Efficient calculation of interval scores for DNA copy number data analysis

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Cancer genomes are characterized by DNA amplifications and deletions, crucial for disease evolution.
• Microarray techniques provide high-resolution DNA copy-number data using fluorescence ratios.
• Analyzing these data is essential for mapping genomic aberrations and identifying significant structures.

Purpose of the Study:

• To develop a statistical framework for DNA copy number data analysis framed as optimization problems.
• To create efficient algorithms for identifying genomic aberrations and their boundaries.
• To apply these methods to both single samples and sets of cancer samples.

Main Methods:

• Formulating DNA copy number analysis as optimization problems over real-valued signal vectors.
• Developing a linear time approximation scheme (O(nepsilon(-2))) for maximizing signal-to-interval-size ratio.
• Implementing practical algorithms that significantly outperform naive quadratic approaches.

Main Results:

• A provable approximation scheme for a key optimization problem in DNA copy number analysis.
• Practical implementations offering substantial performance improvements over existing methods.
• Successful benchmarking on synthetic and real-world breast cancer DNA copy number data.

Conclusions:

• The developed statistical framework and algorithms provide efficient tools for DNA copy number analysis.
• These methods enhance the identification of genomic aberrations in cancer.
• The approach is applicable to diverse cancer datasets, aiding in understanding disease progression and identifying therapeutic targets.