembly pipelines using reads that resemble experimental sequencing.
• Method development: Development and evaluation of bioinformatics methods tailored for third-generation (PacBio) sequencing data.

Methodology:

PaSS applies a context-specific sequencing error model and a learning procedure that derives sequence patterns from existing PacBio datasets to reproduce observed read-length and error-rate distributions and generate simulated reads.