http://bmcmedgenomics.biomedcentral.com/articles/10.1186/s12920-016-0182-9

Comparison of two high-throughput semiconductor chip sequencing platforms in noninvasive prenatal testing for Down syndrome in early pregnancy

Sunshin Kim†, HeeJung Jung†, Sung Hee Han†, SeungJae Lee, JeongSub Kwon, Min Gyun Kim, Hyungsik Chu, Hongliang Chen, Kyudong Han, Hwanjong Kwak, Sunghoon Park, Hee Jae Joo, Byung Chul KimEmail author and Jong BhakEmail author

†Contributed equally

BMC Medical GenomicsBMC series – open, inclusive and trusted20169:22

DOI: 10.1186/s12920-016-0182-9

© Kim et al. 2016

Received: 30 September 2015

Accepted: 18 April 2016

Published: 30 April 2016

Open Peer Review reports

Abstract

Background

Noninvasive prenatal testing (NIPT) to detect fetal aneuploidy using next-generation sequencing on ion semiconductor platforms has become common. There are several sequencers that can generate sufficient DNA reads for NIPT. However, the approval criteria vary among platforms and countries. This can delay the introduction of such devices and systems to clinics. A comparison of the sensitivity and specificity of two different platforms using the same sequencing chemistry could be useful in NIPT for fetal chromosomal aneuploidies. This would improve healthcare authorities’ confidence in decision-making on sequencing-based tests.

Methods

One hundred and one pregnant women who were predicted at high risk of fetal defects using conventional prenatal screening tests, and who underwent definitive diagnosis by full karyotyping, were enrolled from three hospitals in Korea. Most of the pregnant women (69.79 %) received NIPT during weeks 11–13 of gestation and 30.21 % during weeks 14–18. We used Ion Torrent PGM and Proton semi-conductor-based sequencers with 0.3× sequencing coverage depth. The average total reads of 101 samples were approximately 4.5 and 7.6 M for PGM and Proton, respectively. A Burrows-Wheeler Aligner (BWA) algorithm was used for the alignment, and a z-score was used to decide fetal trisomy 21. Interactive dot diagrams from the sequencing data showed minimal z-score values of 2.07 and 2.10 to discriminate negative versus positive cases of fetal trisomy 21 for the two different sequencing systems.

Results

Our z-score-based discrimination method resulted in 100 % positive and negative prediction values for both ion semiconductor PGM and Proton sequencers, regardless of their sequencing chip and chemistry differences. Both platforms performed well at an early stage (11–13 weeks of gestation) compared with previous studies.

Conclusions

These results suggested that, using two different sequencers, NIPT to detect fetal trisomy 21 in early pregnancy is accurate and platform-independent. The data suggested that the amount of sequencing and the application of common, simple, and robust statistical analyses are more important than sequencing chemistry and platform types. This result has practical implications in countries where PGM is approved for NIPT but the Proton system is not.

Keywords Non-invasive prenatal testing – Sequencing – Circulating fetal DNA – Trisomy – Genome