Study on detection of CNVs using human whole genome bisulfite sequencing data

Area of Science:

• Genomics and Bioinformatics
• Epigenetics and DNA Methylation
• Computational Biology

Background:

• Aberrant DNA methylation is linked to copy number variations (CNVs), and CNVs can influence DNA methylation levels.
• Whole genome bisulfite sequencing (WGBS) generates DNA sequencing data with potential for CNV detection, but its performance is not well-established.

Purpose of the Study:

• To evaluate and benchmark the performance of different software tools for detecting CNVs using WGBS data.
• To identify the optimal strategy for CNV detection with WGBS data by comparing various software.

Main Methods:

• Five CNV detection software (BreakDancer, cn.mops, CNVnator, DELLY, Pindel) were selected for analysis.
• Performance was assessed using real (2.62 billion reads) and simulated (12.35 billion reads) human WGBS data.
• Metrics included detection counts, precision, recall, relative ability, memory usage, and CPU runtime, evaluated over 150 runs.

Main Results:

• CNVnator and cn.mops showed excellent performance in CNV detection using WGBS data.
• CNVnator exhibited high precision and recall for both deletions and duplications, potentially outperforming whole genome sequencing data.
• DELLY and BreakDancer had the lowest memory and CPU usage, while CNVnator had the highest.

Conclusions:

• Detecting CNVs using WGBS data is feasible and effective.
• CNVnator and cn.mops are recommended for CNV detection with WGBS data.
• Findings provide valuable information for simultaneous investigation of CNVs and DNA methylation using WGBS.