A comment on ‘Predicting fetoplacental chromosomal mosaicism during noninvasive prenatal testing’ by N. Brison et al. Brison et al.1 developed a mosaicism ‘add-on’ to Non-Invasive Prenatal Testing (NIPT) that identifies samples suspect of mosaicism and quantifies the extent of mosaicism by the expected size of mosaicism. The authors applied the tool within their NIPT pipeline and found that 21 out of 154 samples with confirmed trisomy 21 have the expected size of mosaicism greater than 100%, and hence are suspect of tetrasomic mosaicism. The authors found it intriguing. We believe that such a large number of samples suspect of tetrasomic mosaicism is an artifact of using a suboptimally calibrated NIPT. A NIPT is well-calibrated if the samples with confirmed 100% trisomic placental mosaicism lay in the trisomy band; see the zone between the dashed ‘diagonal’ lines on Figure 1 in Brison et al. Before applying the mosaicism add-on, one should, in our view, ensure that the underlying NIPT is well-calibrated. A key prerequisite for well-calibrated NIPT is a sufficiently precise method of fetal fraction estimation. It is known that the SeqFF2 method, used by the authors, substantially underestimates fetal fraction at the low3 and high ends of the range of fetal fraction values. The use of SeqFF makes the authors NIPT suboptimally calibrated, and invalidates some of their findings – in particular, the finding of the large percentage of tetrasomic mosaicism samples. The fact that SeqFF underestimates fetal fraction at the lower and higher end of values can already be seen on Figure 2A in Kim et al2., where the SeqFF method is compared with the chromosome Y-based method4 (ffY, hereafter), which is known to produce precise estimates of fetal fraction, though, it is applicable only to the male-carrying pregnancies. Data from our study5 confirm this observation; see Figure 1A. In the study a NIPT pipeline5 together with the SeqFF method of fetal fraction estimation lead 8 out of 13 trisomic male-carrying pregnancies outside the trisomy band; see Figure 1B. If the mosaicism add-on was applied to such a NIPT, then it would find 6 samples suspect of tetrasomic mosaicism (i.e., 46% of the trisomic male-carrying pregnancies) and 2 samples suspect of mosaicism (i.e, 15% of the trisomic male-carrying pregnancies). When the pipeline is used together with the ffY method, all the trisomic samples fall within the trisomy band. The NIPT with ffY is well-calibrated.
1. Brison N, Neofytou M, Dehaspe L, Bayindir B, Van Den Bogaert K, Dardour L, Peeters H, Van Esch H, Van Buggenhout G, Vogels A, de Ravel T. Predicting fetoplacental chromosomal mosaicism during non‐invasive prenatal testing. Prenatal diagnosis. 2018 Mar 1;38(4):258-66.
2. Kim SK, Hannum G, Geis J, Tynan J, Hogg G, Zhao C, Jensen TJ, Mazloom AR, Oeth P, Ehrich M, Boom D. Determination of fetal DNA fraction from the plasma of pregnant women using sequence read counts. Prenatal diagnosis. 2015 Aug 1;35(8):810-5. 3. Peng XL, Jiang P. Bioinformatics approaches for fetal DNA fraction estimation in noninvasive prenatal testing. International journal of molecular sciences. 2017 Feb 20;18(2):453. 4. Chiu RW, Akolekar R, Zheng YW, Leung TY, Sun H, Chan KA, Lun FM, Go AT, Lau ET, To WW, Leung WC. Non-invasive prenatal assessment of trisomy 21 by multiplexed maternal plasma DNA sequencing: large scale validity study. BMJ. 2011 Jan 11;342:c7401. 5. Grendár M, Loderer D, Švecová I, Laučeková Z, Hrtánková M, Mendelová A, Nagy B, Lasabová Z, Danko J. Exploring the quantitative foundations of Non-Invasive Prenatal Testing. Under review.
Marian Grendár1*, Dušan Loderer2*, Zora Lasabová1, Ján Danko2
1
Biomedical Center Martin, Jessenius Faculty of Medicine, Comenius University, Slovakia
2
Department of Obstetrics and Gynecology, Jessenius Faculty of Medicine, Comenius University, Slovakia *
These authors contributed equally.
Correspondence to: Marian Grendar. E-mail:
[email protected]
Funding sources: MG and ZL acknowledge support from APVV-16-0066 grant. Conflicts of interest: None
Received: April 11, 2018 Accepted: April 21, 2018
DOI: 10.1002/pd.5277
Figure 1A. SeqFF vs ffY, for 79 male-carrying pregnancies. The 45º line (in red) is added to facilitate comparison of the methods.
Figure 1B. The z-score vs SeqFF, for 13 trisomic male-carrying pregnancies. The ±3-sigma dashed lines are the borders of the trisomy band. There are 6 samples above the trisomy band (the expected size of mosaicism 152%, 136%, 129%, 150%, 131%, 218%) and 2 samples below the trisomy band (the expected size of mosaicism 78%, 86%). Note that Brison et al. work with the chromosome representation whereas we prefer to express the trisomy band in terms of the zscore. The two formulations are equivalent and lead the same conclusions.
Figure 1C. The z-score vs ffY, for 13 male-carrying pregnancies. None of the samples lay outside the trisomy band (delineated by the dashed lines).
