Srfim

Srfim performs model-based quality assessment and base-calling to quantify uncertainties and improve base-calling accuracy in Illumina Genome Analyzer (GA) second-generation sequencing data.

Key Features:

• Model-Based Quality Assessment: Srfim employs a statistical model to quantify and interpret uncertainties in the base-calling process for Illumina GA data.
• Enhanced Base-Calling Performance: It addresses systematic errors introduced by discretization of noisy continuous fluorescence intensity measurements to improve base-calling accuracy.
• Interpretability of Model Parameters: Model parameters have direct interpretations related to the chemistry of the base-calling process.
• Quality Estimates for Downstream Analysis: The method provides read-level and sample-level sequencing quality metrics that support downstream analyses, including rare variant detection.

Scientific Applications:

• Large-scale population genomics (1000 Genomes Project): Applicable to projects such as the 1000 Genomes Project (≈1200 human genomes) to enable comparative single-nucleotide analysis across diverse populations.
• Rare variant detection: Improves reliability of detecting rare variants by providing more accurate base calls and sequencing quality estimates.

Methodology:

Srfim fits a simple statistical model that captures uncertainties arising during base-calling on the Illumina/Solexa Genome Analyzer, accounts for variations in read quality within and across samples, and models the discretization of noisy continuous fluorescence intensities, with model parameters interpretable in terms of base-calling chemistry.