CoNCoS: copy number estimation in cancer with controlled support

Area of Science:

• Genomics
• Cancer Research
• Bioinformatics

Background:

• Somatic copy number (CN) alterations are key drivers in cancer development and progression.
• Next-generation sequencing (NGS) provides deep genomic insights but faces challenges due to data biases, including local read coverage variations.
• Existing algorithms for detecting CN aberrations from NGS data often neglect local coverage quality information.

Purpose of the Study:

• To introduce a novel algorithm, copy number estimation with controlled support (CoNCoS), designed to enhance the accuracy of CN estimation.
• To address the limitations of current methods by incorporating local coverage quality assessment.
• To improve the reliability of somatic CN analysis in paired tumor/normal exome sequencing data.

Main Methods:

• Developed the CoNCoS algorithm, which assesses and optimizes site-specific support for CN estimates.
• Utilized simulations to evaluate the performance of CoNCoS.
• Conducted a benchmarking study comparing CoNCoS against established methods like CNAnorm and VarScan2 using SNP microarray data.

Main Results:

• CoNCoS demonstrates superior performance compared to CNAnorm and VarScan2 in simulations.
• Benchmarking against SNP microarray data confirms the enhanced accuracy of CoNCoS.
• The algorithm effectively improves the precision of somatic CN estimation by optimizing support.

Conclusions:

• CoNCoS offers a significant advancement in the accurate detection of somatic CN alterations from NGS data.
• The support-optimized estimation approach addresses critical biases in genomic data analysis.
• This algorithm is valuable for improving the reliability of cancer genomic studies and clinical applications.