Detection of copy number variations based on a local distance using next-generation sequencing data

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Copy number variation (CNV) is a significant structural variation in the human genome.
• CNVs play a crucial role in the development and progression of human cancers.
• Accurate CNV detection from cancer samples with varying purity and low sequencing coverage remains challenging.

Purpose of the Study:

• To propose an innovative computational approach, LDCNV, for accurate CNV detection using next-generation sequencing (NGS) data.
• To address the challenges of CNV detection in low-purity and low-coverage cancer samples.
• To provide a tool that aids in cancer diagnosis and targeted drug development.

Main Methods:

• Developed Local Distance-based CNV detection (LDCNV) algorithm.
• Calculated average distances between read depths (RDs) and their k-nearest neighbors (KNNs).
• Constructed a local distance score using KNN distances and internal KNN distances, followed by hypothesis testing for CNV prediction.

Main Results:

• LDCNV demonstrated superior performance compared to existing methods in detecting CNVs.
• Validation using both simulated and real cancer genomic data confirmed the method's effectiveness.
• The proposed method shows high accuracy even with challenging sample characteristics like low purity and coverage.

Conclusions:

• LDCNV is an effective and accurate computational method for detecting copy number variations from NGS data.
• The method offers significant advantages over existing techniques, particularly for cancer genomics.
• LDCNV holds promise for improving cancer diagnosis and facilitating targeted drug development.