TrioCNV

TrioCNV detects copy number variations (CNVs) from whole genome sequencing (WGS) data in parent-offspring trios by jointly modeling read depth and Mendelian inheritance to improve CNV detection accuracy.

Key Features:

• Joint Modeling Approach: Jointly models CNV states across parent-offspring trios using familial relationships and Mendelian inheritance.
• Negative Binomial Regression: Models read depth signals with negative binomial regression to accommodate over-dispersion in sequencing data.
• Bias Correction: Adjusts read depth for GC content and mappability biases in WGS data.
• Hidden Markov Model (HMM): Integrates a hidden Markov model to jointly infer CNV states across trio samples.
• Performance Validation: Validated on simulated datasets and a trio from the 1000 Genomes Project and shown to outperform existing methods in CNV detection accuracy.

Scientific Applications:

• CNV discovery in trios: Detection of inherited and de novo CNVs in parent-offspring trios for genetic studies.
• Disease-gene identification: Identification and prioritization of disease-associated genes through accurate CNV calls in hereditary disease research.

Methodology:

Joint modeling of trio data with read depth modeled by negative binomial regression, correction for GC content and mappability bias, and joint CNV inference via a hidden Markov model.