Details

Autor(en) / Beteiligte

• Hwang, Sang Mee
• Lee, Ki Chan
• Lee, Min Seob
• Park, Kyoung Un

Titel

Comparison of Ion Personal Genome Machine Platforms for the Detection of Variants in BRCA1 and BRCA2

Ist Teil von

• Cancer Research and Treatment, 2018, 50(1), , pp.255-264

Ort / Verlag

Korea (South): Korean Cancer Association

Erscheinungsjahr

2018

Link zum Volltext

Quelle

MEDLINE

Beschreibungen/Notizen

• Transition to next generation sequencing (NGS) for / analysis in clinical laboratories is ongoing but different platforms and/or data analysis pipelines give different results resulting in difficulties in implementation. We have evaluated the Ion Personal Genome Machine (PGM) Platforms (Ion PGM, Ion PGM Dx, Thermo Fisher Scientific) for the analysis of /2. The results of Ion PGM with OTG-snpcaller, a pipeline based on Torrent mapping alignment program and Genome Analysis Toolkit, from 75 clinical samples and 14 reference DNA samples were compared with Sanger sequencing for / . Ten clinical samples and 14 reference DNA samples were additionally sequenced by Ion PGM Dx with Torrent Suite. Fifty types of variants including 18 pathogenic or variants of unknown significance were identified from 75 clinical samples and known variants of the reference samples were confirmed by Sanger sequencing and/or NGS. One false-negative results were present for Ion PGM/OTG-snpcaller for an indel variant misidentified as a single nucleotide variant. However, eight discordant results were present for Ion PGM Dx/Torrent Suite with both false-positive and -negative results. A 40-bp deletion, a 4-bp deletion and a 1-bp deletion variant was not called and a false-positive deletion was identified. Four other variants were misidentified as another variant. Ion PGM/OTG-snpcaller showed acceptable performance with good concordance with Sanger sequencing. However, Ion PGM Dx/Torrent Suite showed many discrepant results not suitable for use in a clinical laboratory, requiring further optimization of the data analysis for calling variants.