ABSTRACT
Objective@#To compare the performance of Miseq and Ion Torrent PGM platforms and library construction method for next-generation sequencing (NGS) technology for formalin-fixed and paraffin-embedded (FFPE) samples.@*Methods@#A total of 204 FFPE cancer samples including 100 non-small cell lung cancers at the First Affiliated Hospital of Zhejiang University, and 104 colorectal cancers at West China Hospital of Sichuan University were retrospectively selected from January 2013 to December 2016. By using the same samples, DNA was extracted, and the same amount of DNA was used for library construction with the same kit, and sequenced on Miseq and Ion Torrent PGM respectively, after passing the quality control. Any discordant mutations between two platforms were validated by amplified refractory mutation system-polymerase chain reaction (ARMS-PCR) method and Sanger sequencing.@*Results@#A total of 204 FFPE samples were included and 197 samples were successfully analyzed by both platforms. The number of reads generated by the samples on Miseq platform sequencing was higher than PGM platform (median 391 634 vs. 298 030, P<0.01). Alignment with human reference genome showed that the mapping rate of Miseq platform was higher than PGM platform (median 100.0% vs. 99.7%, P<0.01). The median sequence depth of samples on Miseq was higher than PGM platform (median 853× vs. 698×, P<0.01). A total of 236 mutations were detected by two platforms, of which 221 were detected on both platforms, with a 93.6% concordance. Miseq platform detected 11 mutations not detected on PGM platform, while PGM platform detected 4 more mutations not detected on Miseq platform. With validation by ARMS-PCR and Sanger sequencing, Miseq platform was more reliable for low-frequency mutations. The main reasons for the discordant mutations between two platforms were that mutation frequency on undetected platform was lower than mutation reporting range (5%) and FFPE samples were stored for a long time.@*Conclusions@#Compared with PGM, Miseq platform shows higher sequencing quality in terms of the number of reads, alignment results and coverage depth, and the test results are more reliable. In clinical practice, the appropriate platform should be chosen based on sample size and actual throughput requirements to aid in the molecular characterization of tumors.
ABSTRACT
Objective@#To investigate whether small endoscopic biopsies of colorectal cancer were sufficient for quality and accurate mutational analysis by amplicon-based next-generation sequencing (NGS).@*Methods@#By using an amplicon-based targeted NGS panel for mutational detection on Illumina Miseq platform, a total of 109 formalin-fixed and paraffin-embedded (FFPE) endoscopic biopsies of colorectal cancer were retrospectively selected, based on specific histopathologic criteria, from January 2012 to June 2016 at West China Hospital of Sichuan University and Peking University Third Hospital. Twelve of these biopsies had corresponding FFPE surgical resection specimens. Quality control parameters of NGS testing were analyzed and NGS results were confirmed by other methods. Mutation calls of the 12 paired endoscopic biopsies and surgical resections were compared.@*Results@#Of the endoscopic biopsy specimens, 97.2% (106/109) had sufficient DNA and qualified sequencing library. NGS generated excellent sequencing data, with a median of 848× for median read depth and 95.7% for uniformity. The success rate of NGS was 95.4% (104/109). Conventional methods confirmed the results of NGS for KRAS and BRAF, and the concordance rate was 100.0%. The clinically actionable mutations detected in the 12 paired endoscopic biopsies and surgical resections were concordant.@*Conclusion@#FFPE endoscopic biopsies of colorectal cancer is suitable for targeted NGS, providing quality sequencing data and accurate mutational information to guide targeted therapy.