Related Experiment Video
Updated: May 6, 2026
08:38
Targeted DNA Methylation Analysis by Next-generation Sequencing
Published on: February 24, 2015
38.3K
Performance of common analysis methods for detecting low-frequency single nucleotide variants in targeted
David H Spencer1, Manoj Tyagi2, Francesco Vallania3
1Department of Pathology and Immunology, Washington University, St. Louis, Missouri.
The Journal of Molecular Diagnostics : JMD
|November 12, 2013
Summary
Detecting low-frequency cancer mutations with next-generation sequencing (NGS) is challenging. High-coverage NGS and specific variant callers like VarScan2 and SPLINTER significantly improve the detection of these critical cancer gene mutations.
Area of Science:
- Genomics
- Cancer Research
- Bioinformatics
Background:
- Next-generation sequencing (NGS) is crucial for clinical oncology testing.
- Detecting low-frequency cancer mutations is difficult due to tumor heterogeneity and normal cell contamination.
- Existing NGS analysis tools are often optimized for inherited variants, not somatic mutations.
Purpose of the Study:
- To evaluate the performance of four variant callers (SAMtools, Genome Analysis Toolkit, VarScan2, SPLINTER) for detecting low-frequency variants in oncology specimens.
- To determine the required sequencing coverage for accurate variant detection.
- To assess the utility of high-sensitivity variant callers in clinical lung cancer samples.
Main Methods:
- Generated high-coverage (>1000×) NGS data from synthetic DNA mixtures with low variant allele fractions (VAFs).
- Assessed variant callers SAMtools, Genome Analysis Toolkit, VarScan2, and SPLINTER.
- Analyzed variant detection sensitivity, specificity, and positive predictive value.
- Evaluated performance across different coverage levels and in clinical lung cancer samples.
Main Results:
- SAMtools showed low sensitivity (49% at ~25% VAF).
- Genome Analysis Toolkit, VarScan2, and SPLINTER detected >94% of variants at ~10% VAF.
- VarScan2 (97% sensitivity) and SPLINTER (89% sensitivity) excelled at 1-8% VAF.
- >500× coverage is necessary for optimal performance.
- VarScan2's false positive rate increased with coverage but was mitigated by read filtering.
Conclusions:
- High-coverage NGS combined with appropriate variant callers is essential for accurate detection of low-frequency cancer mutations.
- VarScan2 and SPLINTER demonstrate high sensitivity and positive predictive value for clinically relevant low-frequency variants.
- The findings support the use of advanced variant callers for clinical oncology testing and highlight the importance of sequencing depth.
