PeakCNV

PeakCNV prioritizes copy number variation regions (CNVRs) in genome-wide association studies (GWAS) to distinguish true phenotype-associated CNVs from false positives and pinpoint biologically relevant genomic regions.

Key Features:

• Multi-Feature Ranking Algorithm: Uses a multi-feature ranking algorithm and computes an independence ranking score (IR-score) to assess the likelihood that a CNVR is genuinely associated with the phenotype.
• Reduction of False Positives: Minimizes false positives by emphasizing overlapping genomic regions where CNVs co-occur during CNVR construction.
• Efficiency in Candidate Identification: Benchmarking analyses show it identifies fewer and shorter risk candidate CNVRs than existing tools while covering a greater proportion of cases relative to healthy individuals.

Scientific Applications:

• Prostate Cancer Study: Involving 194 cases and 2,392 healthy controls, PeakCNV identified fewer candidate CNVRs that were more biologically meaningful compared to other tools.
• Neurodevelopmental Disorders Study: Analyzing 19,642 cases and 6,451 controls, candidate CNVRs showed significant overlap with genes exhibiting brain-enriched expression and associations with neurological conditions.

Methodology:

Integrates the FANTOM5 expression atlas and the Clinical Genomic Database, applies a multi-feature ranking algorithm, and computes an independence ranking score (IR-score) while focusing on overlapping CNV regions.