Massively Parallel Sequencing for Chromosomal ... - CiteSeerX

BIOLOGY OF REPRODUCTION (2013) 88(3):69, 1–6 Published online before print 24 January 2013. DOI 10.1095/biolreprod.112.106211

Massively Parallel Sequencing for Chromosomal Abnormality Testing in Trophectoderm Cells of Human Blastocysts1 XuYang Yin,5,6 Ke Tan,5,10 Ga´bor Vajta,5,6,8 Hui Jiang,5,6,9 YueQiu Tan,7,10,11 ChunLei Zhang,6 Fang Chen,6,9 ShengPei Chen,6,13 ChunSheng Zhang,6 XiaoYu Pan,6,14 Chun Gong,6 XuChao Li,6 ChuYu Lin,6 Ya Gao,6 Yu Liang,6 Xin Yi,6 Feng Mu,6 LiJian Zhao,6 HuanHuan Peng,6 Bo Xiong,11 ShuoPing Zhang,7,11,12 DeHua Cheng,11 GuangXiu Lu,7,10,11,12 XiuQing Zhang,4,6 Ge Lin,3,7,10,11,12 and Wei Wang2,6 6

BGI-Shenzhen, Shenzhen, China Institute of Reproductive and Stem Cell Engineering, Central South University, Changsha, China 8 Institute for Resource Industries and Sustainability (IRIS), Central Queensland University, Rockhampton, Queensland, Australia 9 Department of Biology, University of Copenhagen, Copenhagen, Denmark 10 National Engineering and Research Center of Human Stem Cell, Changsha, China 11 CITIC Xiangya Reproductive & Genetic Hospital, Changsha, China 12 Key Laboratory of Stem Cell and Reproductive Engineering, Ministry of Health, Changsha, China 13 State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China 14 School of Bioscience and Bioengineering, South China University of Technology, Guangzhou, China 7

Preimplantation genetic diagnosis and screening are widely accepted for chromosomal abnormality identification to avoid transferring embryos with genetic defects. Massively parallel sequencing (MPS) is a rapidly developing approach for genome analysis with increasing application in clinical practice. The purpose of this study was to use MPS for identification of aneuploidies and unbalanced chromosomal rearrangements after blastocyst biopsy. Trophectoderm (TE) samples of 38 blastocysts from 16 in vitro fertilization cycles were subjected to analysis. Low-coverage whole genome sequencing was performed using the Illumina HiSeq2000 platform with a novel algorithm purposely created for chromosomal analysis. The efficiency of this MPS approach was estimated by comparing results obtained by an Affymetrix single-nucleotide polymorphism (SNP) array. Whole genome amplification (WGA) products of TE cells were detected by MPS, with an average of 0.073 depth and 5.5% coverage of the human genome. Twenty-six embryos (68.4%) were detected as euploid, while six embryos (15.8%) contained uniform aneuploidies. Four of these (10.5%) were with solely unbalanced chromosomal rearrangements, whereas the remaining two embryos (5.3%) showed both aneuploidies and unbalanced rearrangements.

aneuploidy, blastocyst trophectoderm cells, massively parallel sequencing, SNP array, unbalanced chromosomal rearrangement

INTRODUCTION Chromosomal abnormalities are among the major reasons for fetal developmental defects, causing implantation failures, spontaneous abortions, or late anomalies, or necessitating artificial termination of pregnancy [1]. To reduce the incidence of those defects, preimplantation genetic diagnosis and screening (PGD/PGS) have been introduced to detect a suspected abnormality or select the embryos with the highest developmental potential. According to the European Society of Human Reproduction and Embryology (ESHRE) consortium, embryos of over 3000 Robertsonian translocation carriers and reciprocal translocation carriers had been subjected to PGD from 1997 to 2007 [2]. Advanced maternal age, repeated implantation failure, and repeated miscarriage are the main indications for PGS to detect numerical chromosome errors [3]. After the first report of PGD for X-linked diseases in 1990 [4], single-blastomere biopsy from precompaction-stage embryos following fluorescence in situ hybridization (FISH) and polymerase chain reaction (PCR) have become the dominant techniques for embryo analysis [5]. However, these approaches are restricted to selected chromosomes, and the information obtained from a single cell is not always representative of the whole embryo. In contrast, comparative genomic hybridization (CGH) or its advanced form, arraybased CGH, provide more information by analyzing all chromosomes at one time [6–9]. Moreover, with the introduction of single-nucleotide polymorphism (SNP) array

1

Supported by Key Laboratory Project in Guang Dong Province (2011A060906007), Key Laboratory Project in Shenzhen (Shenzhen Municipal Commission of Development and Reform [2011] 861), National Natural Science Funds for Excellent Young Scholar (81222007), and National Basic Research Program of China (973 program 2012CB944901). 2 Correspondence: Wei Wang. E-mail: [email protected] 3 Correspondence: Ge Lin. E-mail: [email protected] 4 Correspondence: XiuQing Zhang. E-mail: [email protected] 5 These authors contributed equally to this work. Received: 11 November 2012. First decision: 20 December 2012. Accepted: 10 January 2013. Ó 2013 by the Society for the Study of Reproduction, Inc. eISSN: 1529-7268 http://www.biolreprod.org ISSN: 0006-3363

1

Article 69

Downloaded from www.biolreprod.org.

Almost all these results were confirmed by the SNP array, with the exception of one sample, where different sizes of unbalanced rearrangements were detected, possibly due to chromosomal GC bias in array analysis. Our study demonstrated MPS could be applied to accurately detect embryonic chromosomal abnormality with a flexible and cost-effective strategy and higher potential accuracy.

ABSTRACT

YIN ET AL. detached from the blastocyst by laser firing on the area of constriction. Samples were then washed three times in PBS and preserved in PCR tubes with 2 ll PBS at 808C.

technologies, magnificent detection accuracy and resolution has been achieved by providing information not only for the chromosomal number but also the subchromosomal alterations [10]. With the combination of trophectoderm (TE) cell biopsy, blastocyst vitrification, and SNP array for diagnosis, Schoolcraft et al. [11] reported a significant improvement of the clinical outcome of implantation and live birth rate. However, the throughput and cost of an SNP array or an array-based CGH may limit widespread application in clinical laboratories. Another rapidly developing and promising technique for chromosomal examination is massively parallel sequencing (MPS). Since the completion of the Human Genome Project [12], the rapid innovation and advancement of sequencing technologies has led to dramatic reduction of the cost and turnaround time, as well as an astonishing improvement in sequencing quality. Compared to traditional capillary electrophoresis sequencing, MPS, which obtains the sequence of over a million DNA fragments, provides a more cost-effective approach for genomic study. Several recent studies have demonstrated the value of MPS in clinical applications, such as noninvasive prenatal diagnosis for fetal aneuploidy through low-coverage whole genome sequencing of maternal plasma [13] and identification of causative mutation in patients with Mendelian disease [14]. Moreover, Navin et al. [15] applied the combination of whole genome amplification (WGA) with MPS to detect copy number variations (CNVs) in single cells from tumor tissues. Another two studies demonstrated the possibility of analyzing a whole genome at single-nucleotide level in single cells [16, 17]. Encouraged by this progress in clinical applications of MPS, the purpose of our work was to explore the potential of MPS in analysis of blastocyst TE biopsy samples for detection of aneuploidy and unbalanced chromosomal rearrangements, and the performance of MPS was evaluated by comparing the results with those of the SNP array using the same samples.

WGA of Genomic DNA of TE Cells WGA was performed using the WGA4 GenomePlex Single Cell Whole Genome Amplification Kit (WGA4; Sigma Aldrich Inc.) according to the manufacturer’s instructions. Briefly, samples were incubated at 508C for 1 h and heated to 998C with Single Cell Lysis & Fragmentation Buffer, then universal oligonucleotide primers were used to amplify the DNA fragment with 25 cycles of PCR. The quality of the amplified DNA was tested on a 2% agarose gel.

Library Construction and DNA Sequencing WGA product (1–2 lg) was used for library construction according to the guidelines of Illumina sample preparation (Illumina). After fragmentation, the sticky ends of the DNA fragments were polished with T4 DNA polymerase and Klenow polymerase, then A base tailing was performed using Klenow exo- (3 0 to 5 0 exo minus). Specific sequencing adaptors with a single T base overhang at the 3 0 end were ligated to the above products and enriched by 10 cycles of PCR with an eight-base barcode of primers. After measuring the concentration and insert size, libraries were processed for single-end 50 cycles high-throughput sequencing using an Illumina HiSeq2000 sequencer (Illumina).

In order to evaluate the accuracy of aneuploidy and unbalanced chromosomal rearrangement detection using an MPS-based approach, donated Day 5 or 6 blastocysts were biopsied to obtain TE cells. After WGA, PCR products were analyzed by both MPS and SNP array. The inconsistencies between MPS and SNP arrays were further validated by quantitative real-time PCR (qPCR).

The initial 20-base sequence of read was omitted for the following analysis to avoid the effects of the common WGA adaptor sequence. The remaining 30base sequences were aligned to Human reference genome (Hg18; NCBIBuild36, http://www.ncbi.nlm.nih.gov/genome/guide/human/) by SOAP2 [18], allowing a maximum of two base pair mismatches. After alignment, an algorithm developed in BGI-Shenzhen was used for chromosomal abnormality analysis. This algorithm included GC correction for WGA-induced bias removal, a binary segmentation algorithm for locating CNV breakpoints, and dynamic threshold determination for final signal filtering. For sex determination, the coverage ratio of the whole genome was used to assess the copy ratio of the Y chromosome, with 0.35 as a threshold. To compare with the results from the SNP array, we filtered the signals larger than 16 Mb in MPS analysis. All results were visualized by digital karyotyping for better presentation [19]. Samples with chromosomal abnormalities were further selected for data simulation to evaluate the sensitivity and specificity of aneuploidy and CNV detection with different data sizes. Effective data from each sample that could be aligned to the reference genome were used for simulation with each data size, obtaining files with similar data size to be analyzed for sensitivity/ specificity evaluation. The positive signals of simulation with size more than half of the region were determined as abnormal, and the positive abnormal signal detected by simulation that was the same as the initial sample result was determined as true positive. The sensitivity was calculated as the ratio of the instances of true positive with the number of simulations. The specificity was calculated as the ratio of true positive signals with total positive signals of the initial sample.

Collection of Samples

SNP Array Analysis

Thirty-eight blastocysts from 16 couples were collected after assisted reproduction treatment procedures in the CITIC Xiangya Reproductive & Genetic Hospital, with an average maternal age of 28.9 yr. Twenty-nine embryos were from 14 couples with structural chromosomal abnormalities such as translocations, and nine embryos were from another two euploid couples. Karyotype data of parents with chromosomal abnormalities are listed in Table 1. All samples were obtained with the approval of the Institutional Review Board of both BGI-Shenzhen and CITIC Xiangya Reproductive & Genetic Hospital. All couples have participated in this trial on an informed-choice basis and have signed a consent form about donation of their leftover blastocysts for the purposes of this experiment.

WGA products were also processed for the SNP array to analyze aneuploidy and unbalanced chromosomal rearrangement in 23 chromosomes (the Y chromosome was not included). The amplified individual embryonic DNA was hybridized to the Gene Chip Mapping Nsp I 262K microarray (Affymetrix Inc.). Approximately 260 000 SNP signal intensities for each test sample were compared computationally with averaged signals from 30 previously evaluated normal female reference samples (data not shown). Copy number analysis was performed by Gene Chip Genotyping Analysis Software (GTYPE; Affymetrix Inc.) with smoothing size at 16M to minimize the noise hybrid signals. As the SNP array is unable to detect the Y chromosome, Y chromosom-specific PCR was performed for sex determination.

TE Cells Biopsy

Quantitative Real-Time PCR Validation

All embryos were cultured in Quinn’s Blastocyst Medium (Sage) to the blastocyst stage. A hole was created in the zona pellucida on Day 3. Hatching blastocysts with the trophectodermal layer partially protruding through the hole of the zona on Day 5 or 6 were biopsied. Three to eight herniating TE cells were aspirated into the lumen of a pipette with internal diameter of 30 lm and

The inconsistent results between the MPS and the SNP array were further validated by qPCR. For each region, three loci primers were designed and triplicate PCR was performed using blastocyst WGA product in 20 ll reaction volume with TaqMan Gene Expression Master Mix (Life Technologies) on a 7900 HT sequence detection system, as recommended by the supplier.

MATERIALS AND METHODS Study Design

2

Article 69


Bioinformatics Analysis for Chromosomal Abnormality Detection

MPS FOR CHROMOSOMAL TESTING IN TE CELLS TABLE 1. Summary of the karyotypes of 38 trophectoderm cell samples by SNP array and MPS, with the karyotype obtained by G-banding karyotyping of the IVF parent with chromosomal abnormalities. Sample

SNP array result

MPS result

46,XY,t(11;22)(q23;q11) 46,XY,t(11;22)(q23;q11) 46,XY,t(11;22)(q23;q11) 46,XY,t(11;22)(q23;q11) 45,XX,der(14;21)(q10;q10) 46,XY,t(3;21)(q21;q22) 46,XY,t(3;21)(q21;q22) 46,XY,t(1;2)(q41;q22) 46,XY,t(1;2)(q41;q22) 46,XY,t(X;22)(q22;q11) 46,XY,t(X;22)(q22;q11) 46,XY,inv(2)(p11q13) 46,XY,t(6;12)(q11;p11) 45,XX,der(14;21)(q10;q10) 46,XY,t(11;17) 46,XY,t(11;17) 46,XY,t(11;17) 46,XX,t(3;19)(q29;q13) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 45,XX,der(13;14)(q10;q10) 46,XX,t(4;19)(p14;p13) 46,XX,t(5;13)(q22;q12) 46,XX,t(5;13)(q22;q12)

46,XY 46,XX 45,XY,-21 46,XX 46,XX 46,XY 46,XX 46,XX 46,XX 46,XY 46,XY,þ11(p15.4-p13), þ22(q12.3-qter) 46,XY 46,XX 46,XX 46,XY,-3(q22-qter),þ21(q21-qter) 46,XY,-3,þ21 46,XY,þ1(q41-qter), 2(q14.3-qter) 46,XY 46,XY 46,XY 46,XY 46,XX 46,XX 46,XX 45,XY,þ17(pter-q11),-21 46,XY,-6(pter-p21),þ17(pter-p12) 47,XX,þ3 46,XX 46,XY 46,XY 47,XY,þ13 46,XX 46,XX 46,XY 46,XY 47,XX,þ14 47,XX,þ1,þ5(q22-qter), 13(q14-qter) 45,XY,-7

46,XY 46,XX 45,XY,-21 46,XX 46,XX 46,XY 46,XX 46,XX 46,XX 46,XY 46,XY,þ11(pter-q23.3),-11(q23.3-qter),þ22(q11.21-qter) 46,XY 46,XX 46,XX 46,XY,-3(q22-qter), þ21(q22.12-qter) 46,XY,-3,þ21 46,XY,þ1(q41-qter),-2(q21.2-qter) 46,XY 46,XY 46,XY 46,XY 46,XX 46,XX 46,XX 45,XY,þ17(pter-p11.2),-21 46,XY,-6(pter-p12.3), þ17(pter-p11.2) 47,XX,þ3 46,XX 46,XY 46,XY 47,XY,þ13 46,XX 46,XX 46,XY 46,XY 47,XX,þ14 47,XX,þ1,þ5(q22.1-qter), 13(q14.12-qter) 45,XY,-7

* The samples in which one of both derivative chromosomes of the IVF patient was transmitted or partly transmitted to the offspring blastocyst.

Blastocysts with euploid karyotype were used as controls. The standard delta delta threshold cycle (DDCt) method of relative quantitation was applied to calculate the copy ratio [20].

chromosomes of the IVF parent was transmitted, including BLS11, BLS15, BLS17, and BLS37 (Table 1). Twenty one male and 17 female embryos were identified according to the coverage ratio of chromosome Y. Combining the low-coverage whole genome sequencing with the algorithm for analysis, 68% (26/38) of embryos were found euploid. Thirty-two percent (12/38) of embryos contained chromosomal abnormalities, including 15.8% (6/38) with uniform aneuploidies, 10.5% (4/38) with solely unbalanced rearrangements, and 5.3% (2/38) with both aneuploidies and unbalanced rearrangements. In total, four monosomies and five trisomies were identified in chromosomes 1, 3, 7, 13, 14, and 21, respectively, and 14 events of unbalanced rearrangements were observed in chromosomes 1, 2, 3, 5, 6, 11, 13, 17, and 21, respectively (Table 1). Using and MPS-based approach, 12 embryos were classified as containing at least one chromosomal error.

RESULTS Data Production An average of 9.3 million reads were obtained for each embryo by MPS, covering 5.6% 6 1.9% of the whole human genome, with an average of 0.073 sequencing depth; 80.4% 6 3.5% of reads were aligned to the reference sequence and 71.9% 6 3.6% of reads were uniquely mapped by SOAP2, with an average duplication rate of 6.9%. The GC content was 41.3% 6 0.6% in accordance with that of the Human reference genome. The average coverage for mitochondrial DNA was 97.9% 6 2.6%, with an average 16.33 sequencing depth, which was 233 times higher than the depth of the whole genome.

Comparison of MPS and SNP Array Results Identification of MPS-Based Aneuploidies and Unbalanced Chromosomal Rearrangements

The SNP array also detected 26 euploid blastocysts and 12 embryos with chromosomal abnormalities (Table 1). All 26 euploid embryos as well as six embryos with uniform aneuploidies were correctly identified both by the MPS and SNP arrays (Table 1). Eighty-three percent (5/6) of embryos

Karyotypes of in vitro fertilization (IVF) patients were identified by G-banding karyotyping of peripheral blood. There were four blastocysts for which one of both derivative 3

Article 69


BLS1 BLS2 BLS3 BLS4 BLS5 BLS6 BLS7 BLS8 BLS9 BLS10 BLS11* BLS12 BLS13 BLS14 BLS15* BLS16 BLS17* BLS18 BLS19 BLS20 BLS21 BLS22 BLS23 BLS24 BLS25 BLS26 BLS27 BLS28 BLS29 BLS30 BLS31 BLS32 BLS33 BLS34 BLS35 BLS36 BLS37* BLS38

Karyotype of the IVF parent

YIN ET AL. TABLE 2. ISCN nomenclature for SNP array and MPS results of the 12 samples with chromosomal abnormalities. Sample

Finding

BLS3 BLS11

Monosomy 21 11p gain 11q gain 11q loss 22q gain 3q loss 21q gain Monosomy 3 Trisomy 21 1q gain 2q loss 17 chr gain Monosomy 21 6p loss 17p gain Trisomy 3 Trisomy 13 Trisomy 14 Trisomy 1 5q gain 13q loss Monosomy 7

BLS15 BLS16 BLS17 BLS25 BLS26 BLS27 BLS31 BLS36 BLS37 BLS38

SNP array result snp arr snp arr None None snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr snp arr

MPS result

21pter!qter (9887804–46894358)31 11p15.4.11p13 (2801017–35947626)x3 22q12.2!qter (44755205–49465770)x3 3q22!qter (132564802–199322068)x1 21q21!qter (28884248–46894358)x3 3pter!qter (48603–199322068)x1 21pter!qter (9887804–46894358)33 1q41!qter(214860843–245353397)x3 2q14.3!qter(127558160–242717659)x1 17pter!q11 (18901–26858750)x3 21pter!qter (9887804–46894358)31 6pter!p21 (119769–45645653)x1 17pter!p12 (18901–15357533)x3 3pter!qter (48603–199322068)33 13pter!qter (17960319–114092980)33 14pter!qter (19336854–106356482)33 1pter!qter (825852–245353397)33 5q22!qter(116910275–180629495)x3 13q14!qter(44552688–114092980)x1 7pter!qter (141322–158605053)31

seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq seq

21pter!qter (1–46944322)x1 11pter!11p11.2 (1–50450780)x3 11q11!q23.3 (55450782–116045647)x3 11q23.3!qter (116045647–134366790)x1 22q11.21!qter(18589455–49403974)x3 3q22!qter (133226185–199501826)x1 21q22.12!qter (35823268–46944322)x3 3pter!qter (1–199501826)x1 21pter!qter (1–46944322)x3 1q41!qter (215429472–247249718)x3 2q21.2!qter (127731682–242951148)x1 17pter!p11.2(1–22089976)x3 21pter!qter (1–46944322)x1 6pter!p12.3(1–46293963)x1 17pter!p11.2(1–22662667)x3 3pter!qter (1–199501826)x3 13pter!qter (1–114142979)x3 14pter!qter (1–106368584)x3 1pter!qter (1–247249718)x3 5q22.1!qter(111022764–180857865)x3 13q14.12!qter(44523338–114142979)x1 7pter!qter (1–158821423)x1

FIG. 1. Examples of MPS and SNP array results of BLS11 from an IVF father with 46,XY,t(11;22)(q23;q11). A) Digital karyotyping of MPS result of BLS11. There are þ11 (pter-q23.3) and a þ22 (q11.21-qter) duplications detected as green regions, and a 11 (q23.3-qter) deletion detected as red regions. Both the imbalanced segments were validated by qPCR. B) SNP array result of BLS11. There is a þ11 (p15.4-p13) duplication and a þ22 (q12.3-qter) duplication detected as a shift of copy number in the region from 2 to 3 (gain). Failed detection for the breakpoint in q23.3 of chromosome 11 in the SNP array platform may be due to the bias from WGA that affected the result, but it can be corrected by MPS during bioinformatics analysis.

4

Article 69


There were some discrepancies in the size of imbalanced segments between MPS and the SNP array, which might be related to the density of the probes. For instance, in sample BLS15 (Fig. 2), we identified a 14.0-Mb duplication in 21 (q22.12-qter) by sequencing (Fig. 2A), but a 21.1-Mb duplication in 21 (q21-qter) by the SNP array (Fig. 2B). The duplication in chromosome 21 of BLS15 was validated by qPCR as true, which was most possible from the father with a karyotype of 46, XY, t(3;21)(q21;q22). Results above demonstrated that both MPS and the SNP array can identify embryo aneuploidy with 100% consistency. For the embryos with unbalanced chromosomal rearrangements, an MPS-based approach provided higher accuracy in some regions.

with unbalanced chromosome abnormalities showed consistent testing results as well (Table 2). MPS revealed þ11 (pter-q23.3), 11 (q23.3-qter), and þ22 (q11.21-qter) of BLS11 (Fig. 1). However, this anomaly of duplication in 11 (q11!q23.3) and deletion in 11 (q23.3!qter) were not detected in the SNP array for this sample (Table 2). The two incongruous segments and 22 (q11.21-qter) in BLS11 were validated by qPCR, where the copy ratio of 11 (pter-q23.3) and 22 (q11.21-qter) was identified as three by all three loci, indicating a duplication, and the copy ratio of 11 (q23.3-qter) was identified as one by all three loci, indicating a deletion. The failed detection for the imbalance and breakpoint above in the SNP array may due to the bias from WGA that affected the SNP array result, but was corrected by MPS during bioinformatics analysis.

MPS FOR CHROMOSOMAL TESTING IN TE CELLS

obtained with and SNP array. For the first time, we explored the use of MPS in chromosomal abnormality testing of human embryos, and demonstrated that low-coverage sequencing combined with WGA and bioinformatics analysis can effectively detect aneuploidies and unbalanced chromosomal rearrangements in TE cells. A major advantage of this MPS-based approach to detection of chromosomal abnormality is the high accuracy of the approach, because it can correct the WGA bias during data analysis. The WGA bias cannot be avoided during SNP array analysis and may affect the reliability of the array results. Moreover, this new approach could provide relatively higher resolution for chromosomal abnormality detection. For each test sample, around 10 million reads were obtained as tags using the MPS-based approach, whereas the maximum density for SNP arrays is five million using Illumina. Moreover, extremely high genome coverage (more than 95%) and depth (over 163) was also observed in mitochondrial DNA by MPS. Accordingly, the mitochondrial DNA can be evaluated in parallel, which may act as another marker for evaluation of embryo quality, with potential impact on developmental competence [21, 22]. Besides the accuracy, the costs of reagents, turnaround time, and throughput are also important factors when new technologies are suggested for application in clinical practice. In this study, the whole MPS-based procedure required 7–10 days in total and required cryopreservation of blastocysts to postpone transfer to the subsequent cycles. However, with the introduction of vitrification, pregnancy rates obtained with cryopreserved blastocysts after minimal or no stimulation may exceed those with fresh embryos [23, 24], and may also decrease the incidence of ectopic pregnancies [25]. Moreover, several studies indicated that the combined approach of biopsy at the blastocyst stage, comprehensive chromosome analysis, vitrification, and warming and transfer of a single normal embryo might dramatically improve the overall efficiency [9, 11, 26, 27]. On the other hand, the sustained technical improvement for the benchtop sequencing platforms, such as Ion Torrent and MiSeq, could generate modest sequencing data in just a few hours and may make fresh transfer possible in the near future.

MPS Simulation for Sensitivity/Specificity Evaluation of Chromosomal Abnormalities In order to determine the effective data size for detection of different types of chromosomal abnormalities by MPS, we selected all the abnormal samples to estimate the sensitivity/ specificity of this MPS-based approach with various sequencing sizes. We randomly selected 3 million, 2 million, 1 million, 0.5 million, and 0.1 million reads, and the specificity and sensitivity of aneuploidy and CNV detection with those reads were analyzed and illustrated in Figure 3. It was found that both the specificity and sensitivity of aneuploidy detection were 100% with all different sequencing sizes, demonstrating that using as few as 0.1 million sequencing data reads, theoretically all types of aneuploidy could be identified correctly. While for CNV (.16 M) detection, with 0.5 million sequencing data reads, the sensitivity and specificity could achieve 95% and 96%, respectively. DISCUSSION TE cell samples derived from a total of 38 blastocysts were examined by MPS, and results were compared with those

FIG. 3. The MPS simulation for specificity/sensitivity evaluation of CNVs and aneuploidy with different data sizes.

5

Article 69


FIG. 2. Examples of MPS and SNP array results of BLS15 from an IVF father with 46, XY, t(3;21)(q21;q22). A) Digital karyotyping of MPS result of BLS15. There is a 3 (q22-qter) deletion detected as a red region, and a þ21 (q22.12-qter) duplication detected as a green region. B) SNP array result of BLS15. There is a 3 (q22-qter) deletion detected as a shift of copy number in the region from 2 to 1 (loss), and a þ21 (q21-qter) duplication detected as a shift of copy number from 2 to 3 (gain).

YIN ET AL.

ACKNOWLEDGMENTS We sincerely thank our colleagues at the BGI-Shenzhen for sequencing. We thank Lotte Andreasen for excellent advice and revision of the manuscript.

REFERENCES 1. Cooper AR, Jungheim ES. Preimplantation genetic testing: indications and controversies. Clin Lab Med 2010; 30:519–531. 2. Harper JC, Wilton L, Traeger-Synodinos J, Goossens V, Moutou C, SenGupta SB, Pehlivan Budak T, Renwick P, De Rycke M, Geraedts JP, Harton G. The ESHRE PGD Consortium: 10 years of data collection. Hum Reprod Update 2012; 18:234–247. 3. Wang N, Zheng YM, Li L, Jin F. Preimplantation genetic screening: an effective testing for infertile and repeated miscarriage patients? Obstet Gynecol Int 2010; 2010:120–130. 4. Handyside AH, Kontogianni EH, Hardy K, Winston RM. Pregnancies from biopsied human preimplantation embryos sexed by Y-specific DNA amplification. Nature 1990; 344:768–770. 5. Handyside AH. Preimplantation genetic diagnosis after 20 years. Reprod Biomed Online 2010; 21:280–282. 6. Wilton L, Williamson R, McBain J, Edgar D, Voullaire L. Birth of a healthy infant after preimplantation confirmation of euploidy by comparative genomic hybridization. N Engl J Med 2001; 345:1537–1541. 7. Wells D, Escudero T, Levy B, Hirschhorn K, Delhanty JD, Munne S. First clinical application of comparative genomic hybridization and polar body testing for preimplantation genetic diagnosis of aneuploidy. Fertil Steril 2002; 78:543–549. 8. Wilton L, Voullaire L, Sargeant P, Williamson R, McBain J. Preimplantation aneuploidy screening using comparative genomic hybridization or fluorescence in situ hybridization of embryos from patients with recurrent implantation failure. Fertil Steril 2003; 80:860–868. 9. Schoolcraft WB, Fragouli E, Stevens J, Munne S, Katz-Jaffe MG, Wells D. Clinical application of comprehensive chromosomal screening at the blastocyst stage. Fertil Steril 2010; 94:1700–1706.

6

Article 69


10. Treff NR, Su J, Tao X, Levy B, Scott RT Jr. Accurate single cell 24 chromosome aneuploidy screening using whole genome amplification and single nucleotide polymorphism microarrays. Fertil Steril 2010; 94: 2017–2021. 11. Schoolcraft WB, Treff NR, Stevens JM, Ferry K, Katz-Jaffe M, Scott RT Jr. Live birth outcome with trophectoderm biopsy, blastocyst vitrification, and single-nucleotide polymorphism microarray-based comprehensive chromosome screening in infertile patients. Fertil Steril 2011; 96:638–640. 12. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, et al. Initial sequencing and analysis of the human genome. Nature 2001; 409: 860–921. 13. Lau TK, Chen F, Pan X, Pooh RK, Jiang F, Li Y, Jiang H, Li X, Chen S, Zhang X. Noninvasive prenatal diagnosis of common fetal chromosomal aneuploidies by maternal plasma DNA sequencing. J Matern Fetal Neonatal Med 2012; 25:1370–1374. 14. Wei X, Jin F, Ye Y, Xu C, Qu N, Ju X, Yi X. A novel mutation of IDS gene in a Chinese patient with mucopolysaccharidosis II by nextgeneration sequencing. Clin Chim Acta 2011; 412:2340–2342. 15. Navin N, Kendall J, Troge J, Andrews P, Rodgers L, McIndoo J, Cook K, Stepansky A, Levy D, Esposito D, Muthuswamy L, Krasnitz A, et al. Tumour evolution inferred by single-cell sequencing. Nature 2011; 472: 90–94. 16. Xu X, Hou Y, Yin X, Bao L, Tang A, Song L, Li F, Tsang S, Wu K, Wu H, He W, Zeng L, et al. Single-cell exome sequencing reveals singlenucleotide mutation characteristics of a kidney tumor. Cell 2012; 148: 886–895. 17. Hou Y, Song L, Zhu P, Zhang B, Tao Y, Xu X, Li F, Wu K, Liang J, Shao D, Wu H, Ye X, et al. Single-cell exome sequencing and monoclonal evolution of a JAK2-negative myeloproliferative neoplasm. Cell 2012; 148:873–885. 18. Li R, Yu C, Li Y, Lam TW, Yiu SM, Kristiansen K, Wang J. SOAP2: an improved ultrafast tool for short read alignment. Bioinformatics 2009; 25: 1966–1967. 19. Leary RJ, Cummins J, Wang TL, Velculescu VE. Digital karyotyping. Nat Protoc 2007; 2:1973–1986. 20. Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc 2008; 3:1101–1108. 21. Tsai HD, Hsieh YY, Hsieh JN, Chang CC, Yang CY, Yang JG, Cheng WL, Tsai FJ, Liu CS. Mitochondria DNA deletion and copy numbers of cumulus cells associated with in vitro fertilization outcomes. J Reprod Med 2010; 55:491–497. 22. Ge H, Tollner TL, Hu Z, Dai M, Li X, Guan H, Shan D, Zhang X, Lv J, Huang C, Dong Q. The importance of mitochondrial metabolic activity and mitochondrial DNA replication during oocyte maturation in vitro on oocyte quality and subsequent embryo developmental competence. Mol Reprod Dev 2012; 79:392–401. 23. Roque M, Lattes K, Serra S, Sola I, Geber S, Carreras R, Checa MA. Fresh embryo transfer versus frozen embryo transfer in in vitro fertilization cycles: a systematic review and meta-analysis. Fertil Steril 2013; 99: 156–162. 24. Shapiro BS, Daneshmand ST, Restrepo H, Garner FC, Aguirre M, Hudson C. Matched-cohort comparison of single-embryo transfers in fresh and frozen-thawed embryo transfer cycles. Fertil Steril 2013; 99:389–392. 25. Shapiro BS, Daneshmand ST, De Leon L, Garner FC, Aguirre M, Hudson C. Frozen-thawed embryo transfer is associated with a significantly reduced incidence of ectopic pregnancy. Fertil Steril 2012; 98:1490–1494. 26. Feng G, Zhang B, Zhou H, Shu J, Gan X, Wu F, Deng X. Comparable clinical outcomes and live births after single vitrified-warmed and fresh blastocyst transfer. Reprod Biomed Online 2012; 25:466–473. 27. Lathi RB, Massie JA, Gilani M, Milki AA, Westphal LM, Baker VL, Behr B. Outcomes of trophectoderm biopsy on cryopreserved blastocysts: a case series. Reprod Biomed Online 2012; 25:504–507. 28. Quail MA, Smith M, Coupland P, Otto TD, Harris SR, Connor TR, Bertoni A, Swerdlow HP, Gu Y. A tale of three next generation sequencing platforms: comparison of Ion Torrent, Pacific Biosciences and Illumina MiSeq sequencers. BMC Genomics 2012; 13:341. 29. Fragouli E, Alfarawati S, Daphnis DD, Goodall NN, Mania A, Griffiths T, Gordon A, Wells D. Cytogenetic analysis of human blastocysts with the use of FISH, CGH and aCGH: scientific data and technical evaluation. Hum Reprod 2011; 26:480–490. 30. Harper JC, Sengupta SB. Preimplantation genetic diagnosis: state of the art 2011. Hum Genet 2012; 131:175–186.

Regarding the costs, it take only about $41 to generate 1 Gb sequencing data using Illumina HiSeq 2000 with paired-end 100-bp sequencing strategy [28], which makes the MPS-based test comparable with conventional approaches. In our study, we obtained around 10 million sequencing data reads for each test sample. However, the sensitivity/specificity evaluation by MPS simulation has demonstrated that a minimum 500 and 100 K effective data size is enough for detection of unbalanced rearrangements and aneuploidy, respectively, both with high sensitivity and specificity. Accordingly, the estimated reagent cost for chromosomal abnormality detection is less than $100. Further development of sequencing technology still has a huge potential to reduce the cost to or below that of conventional approaches. The high throughput is another notable advantage of this MPS-based approach. By using the Illumina HiSeq2000 platform, more than 160 embryos could be processed in parallel in a labor-saving manner. This study demonstrated a promising application of MPS in PGD/PGS. Apart from the expected further improvement in efficiency and sensitivity of MPS, it can also be used for complex clinical abnormalities, such as mosaicism [29]. In the present study, TE biopsy was used to obtain samples for multiple cells [30]; however, MPS may also be useful for analysis of biopsied blastomeres and polar bodies. In summary, with the rapid development of sequencing technology and continuously decreasing cost and time of sequencing, it is foreseeable that MPS will play an increasingly significant role in clinical laboratories for human-assisted reproduction applications and studies. Our work was the first to demonstrate the possibilities of this application. Further studies with large sample sizes and in vivo outcomes are needed for comprehensive evaluation to outline the potential for routine clinical application.