The use of high-resolution SNP arrays to detect congenital cardiac defects

Abstract

Objective

Copy number variations (CNVs) detected by high-resolution single nucleotide polymorphism microarrays (SNP arrays) have been associated with congenital heart defects (CHDs). The genetic mechanism underlying the development of CHDs remains unclear.

Methods

High-resolution SNP arrays were used to detect CNVs and traditional chromosomal analyses, respectively, were carried out on 60 and 249 fetuses from gestational 12–37 weeks old, having isolated or complex CHDs that were diagnosed using prenatal ultrasound.

Results

Twenty of the 60 fetuses (33.5%) had abnormalities, of which 23 CNVs (12 pathogenic, five probable pathogenic and six of undetermined significance) were detected by SNP arrays, and two distinct CNVs were present in three of these fetuses. In addition, in 39 patients with isolated congenital heart disease who had normal karyotypes, abnormal CNVs were present in 28.2% (11/39), and in patients with complex coronary artery disease, 19.0% (4/21) had abnormal karyotypes and 42.9% (9/21) had abnormal CNVs. In patients with complex coronary artery disease, 19.0% (4/21) had abnormal karyotypes and 42.9% (9/21) had abnormal CNVs.

Conclusions

In conclusion, genome-wide high-resolution SNP array can improve the diagnostic rate and uncover additional pathogenic CNVs. The submicroscopic deletions and duplications of Online Mendelian Inheritance in Man (OMIM) genes found in this study have haploinsufficient (deletion) or triplosensitive (duplication) traits, which further clarify the etiology and inheritance of CHDs.

Keywords:

• Chromosomal microarray analysis
• congenital heart defects
• prenatal diagnosis
• copy number variation
• single nucleotide polymorphism microarray

Introduction

Congenital heart defects (CHDs) are the most common type of birth defect in the population, and are responsible for the majority of morbidity and mortality in children, with the incidence of CHDs varying from 4 to 50 cases per 1000 live births [1]. Several phenotype–genotype correlation studies have shown that specific morphogenetic mechanisms regulated by genes can lead to specific cardiac phenotypes [2]. However, due to the genetic and clinical heterogeneity of CHDs, it is difficult to precisely define specific genetic defects, as well as the discriminatory power of current clinical testing methods. For example, 44% of CHDs go undetected in the typical prenatal setting, despite the sensitivity of ultrasound technology and well-trained sonographers [3]. Therefore, genetic testing and ultrasound are now strongly recommended for fetuses with CHDs. Several methods are available for genetic detection of CHDs, including cytogenetic analysis, linkage and association studies, microarray analysis including single nucleotide polymorphism array (SNP array) and array comparative genomic hybridization (aCGH), CNV sequencing (CNV-Seq), and whole exome sequencing. However, few studies have investigated the use of high-resolution SNP arrays (Affymetrix CytoScan HD Array, Affymetrix, Santa Clara, CA) for the prenatal diagnosis of fetal CHDs.

Recent publications suggest that genomic copy number variations (CNVs) are involved in the etiology of CHDs [4]. CNVs are a phenomenon of variation formed by gene duplication, gene deletion, and other genomic rearrangements, with different number of genes duplicated or deleted in the genome varying between individuals [5]. The association of CNV abnormalities in patients with CHDs provided the opportunity to further delineate the etiology of syndromic and non-syndromic cardiac abnormalities [6]. However, in the absence of longitudinal studies, the mechanisms by which CNVs may contribute to this developmental defect are often limited. Furthermore, in our understanding of the natural history of these conditions, there is a major influence of the deterministic bias. Due to technical limitations, karyotyping cannot be used for (<10 Mb) assay microdeletion/microduplication, while CMA can detect microdeletion/microduplication as low as 50–100 kb. It is a high-resolution and high-throughput molecular analysis technology for scanning the whole genome, which can detect chromosome polyploids, aneuploid, CNVs, uniparental diploid, and mosaic (≥30%). In this study, to provide more evidence helpful to establishing phenotype–genotype correlation studies for CHD, high-resolution SNP arrays were used to detect CNVs in 60 CHD fetuses diagnosed by prenatal ultrasound, with comparison results with those obtained by traditional chromosome analysis, which indicated microdeletions/duplications involving Online Mendelian Inheritance in Man (OMIM) genes will further delineate the etiology and inheritance patterns of CHDs.

Materials and methods

Subjects

Two hundred and fifty-two fetal CHD samples were obtained between 12 and 37 weeks of gestation. The fetuses had isolated, or complex CHD diagnosed by two experienced ultrasounds specialists (Figure 1(A)). CHDs were classified according to the method detailed by Rocheleau et al. [7]. Out of the 252 fetuses, 28 pairs were twins (12 monochorionic diamniotic twin pairs, 15 dichorionic diamniotic twin pairs, and one monochorionic monoamniotic twin pair) along with 196 singleton fetuses. Several fetal specimens were obtained by the following methods: two chorionic villus sampling (CVS); 83 amniocenteses; and 167 cordocentesis. Parental blood samples were taken to exclude maternal contamination and assist in the interpretation of CNVs, if necessary. Among the 252 cases, four cases were lost to follow-up (follow-up rate 98.4%), 113 underwent termination of pregnancy, and 135 were live born. Further analyses for parental verification or WES were applied to analyze the relationship between pregnancy outcome and chromosomal abnormalities. All samples were tested and analyzed in the fetal medical center of the Affiliated Hospital of Sun Yat-Sen University. The study was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat-Sen University (Guangzhou, China). All couples obtained informed consent for invasive prenatal diagnosis.

Figure 1. (A) A flowchart for the research methodology. (B) Summary of CNVs analyzed by SNP array in 60 fetuses with CHDs. (C) Fifty-three cases with/without CNVs analyzed by SNP-array. (D) The types and frequencies of CHDs and the detection rates of CNVs. (E) del11q23.3q25. (F) del17p11.2. (G) del17p11.2. (H) Ventricular septal defect. (I) Tetralogy of fallot. (J) Hypoplastic right heart.

Karyotyping

Two hundred and forty-nine samples were successful in obtaining chromosomal analysis. Samples without karyotypes were attributed to a failure of cell culture. G-banded chromosomes were performed according to the standard procedures described by Wu et al. [8].

High-resolution SNP-array analysis

Whole genome SNP arrays were performed using commercially available CytoScan HD arrays (Affymetrix, Santa Clara, CA) according to the manufacturer’s protocol. DNA was isolated from samples of 60 CHD fetuses using the QIAamp DNA Mini Kit (Qiagen BeneluxBV, Venlo, The Netherlands). The Affymetrix CytoScan HD array contains more than 2.6 million copy number markers, of which 750,000 are SNP probes and 1.9 million are non-polymorphic probes. All analyses were performed using the Affymetrix Chromosome Suite (ChAS) software package. The Affymetrix Chromosome Suite (ChAS) software package was used for all analyses. CNV was detected by visual inspection of the normalized log2 intensity plot and numerical analysis of the log2 intensity ratio of Snps.

CNVs data analysis

The gene content of CNVs of the target genes is determined in the UCSC browser based on GRCH37. All observed copy number changes were compared with the genome variation database (DGV; http://projects.tcag.ca/variation/) and the UCSC genome browser (http://genome.ucsc.edu/). For a hypothetical candidate region containing at least one gene, each evaluation is included in DECIPHER (https://decipher.sanger.ac.uk/) and PubMed (http://www.ncbi.nlm.nih.gov/pubmed/).

The cutoff of detection criteria for CNV gain, loss, and regions of homozygosity (ROHs) was 200 kb, 100 kb, and 1000 kb, respectively. In addition, the criteria for the reporting of CNVs depended on the gene content and size. The breakpoint location for each anomaly region was converted to UCSC hg19 (UCSC Genome Browser, released February 2009). All CNVs were confirmed by real-time polymerase chain reaction. CNV is classified as the ACMG/ClinGen guide: (i) pathogenic (P), (ii) likely pathogenic (LP), (iii) variants of uncertain significance (VUS), (iv) likely benign (LB), and (v) benign.

Statistical analysis

The data are expressed as the mean ± standard deviation of SPSS 17.0 (SPSS, Inc., Chicago, IL). Chi-square test or t-test was used for statistical analysis. p < .05 means the difference is statistically significant.

Results

The distribution of CHD

The cohort included a total of 252 fetuses with CHD. The most common types of CHDs were atrioventricular septal abnormalities (53.1%), defects of the great vessels, and conotruncal heart defects (31.1%), followed by right heart abnormalities (6.9%) and left heart abnormalities (6.0%) (Figure 1(A)). The chromosomal analysis performed on 249 fetuses showed that 56 (22.5%) were abnormal and 48 were aneuploid (19.3%), and the most prominent karyotypes were trisomy 18 (8.0%) and trisomy 21 (7.6%) and six cases of chromosomal structural abnormalities (Table 1).

Results of SNP array among different types of CHDs detected by ultrasound

Ultrasound abnormalities include structural and ultrasound soft index abnormalities. Ultrasonic soft indexes include NT layer thickening, lateral ventricular dilatation, choroidal cyst, intracardiac punctate, strong echo, renal pelvic dilatation, intestinal strong echo, nasal bone loss, etc. Among the CHD abnormalities, we observed left heart abnormalities, right heart abnormalities, atrioventricular septal abnormalities, great vessels, and construal defects (Figure 1(D,H–J)). Among the 60 cases of CHDs, the detection rates of CNVs for left heart abnormalities, right heart abnormalities, atrioventricular septal abnormalities, great vessels, and construct defects were 15.8%, 13.6%, 11.2%, and 11.1%, respectively (Figure 1(B)). However, in 39 cases of isolated CHDs, 11 cases (P (6), LP (3), VUS (2)) (28.2%, 11/39) had abnormal CNVs (Figure 1(B)). In contrast, in 21 cases of complex CHD, nine cases (P (5), LP (1), VUS (3)) (42.9% (9/21)) had abnormal CNVs (Figure1(B)).

Table 1. Karyotypes and CMA results for six CHD fetuses with chromosomal structural abnormalities.

Case no.	Chromosome karyotype	CMA result	CNV size	Clinical information
16	46,XY,del(13)(qter→p10)	NA	NA	Ventricular septal defect, absence of nasal bone
21	46, XY, add(18)(q11.2→qter)	NA	NA	Ventricular septal defect, absence of left radius, absence of corpus callosum, the posterior cranial fossa increases, small cerebellum, left-diaphragmatic hernia, right strephenopodia
49	46,XX,del(11)(pter→q23:)	NA	NA	Hypoplastic left heart syndrome, premature closure of foramen ovale, ectopic drainage of umbilical vein
190	46,XY,del(13)(pter→q22)	arr13q31.1q34(84,045,912-114,981,524)×1	30 Mb	Holoprosencephaly, absence of nasal bone, ventricular septal defect
223	46, XX,del(4)(p16)	arr4p16.3p16.1(68,345-9,597,692)×1	9.5 Mb	Single atrium, long bones short, fetal growth restriction, polyhydramnios
237	46, XX,del(5)(p14.2)	arr5p15.3p14.2(113,576-24,599,883)×1	24 Mb	Ventricular septal defect, small cerebellum, thickened nuchal translucency

NA: not accessible; CNVs: copy number variations; CMA: chromosomal microarray analysis.

SNP array detection values among CHD fetuses

In order to identify the potential CHD-related CNVs, genomic CNV screening using SNP arrays were performed and 23 significant CNVs were detected. Of these, three cases had both abnormal CNVs, and an abnormal karyotype and 17 cases had no abnormal karyotype, involving more than one OMIM gene were found in 20 (33. 3%, 20/60) of the 60 CHD fetuses (Table 2, Figure 1(B)). Two CNVs were present in three CHD affected fetuses, and one of these genes was reported as haploinsufficient (deletion) or triplosensitive (duplication) (Table 3) (https://dosage.clinicalgenome.org/).

Table 2. Variants of CNVs in fetuses with CHD.

6 Variants of uncertain significance CNVs detected in 5 fetuses with CHDs
Case no.	Cardiac phenotype	Extra cardiac malformation	CNVs	Initiation or stop sites (NCBI37/hg19)	Size	Main genes involved in	Pregnancy outcome
196	VSD	—	dup16p11.2	32,524,764; 33,814,547	1.2Mb	TP53TG3D, TP53TG3	Live birth, surgery
206	HRH	—	dup 17q12 × 4	34,447,401; 34,822,466	375kb	TBC1D3B, CCL3L3, CCL3L1, CCL4L1, CCL4L2, TBC1D3C, TBC1D3H, TBC1D3G	Terminate pregnancy
215	SA, SV, TAPVR	Magenblase, gallbladder trans, asplenia (asplenia syndrome)	dup 18p11.31p11.23	7,076,087; 7,581,477	505kb	LAMA1, LRRC30, PTPRM	Terminate pregnancy
236	VSD	—	dup2p24.3 dup16p11.2 × 4	13,549,727; 14,325,998 31,975,035; 32,926,233	776 kb; 951kb	NACAP4 TP53TG3D, TP53TG3	Live birth, surgery
247	TOF	—	dup2p24.3	13,529,731; 14,360,751	831kb	LINC00276	Terminate pregnancy
5 Likely pathogenic CNVs detected in 4 fetuses with CHDs
194	VSD	—	dup4q32.3 × 4	167,019,071; 167,591,578	570kb	TLL1	Live birth
216	SA, DORV, HLH, TAPVR	Magenblase, gallbladder trans, asplenia (asplenia syndrome)	del14q11.2 [25%]	20,512,609; 22,347,693	1.8Mb	CHD8, SUPT16H, and other OMIM genes	Terminate pregnancy
218	VSD	—	dup1p36.33p36.32 [23%]	849,466; 4,456,277	3.6Mb	75 coding genes	Terminate pregnancy
232	VSD	Micrognathia, long bone dysplasias	dup4q35.2 del16p13.3	189,905,481; 190,957,473 85,880; 320,729	1.05Mb 235kb	FRG1, FRG2, and other OMIM genes HBA1, HBA2, NPRL3, and other OMIM genes	Terminate pregnancy
12 Pathogenic CNVs detected in 11 fetuses with CHDs
Number of fetuses
3	VSD, TGA, PS, DORV	22q11.2 microdeletion syndrome	del22q11.2	/	/	PRODH, DGCR2, CLTCL1, TBX1 and other 35 genes	Terminate pregnancy
2	VSD, AC, PS	Smith-Magenis syndrome	del17p11.2	/	/	PRODH, DGCR2, CLTCL1, TBX1 and other 35 genes	Terminate pregnancy
1	VSD	Cri du chat syndrome	del5p15.3p14.2	/	/	Containing more than one OMIM genes	Terminate pregnancy
1	VSD, HLH	11q terminal deletion syndrome	del11q23.3q25	/	/	Containing more than one OMIM gene	Terminate pregnancy
1	VSD	13q terminal deletion syndrome	del13q31.1q34	/	/	Containing more than one OMIM gene	Terminate pregnancy
1	VSD, DORV	16p terminal microduplication syndrome	dup16p13.3	/	/	LUC7L, RGS11, ARHGDIG, PDIA2, AXIN1, MRPL28, NME4, RAB11FIP3	Terminate pregnancy
1	SA	—	del4p16.3p16.1	/	/	Containing more than one OMIM gene	Terminate pregnancy
1	VSD	Cri du chat syndrome	delXp22.33p21.1 dup3q25.1q29	/	/	Containing more than one OMIM gene /	Terminate pregnancy

Abbreviations. CHDs, congenital heart defects; CNVs, copy number variations; VSD, ventricular septal defect; HRH, hypoplastic right heart; SA, single atrium; SV, single ventricle; TAPVR, total anomalous pulmonary venous connection; TOF, tetralogy of Fallot.

Table 3. Dosage sensitivity map in Clinical genome Database of CNVs with deletion-genes found in our study.

Dosage sensitivity map in Clinical genome Database of CNVs with deletion-genes
Gene	Gene type:	OMIM	ClinGen haploinsufficiency score	ClinGen triplosensitivity score	ExAC pLI score
FKBP6	Protein-coding	https://omim.org/entry/604839	Awaiting review	Awaiting review	0.131
ELN	Protein-coding	https://omim.org/entry/194050	3	0	NA
GTF2IRD1	Protein-coding	https://omim.org/entry/604318	0	0	0.996
GATA4	Protein-coding	https://omim.org/entry/600576	3	0	0.802
CRKL	Protein-coding	https://omim.org/entry/602007	1	0	0.45
TBX1	Protein-coding	https://omim.org/entry/602054	2	0	0.983
ATRX	Protein-coding	https://omim.org/entry/300032	3	0	1.0
GPC3	Protein-coding	https://omim.org/entry/300037	3	0	0.988
BCOR	Protein-coding	https://omim.org/entry/300485	3	0	1.0
ZIC3	Protein-coding	https://omim.org/entry/300265	Awaiting review	Awaiting review	0.823
FLNA	Protein-coding	https://omim.org/entry/300017	Awaiting review	Awaiting review	1.0
MID1	Protein-coding	https://omim.org/entry/300552	3	0	0.501
Dosage sensitivity map in Clinical genome Database of CNVs with duplication-genes
PRKAB2	Protein-coding	https://omim.org/entry/602741	Awaiting review	Awaiting review	0.767
FMO5	Protein-coding	https://omim.org/entry/603957	Awaiting review	Awaiting review	NA
CHD1L	Protein-coding	https://omim.org/entry/613039	0	0	NA
BCL9	Protein-coding	https://omim.org/entry/602597	Awaiting review	Awaiting review	0.832
ACP6	Protein-coding	https://omim.org/entry/611471	Awaiting review	Awaiting review	NA
GJA5	Protein-coding	https://omim.org/entry/121013	1	0	0.098
HRAS	Protein-coding	https://omim.org/entry/218040	0	0	0.008
GATA6	Protein-coding	https://omim.org/entry/601656	3	0	NA
RUNX1	Protein-coding	https://omim.org/entry/151385	3	0	0.449

The detailed information of the CNVs is listed as follows: (i) 12 pathogenic CNVs were 5p deletion syndrome, 11q deletion syndrome, 13q deletion syndrome, 16p duplication syndrome 17p11.2 deletion syndrome (two cases), 22q11.2 deletion syndrome (three cases), and so on (Table 2). (ii) Five likely pathogenic CNVs were deletions at 14q and 16p, 24 Kb to 1.8 Mb; duplication at 1p and 4q, 57 Kb to 3.6 Mb (Table 2). (iii) Six VUS CNVs from five fetuses were all duplication, distributing at 2p, 16p, 17q and 18p, 375 Kb to 1.2 Mb (Figure 1(E–G), Table 2). The cardiac phenotype and extra cardiac malformation were also listed.

Compared results of karyotype

CMA was performed in 60 cases, and karyotype analysis was not performed in three cases due to cell culture failure, so a total of 57 cases underwent both karyotyping and CMA. Of those, four cases had abnormal results for both tests: one was 45, X; the other cases were CNVs that had sizes of 9.5 Mb, 24 Mb, and 30 Mb. Among the remaining 53 cases with normal karyotypes, 15 cases (28.3%, 15/53) had abnormal CNVs, including six pathogenic (16p duplication syndrome, 22q11.2 deletion syndrome and Smith-Magenis syndrome) and four LP (Figure 1(C)). The above results demonstrate that there are certain false positive rates in karyotyping. SNP arrays can significantly improve the specificity and sensitivity of fetal CNVs detection.

Follow-ups of pregnancy outcome

In this study, 248 study participants were successfully followed. Among the cases with CNVs, 113 of them elected to terminate their pregnancy, 135 cases continued their pregnancy without obvious abnormalities observed after birth. Of the cases with VUS, three cases chose to terminate their pregnancy, two cases continued their pregnancy and resulted in live births (Table 2). The follow-up results showed that two cases exhibited ventricular septal defects requiring surgical intervention, and no cases exhibited language and motor developmental delay. There were 113 induced deliveries, of which 70 had karyotypic abnormalities or (and) abnormal CNVs (see Figure 1(A) – a redrawn figure supplemented with follow-up outcomes has been substituted for the original Figure 1(A)), and 43 were induced due to structural malformations of the fetus; none of the induced fetuses were pathologically investigated. No obvious abnormalities were observed in the remaining cases.

Discussion

In recent years, with increasing clinical evidence, SNP arrays have been commercialized in the clinical environment, which is convenient to use chromosome microarray analysis to diagnose unexplained CHDs [9]. The ability to identify CNVs in CHD cases was relatively high compared to the previously mentioned reports, suggesting that the higher density of the high-resolution array corresponds to a proportional increase in the number of detected CNVs of significant clinical significance [10].

CNVs in fetuses with structural abnormalities indicated by ultrasound can clarify the underlying genetic causes of fetal structural abnormalities and further improve the efficiency of pathogenic genetics detection. Among the 23 abnormal CNVs detected in 20 cases, we found that the frequency of copy number losses was approximately equal to that of copy number gains (12:11). This may be because most copy number changes come from non-allelic homologous recombination (NAHR). Theoretically, the incidence of gain and loss events generated by NAHR should be similar. Among these abnormal CNVs, 11 fetuses (18.3%) found well-described microdeletion or microduplication syndrome. The most significant finding was 22q11.2 deletion syndrome in three fetuses (5%), which was consistent with previous studies (6.4%) [11]. And the 23 CNVs overlapped with some OMIM genes with reported CHD phenotype, such as DGCR2, CLTCL1, and TBX1. In a previous study, the CNVs at 1q21.1 and 15q11.2 were strongly associated with the risk of sporadic non-syndromic CHD [12]. Although most clinical phenotypes of the fetuses in our research were isolated CHDs (39/60, 65%), we did not find these two CNVs. This incongruity may arise from the disparity of subjects. Meanwhile, the abnormal CNVs detection rate of complex CHDs (9/21, 42.9%; four cases with abnormal chromosomal karyotype) by SNP array was significantly more frequent than that of isolated CHDs (11/39, 28.2%; all had normal chromosomal karyotype). León et al. showed that KANSL1 and CRKL are genes covered by the common deletion regions in 22q11.2 microdeletion syndrome, and are part of the miRNA-mRNA network regulation module, which may play a role in modifying genes critical to the molecular mechanism of 22q11.2 DS patients [13]. This can be explained by the fact that patients with complex CHDs would be based on a genomic-related factor which generate a higher diagnostic value for the SNP array, suggesting that SNP arrays might be more suitable for the diagnosis of complex CHDs.

To identify the diagnostic efficiency of SNP arrays for CHDs, a comparison was made between chromosomal analysis and SNP arrays, which revealed that there were four cases with abnormal results for both types of diagnostic tests. However, 15 (28.3%) of the remaining 53 CHD cases with normal chromosomal karyotype had abnormal CNVs detected by the high-resolution SNP arrays. Compared to traditional karyotype analysis, SNP arrays have significantly higher resolution, because traditional karyotype analysis can only detect large chromosomal aberrations of more than 5–10 MB, but not small amplification and deletion regions. In comparison, SNP arrays can detect submicroscopic deletions and duplications [14]. Recent studies on the etiology of congenital heart disease in relation to patterns of chromosomal imbalance have shown that many genomic loci may be associated with normal heart development: some have very strong direct effects, while others have less direct effects [15].

VUS is a major challenge in clinical consultation; a comprehensive interpretation of pathogenicity in VUS patients is difficult due to low parental validation rates and incomplete databases. In our study, the discovered five variation of possible significance (VPS) CNVs and six VUS CNVs contained plenty of OMIM genes, which resulted in a variety of CHDs. Most of these genes are associated with cardiac development, including structure formation (e.g. collagen synthesis, sarcomere formation, and mitochondria) and cardiac function (e.g. ion channels and gap junctions), which indicate a precise, high rate of detection. Our results were supported by previous studies that found numerous genes in these CNVs, including ATRX, BCOR, CRKL, ELN, FLNA, FKBP6, GTF2IRD1, GATA4, TBX1, GPC3, MID1, and ZIC3 at the deletion site, and ACP6, BCL9, CHD1L, FMO5, GATA6, GJA5, PRKAB2, HRAS, and RUNX1 at the duplication site [16]. The information in Database (https://dosage.clinicalgenome.org/) for a dosage sensitivity map of these genes also shown that these gene were protein-coding genes, most of them were haploinsufficient (deletion) or triplosensitive (duplication) (Table 3) which may explain the genetic cause in these CHD cases.

In conclusion, the use of genome-wide high-resolution SNP arrays improves the diagnostic capabilities of CHDs, while revealing additional pathogenic CNVs. Submicroscopic deletions and duplications involving OMIM genes, protein-coding genes with haploinsufficient (deletion) or triploid (duplication) characteristics, will further delineate the etiology and inheritance of CHDs. High-resolution SNP arrays are a major advantage in the study of CHD, as they provide a clearer insight into the causes of CHD. These data are consistent with this rare allele being associated with coronary heart disease risk, but our study, apart from the limited number of cases, does not have the ability to formally address or measure the possible association with CHD risk. Therefore, it is unclear whether additional genes disrupted by rare deletions in cases play a causal role in CHD risk. Limitations of this retrospective study include the small sample size due to the acquisition of detailed genotype–phenotype information of some fetuses. In addition, differences in detection rates may be related to patient selection bias, number of cases, and exclusion of cases with abnormal karyotypes. Therefore, larger studies are needed to further investigate these possible associations and to understand the role of CNVs in the development of CHDs.

Author contributions

Huang Linhuan and Xie Yingjun carried out study design; Cai Danlei, He Zhiming, Luo Yanmin, and Kong Shu performed the experiments; Chen Jiayi, Peng Jiayi, Su Chuqi, and Yang Yinghong prepared Tables 1–3. Huang Linhuan and Xie Yingjun wrote the paper. All authors read and approved the final manuscript.

Ethical approval

Ethical approval was obtained for this study from the First Affiliated Hospital of Sun Yat-sen University.

Consent form

Not applicable.

Acknowledgements

The authors would like to thank the Clinical Cytogenetics laboratory for helping the collection of data presented here. We are also grateful to the individuals included in this study as well as their families. We would like to express our sincere gratitude to Professor Richard H. Finnell for his invaluable contributions to the refinement of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Funding

This study was supported by the National Natural Science Foundation of China (No. 81671464), Guangdong Municipal Department of Science and Technology, Municipal Schools (Institutes) Jointly Funded Project, China (No. 2023A03J0386), Guangzhou Medical University, First-class Professional Construction Project in 2022-Enhancement of Undergraduates’ Scientific Research and Innovation Ability Project (No. 02-408-2203-2059), and Key R&D Program of Zhejiang Province of China (No. 2021C03030).