Robust regression analysis of copy number variation data based on a univariate score

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Copy number variants (CNVs) are prevalent in the human genome.
• Existing algorithms for CNV detection from intensity data suffer from high false discovery rates.
• Prioritizing CNV calls for experimental validation is challenging due to their dual dimensions (length and intensity).

Purpose of the Study:

• To develop a robust method for identifying and prioritizing true copy number variants (CNVs).
• To create a computationally efficient algorithm for CNV calling with reduced false discovery rates.

Main Methods:

• Development of a univariate score correlating with the likelihood of a true CNV.
• Implementation of cnv.beast, an algorithm utilizing backward elimination regression for CNV calling.
• Validation of the scoring method and algorithm on an independent dataset from a different platform.

Main Results:

• The developed univariate score effectively orders CNV calls, prioritizing those more likely to be true.
• cnv.beast demonstrated superior performance compared to six other existing CNV detection methods.
• The algorithm successfully reduced false discovery rates in CNV detection.

Conclusions:

• The proposed scoring system and cnv.beast algorithm offer a significant improvement in the accuracy and prioritization of CNV detection.
• This approach enhances the reliability of genomic analyses involving copy number variations.
• cnv.beast provides a computationally efficient and validated tool for researchers in the field.