CNVrd2

CNVrd2 assigns copy-number states from next-generation sequencing (NGS) read-depth data to detect and characterize copy number variation (CNV) at variable genomic loci.

Key Features:

• Read depth-based segmentation: Utilizes observed read-count ratios to segment genomic regions within a population and refine segmentation results.
• Cross-population linear regression: Applies linear regression to adjust segmentation results across multiple populations for consistent copy-number calls.
• Bayesian normal mixture clustering: Employs a Bayesian normal mixture model to cluster segmentation scores into discrete copy-number groups.
• Performance validation: Validated on 1000 Genomes Project data at CCL3L1 and DEFB103A, achieving concordance rates of 77.8% and 90.4% against paralog ratio test data and outperforming cn.mops (36.7%/4.8%) and CNVnator (7.2%/1%).
• Complex CN handling: Targets regions with complex and variable copy number to improve accuracy of CN assignments.

Scientific Applications:

• Population genomics: Supports studies of genetic diversity and population-level variation in copy number.
• Disease-associated CNV analysis: Enables investigation and characterization of CNVs associated with phenotypic differences or disease susceptibility.
• NGS-based CNV studies: Provides robust copy-number assignment for genomic studies relying on NGS read-depth data.

Methodology:

Initial segmentation using read-count ratios within a population; cross-population adjustment using linear regression; clustering of segmentation scores with a Bayesian normal mixture model.