BACKGROUND: Noninvasive trisomy 21 detection performed by use of massively parallel sequencing is achievable with high diagnostic sensitivity and low false-positive rates. Detection of fetal trisomy 18 and 13 has been reported as well but seems to be less accurate with the use of this approach. The reduced accuracy can be explained by PCR-introduced guanine-cytosine (GC) bias influencing sequencing data. Previously, we demonstrated that sequence data generated by single molecule sequencing show virtually no GC bias and result in a more pronounced noninvasive detection of fetal trisomy 21. In this study, single molecule sequencing was used for noninvasive detection of trisomy 18 and 13.
METHODS: Single molecule sequencing was performed on the Helicos platform with free DNA isolated from maternal plasma from 11 weeks of gestation onward (n = 17). Relative sequence tag density ratios were calculated against male control plasma samples and results were compared to those of previous karyotyping.
RESULTS: All trisomy 18 fetuses were identified correctly with a diagnostic sensitivity and specificity of 100%. However, low diagnostic sensitivity and specificity were observed for fetal trisomy 13 detection.
CONCLUSIONS: We successfully applied single molecule sequencing in combination with relative sequence tag density calculations for noninvasive trisomy 18 detection using free DNA from maternal plasma. However, noninvasive trisomy 13 detection was not accurate and seemed to be influenced by more than just GC content.
Recent large studies have confirmed that noninvasive prenatal diagnosis for fetal aneuploidies is achievable (1–7). With the use of massively parallel sequencing (MPS)4 and subsequent quantification of chromosome-specific sequences, overrepresentation of a specific chromosome can be determined with high diagnostic accuracy. Successful detection of fetal trisomy 21 in maternal plasma was shown in several clinical validation studies (3–7). For noninvasive detection of trisomy 18 (Edwards syndrome) and trisomy 13 (Patau syndrome), however, it seems to be more difficult to achieve similar results (7–10). Although theoretically molecules from different regions of a genome should be sequenced uniformly by MPS, preferential amplification of sequences, depending on different guanine-cytosine (GC) content, has been observed (1, 11, 12). In contrast to the average GC content of chromosome 21, relatively low GC content occurs in chromosomes 13 and 18 (1, 12). Therefore, nonuniform amplification of these chromosomes could occur on PCR-based MPS platforms. As a result, several studies have used specific algorithms or internal references to correct for GC content to optimize noninvasive detection rates for trisomy 18 and trisomy 13 (8–10, 13).
We previously demonstrated that sequence data generated by single molecule sequencing show virtually no GC bias (12). This specific method of sequencing requires no PCR amplification step during sample preparation or during flow cell processing and results in a more pronounced noninvasive detection of trisomy 21. Therefore, this approach could also be applicable for the detection of other common fetal aneuploidies such as trisomy 18 and trisomy 13.
To test this hypothesis, we performed a retrospective study on first- and second-trimester pregnant women with an increased risk for fetal aneuploidy based on previous serum screening and/or ultrasound results. Maternal peripheral blood samples were collected in EDTA-coated tubes and processed within 24 h after collection. All blood samples were drawn at a median gestational age of 12 weeks + 6 days (range 11 weeks +4 days to 22 weeks +1 day, see Table 1) before an invasive procedure, except for 1 sample, which was obtained 6 days after amniocentesis. Plasma was obtained by double centrifugation of the blood samples and stored at −80°C until further processing. Material from all invasive procedures was sent to our cytogenetics laboratory for karyotyping as the gold standard.
A total of 21 plasma samples were used in this study. These samples included 4 plasma control samples from anonymous male blood donors and 17 samples from women with singleton pregnancies (Table 1), including 9 samples from pregnant women with trisomy 18 fetuses (2 female and 7 male fetuses), 4 with trisomy 13 fetuses (2 female and 2 male fetuses), and 4 with euploid fetuses (all male fetuses). Cell-free DNA was isolated from plasma by use of the EZ1 Virus Mini Kit v2.0. For QC purposes, fetal sex and the total amount of free DNA in maternal plasma were determined by real-time Taqman PCR assays as described previously (12, 14). In addition, using this data we estimated the percentage of cell-free fetal DNA (cffDNA) for male pregnancies (12). All samples were deidentified to the investigators before sample preparation and data analysis. Libraries were prepared according to the manufacturer's ChipSeq protocol and a standard 120-cycle sequencing run was performed on the Helicos platform (Helicos BioSciences, www.helicosbio.com). Raw data analysis was performed with the HeliSphere software package. Ratio calculations and statistical analyses were executed as described previously (12). In short, for fetal trisomy detection, ratios of relative sequence tag density (RSTD) were calculated by dividing the normalized total summed number of reads for each sample by the normalized mean of male plasma controls for each chromosome of interest. After alignment against the hg19 reference genome and filtering of gaps and repeats, a mean (SD) of 1.21 × 106 (0.69 × 106) reads, with a median of 1.12 × 106, were obtained per sample. In 11 maternal plasma samples from women carrying a male fetus, the percentage of cffDNA was estimated, resulting in a mean percentage of 11% (Table 1).
For noninvasive trisomy 18 detection we showed that with the use of RSTD calculations for chromosome 18, all trisomy 18 samples (n = 9) were correctly identified as being aneuploid and all euploid controls and trisomy 13 samples as being disomic for chromosome 18. When constructing a 99% CI from all samples disomic for chromosome 18 (n = 8), we found that all trisomy 18 samples were outside the upper boundary of the 99% CI (0.991–1.016), while all euploid controls and trisomy 13 samples were on or below this upper boundary (Fig. 1). For noninvasive trisomy 13 detection, only 1 of 4 trisomy 13 samples was correctly identified. False-positive results (4/13) were observed in both euploid (n = 2) and trisomy 18 (n = 2) samples when we used the RSTD ratio and 99% CI calculations for chromosome 13, resulting in a diagnostic sensitivity and specificity of 25% and 69%, respectively (Fig. 1). As a control we calculated RSTD ratios for chromosome 21 for all samples tested in this study (n = 17) using the 99% CI previously published (12). All samples tested in this study were indeed identified as disomic for chromosome 21 (Fig. 1). When we calculated a 99% CI using RSTD results from this study a similar upper boundary was obtained, thus confirming this result.
As a follow-up on noninvasive trisomy 21 detection using single molecule sequencing, in the present study we demonstrated successful noninvasive detection of trisomy 18 (100% diagnostic sensitivity and specificity) using free DNA from maternal plasma from 11 weeks + 4 days of gestation onwards. The mean percentage of cffDNA in maternal plasma in the first trimester was 4.0%, and we observed an increase in the fetal fraction during the second trimester, with a mean percentage of 21.1%. This observation is concordant with previous reports (15, 16). Although the percentage increased, we still observed quite a large range between individuals, with an approximate 4-fold change for the second-trimester pregnancies and up to a 13-fold difference between first-trimester samples.
Compared to noninvasive detection of trisomy 18, our data showed low diagnostic sensitivity and specificity for detection of trisomy 13 with the use of single molecule sequencing. In previous publications other groups also reported reduced diagnostic sensitivity and/or specificity for noninvasive trisomy 13 detection (7–10). However, the values were not as low as those observed in this study. Furthermore, in the previous cases the reductions in diagnostic sensitivity and/or specificity were thought to be related to the GC content of chromosome 13. given that PCR-based next generation sequencing (NGS) platforms were used. As shown in our previous study, data for chromosome 13 are biased on such platforms (12). Chromosome 13, compared to 18 and 21, has the lowest GC content of all 3 (38.5%) (17). This low GC content could be the reason for a misrepresentation of the amount of sequencing reads coming from these PCR-based NGS platforms. However, in the current study, single molecule sequencing results were not influenced by a chromosome's GC content, implying that other factors might be involved in lowering the diagnostic sensitivity and specificity for noninvasive trisomy 13 detection.
The fetal contribution of free DNA in maternal plasma is derived from syncytiotrophoblasts undergoing apoptosis (18). Placental apoptosis is a naturally occurring process during gestation in both normal and abnormal pregnancies, resulting in fragmented fetal DNA circulating in the maternal circulation (18, 19). Some studies have demonstrated the difference in size between fetal and maternal free DNA fragments and have even shown that the entire fetal genome is present (20). However, virtually no studies have considered that fetal DNA from chromosomes of different sizes and/or those of differing GC contents may fragment at different rates. Considering that chromosome 13 is the largest acrocentric chromosome with the lowest gene density among all human chromosomes (17), its stability may differ from those of chromosomes 18 and 21. A less stable chromosome is hypothesized to degrade faster, which could lead to a skewed number of DNA fragments from this particular chromosome in maternal plasma. Also, several segmental duplications with at least 90% homology and regions with a high single-nucleotide polymorphism density due to the presence of paralogous sequence variants have been shown for chromosome 13 (17). These features may influence data analysis, resulting in the improper assignment of reads to a certain chromosome during alignment. Which factors exactly play a role is not clear at this point and must be studied in more detail; however, our study suggests that sequencing data are influenced by more than just GC content alone. Data analysis for noninvasive fetal trisomy 13 detection may therefore require a different approach.
In summary, we demonstrated successful noninvasive trisomy 18 detection using a combination of single molecule sequencing and relative sequence-tag density ratio calculations, whereas noninvasive trisomy 13 detection was not accurate using this approach.
The authors thank all participants in this study, Jennie Verdoes from the Department of Obstetrics of the LUMC for including the pregnant women, and Michiel van Galen and Henk Buermans for bioinformatics support.
↵4 Nonstandard abbreviations:
- massively parallel sequencing;
- cell-free fetal;
- relative sequence tag density;
- next generation sequencing.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: EuroGentest2 grant 261469.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
- © 2013 The American Association for Clinical Chemistry