Robust and exact structural variation detection with paired-end and soft-clipped alignments: SoftSV compared with...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Structural variations (SVs) significantly impact genetic diversity and are implicated in diseases like cancer.
• Next-generation sequencing (NGS) enables high-resolution detection of SV breakpoints using split-read analysis.
• Challenges include ambiguous alignments and the sensitivity-specificity trade-off in current SV detection methods.

Purpose of the Study:

• To introduce SoftSV, a novel method for exact breakpoint detection of various structural variations.
• To evaluate SoftSV's performance against existing SV detection algorithms.
• To demonstrate SoftSV's robustness and sensitivity across different sequencing and SV types.

Main Methods:

• SoftSV utilizes the mutual alignment of soft-clipped reads near discordantly mapped paired-end reads for breakpoint detection.
• The method does not rely on sequencing coverage or mapping quality thresholds.
• Comparative analysis with eight other SV detection tools on simulated and real datasets.

Main Results:

• SoftSV demonstrated robust and sensitive detection across various SV types (deletions, inversions, duplications, translocations).
• High sensitivities (>85%) and positive predictive values (80-100%) were achieved consistently, even with short reads (75 bp) and low coverage (10-15×).
• SoftSV proved more robust and sensitive than most evaluated methods across diverse experimental setups.

Conclusions:

• SoftSV offers a reliable and threshold-independent approach for precise structural variation detection.
• The method is effective across a range of SV types and sequencing parameters, improving upon existing tools.
• SoftSV advances the analysis of genetic diversity and cancer genomics through accurate SV identification.