Related Experiment Video
Updated: Jan 1, 2026
11:02
Detecting Somatic Genetic Alterations in Tumor Specimens by Exon Capture and Massively Parallel Sequencing
Published on: October 18, 2013
19.8K
Comparison of somatic variant detection algorithms using Ion Torrent targeted deep sequencing data
Qing Wang1, Vassiliki Kotoula2,3, Pei-Chen Hsu1,4
1Victor Chang Cardiac Research Institute, Darlinghurst, Australia.
BMC Medical Genomics
|December 26, 2019
Summary
Somatic variant detection algorithms show low concordance on Ion Torrent deep sequencing data. Combining multiple methods and improving quality control are crucial for accurate cancer variant identification in research and clinics.
Area of Science:
- Genomics
- Bioinformatics
- Oncology
Background:
- Next-generation sequencing (NGS) is vital in cancer research, revealing genomic landscapes.
- Somatic variant detection algorithms are common but often evaluated on Illumina data.
- Comprehensive evaluation on Ion Torrent targeted deep sequencing data is lacking.
Purpose of the Study:
- Evaluate three somatic variant detection algorithms on Ion Torrent deep sequencing data.
- Assess algorithm concordance and performance in ovarian cancer.
- Provide recommendations for robust variant calling.
Main Methods:
- Applied Torrent Variant Caller, MuTect2, and VarScan2 to 208 paired tumor-blood and 253 tumor-only ovarian cancer samples.
- Utilized Ion Torrent Proton platform with deep sequencing across 330 amplicons.
- Assessed concordance, performance, and variant characteristics.
Main Results:
- Observed low concordance: 0.5% SNVs and 0.02% INDELs common across all three algorithms.
- Combined methods showed better performance correlating with mutational signatures and COSMIC database.
- Torrent Variant Caller demonstrated good performance and lower type II error rates.
Conclusions:
- Caution is advised when using current somatic variant algorithms with Ion Torrent deep sequencing data.
- Improved quality control and multi-method approaches are essential for accuracy.
- Robust bioinformatics pipelines are critical for clinical and research applications.
