CNVPanelizer

CNVPanelizer performs gene-level detection and classification of copy number variations from targeted sequencing data using collections of non-matched normal tissue samples.

Key Features:

• Non-Parametric Bootstrap Subsampling: Employs non-parametric bootstrap subsampling of reference samples to estimate the distribution of read counts from targeted sequencing data.
• Amplicon Subsampling (inspired by Random Forests): Subsamples amplicons associated with each targeted gene using a procedure inspired by random forest algorithms to improve classification stability.
• Support for Non-Matched Normal Reference Samples: Utilizes collections of non-matched normal tissue samples as reference for CNV inference rather than requiring matched normals.
• Gene-Level Classification: Integrates the above statistical procedures to classify copy number aberrations at the gene level.

Scientific Applications:

• Cancer Genomics: Detection and classification of gene-level CNVs in cancer sequencing studies using targeted amplicon panels.
• Targeted Sequencing Studies with Unmatched References: Analyses of CNVs in experiments that rely on targeted sequencing and collections of non-matched normal samples.

Methodology:

Performs non-parametric bootstrap subsampling of reference samples to estimate read-count distributions and applies an amplicon subsampling procedure inspired by random forest methods to refine gene-level CNV classification.