SM-RCNV: a statistical method to detect recurrent copy number variations in sequenced samples.

Area of Science:

• Genomics
• Computational Biology
• Statistical Genetics

Background:

• Copy number variation (CNV) represents a key genomic structural variation implicated in numerous human diseases.
• Next-generation sequencing (NGS) data analysis and computational method development are crucial for elucidating disease mechanisms driven by structural variants.

Purpose of the Study:

• To introduce SM-RCNV, a novel statistical approach for identifying recurrent CNVs within multiple genomic samples.
• To enhance the understanding of genomic variations associated with human diseases through advanced computational methods.

Main Methods:

• Developed SM-RCNV, a statistical method integrating variation frequency across samples and correlation between consecutive genomic locations.
• Trained model weights using real-world datasets containing known CNVs.
• Employed permutation testing to assess the statistical significance (P-value) for each genomic location.

Main Results:

• SM-RCNV demonstrated superior performance compared to six existing methods, as evidenced by receiver operating characteristic (ROC) curves.
• Identified numerous biologically significant recurrent CNVs, many associated with known disease-related genes.
• Achieved high validation rates: 79% in the CEU dataset and 51% in the YRI dataset against the Database of Genomic Variants.

Conclusions:

• SM-RCNV provides a robust statistical framework for recurrent CNV detection from genomic sequences.
• The method offers valuable insights for studying the role of genomic variations in human diseases.
• Source code for SM-RCNV is publicly accessible for further research and application.