CNVcaller: highly efficient and widely applicable software for detecting copy number variations in large populations

Area of Science:

• Genomics
• Population Genetics
• Bioinformatics

Background:

• Population-level copy number variation (CNV) detection is limited by large datasets and complex nonhuman genomes.
• Existing human-focused CNV tools struggle with efficiency and accuracy in diverse species.

Purpose of the Study:

• To develop a robust and efficient CNV detection method for population sequencing data.
• To address the computational challenges posed by nonhuman genomes.

Main Methods:

• CNVcaller, a novel read-depth based method for CNV discovery.
• Utilized generalized detection algorithms for speed and scalability.

Main Results:

• CNVcaller demonstrated 1-2 orders of magnitude faster computational speed compared to existing tools for complex genomes.
• Accurate CNV detection in 232 goats was achieved in 1.4 days on a single node.
• Showcased high Mendelian consistency in sheep trios, mitigating reference genome assembly issues.
• Achieved superior accuracy and sensitivity in detecting duplications using sheep and human data.

Conclusions:

• CNVcaller overcomes computational barriers for CNV detection in large-scale, complex genomic datasets.
• The method enhances population genetic analyses of functional CNVs across a wider range of species.