CMA is a high-throughput microarray-based technology. Through specifically designed probes that hybridize with sample DNA, CMA detects various genomic variations including CNVs, single nucleotide polymorphisms (SNPs), loss of heterozygosity and uniparental disomy through fluorescence signal intensity analysis (18). At present, CMA is recommended as the effective genomic test for fetuses with SBDs in clinical practice (19,20).

In fetuses with structurally abnormal findings detected by ultrasound, CMA can effectively enable high-resolution genomic screening to identify chromosomal abnormalities undetectable by conventional karyotyping, including clinically relevant microdeletions and microduplications (21-23). Besides, for fetuses with ultrasound-detected multi-system structural abnormalities, CMA demonstrates a notably higher detection rate of 41.2% for chromosomal abnormalities, markedly surpassing the rates observed in isolated system anomalies (such as 25.4% for cardiovascular system abnormalities and 18.6% for central nervous system abnormalities) (24). Furthermore, CMA offers particular diagnostic value by identifying ~1.2% of clinically relevant abnormalities in low-risk pregnancies with normal NIPT results (25). Furthermore, CMA technology is relatively mature and easy to operate, making it conducive to promoting prenatal genetic testing in regions with limited medical resources (26).

However, CMA has inherent technical limitations in clinical practice. The methodology carries detection risks for variants <50 kb and cannot identify balanced translocations or low-level mosaicism (26). Furthermore, CMA lacks diagnostic capability for balanced chromosomal rearrangements such as inversions and translocations (27).

CNV-seq is an NGS technique that utilizes low-coverage WGS (0.1-1x depth) to analyze fetal DNA for chromosomal abnormalities and CNVs (28). With the development of NGS technology, CNV-seq is emerging as a viable alternative to CMA for the prenatal diagnosis (29).

CNV-seq demonstrates superior detection efficiency for both CNVs and chromosomal abnormalities compared with NIPT in SBDs (30,31). For fetuses with ultrasound-detected abnormalities, CNV-seq exhibits a 2-4% higher sensitivity compared with karyotyping in identifying chromosomal aberrations (32,33). Moreover, CNV-seq provides superior resolution compared with CMA, allowing for precise detection of microdeletions, microduplications and low-level mosaicism (<20%) (34,35). Notably, CNV-seq not only detects all chromosomal aneuploidies and large-scale genomic rearrangements identifiable by CMA but also uncovers an additional 34.88% of clinically relevant variants (34). In clinical practice, due to its faster detection speed and smaller sample requirement, CNV-seq can serve as a preliminary rapid test for early clinical intervention. Meanwhile, its high compatibility with other NGS platforms greatly simplifies testing workflows and reduces costs, making the combination of CNV-seq with other NGS technologies a preferred strategy in genetic testing for SBDs (36).

However, in clinical testing, CNV-seq cannot detect balanced chromosomal translocations or inversions and has restricted diagnostic utility in identifying polyploidy and uniparental disomy associated with syndromic SBDs (37). Furthermore, CNV-seq exhibits reduced sensitivity and undetermined specificity in CNV detection when benchmarked against CMA (30,38).

Targeted sequencing represents a precision detection methodology that employs NGS technology. This hypothesis-driven approach utilizes customized probe or primer panels to selectively capture and enrich predefined genomic regions for high-throughput analysis, focusing primarily on functionally characterized regions with established or putative disease associations (39). Through target enrichment, targeted sequencing achieves notably enhance sequencing depth, enabling sensitive detection of rare pathogenic SNVs/indels and CNVs in known disease-related genes among fetuses with SBDs (40,41). Therefore, in clinical practice, specific genes or gene combinations can be selected based on fetal phenotypes to perform targeted deep sequencing analysis and diagnosis, while avoiding the blind use of WES or WGS (42).

In prenatal diagnosis, targeted sequencing serves as a supplementary diagnostic tool for CMA-negative cases and is particularly advantageous for fetuses with specific phenotypes, detecting ~13.6% of pathogenic variants in those with SBDs (42,43). Besides, the technology demonstrates particular utility in diagnosing genetically heterogeneous SBDs, allowing simultaneous analysis of multiple candidate genes (44). Targeted sequencing offers a cost-effective diagnostic solution for SBDs, with core advantages of lower cost and short turnaround time. When prenatal examinations suggest a fetus may have specific types of structural abnormalities, this method can quickly confirm relevant genetic subtypes and provide key evidence for genetic counseling and clinical decision-making (41,45,46).

However, targeted sequencing has notable limitations inherent to its design, which limits its application in certain clinical scenarios; its diagnostic scope remains constrained by current genomic knowledge, potentially missing novel pathogenic genes or functionally relevant non-coding variants beyond established disease associations (47). Furthermore, its capacity for systematic investigation of complex structural variations is intrinsically limited by restricted genomic coverage, which impedes comprehensive genome-wide analyses (48).

WES employs an NGS technology coupled with probe-based hybridization capture systems to specifically target and enrich exonic regions and adjacent splice sites (±20 bp). Although these regions constitute merely 1-2% of the human genome, they harbor ~85% of known pathogenic variants. WES of these regions enables the identification of clinical relevance for the majority of SNVs/indels and CNVs in the human genome (49). Clinically, trio-WES is the most effective exome testing method, as it enhances diagnostic rates, elucidates variant origins and permits accurate recurrence risk assessment and early prenatal diagnosis (50).

In SBDs, WES serves as a supplementary diagnostic tool following inconclusive initial tests (such as CMA/CNV-seq) or in special cases such as recurrent or lethal fetal abnormalities, offering comprehensive exonic coverage that enhances the detection of pathogenic variants compared with these methods (50,51). For fetuses with isolated SBDs (such as CHDs or NTDs), WES offers an incremental diagnostic yield of 10-50%, increasing to ~33% in cases with multi-system SBDs (52-55). This approach is particularly advantageous for disorders exhibiting high genetic and phenotypic heterogeneity. Moreover, in fetuses with ultrasound-detected cardiac or central nervous system abnormalities but negative CMA and karyotyping results, WES effectively identifies recessive or de novo dominant pathogenic mutations undetectable by conventional methods, demonstrating superior diagnostic sensitivity (56,57).

However, as WES relies on short-read sequencing platforms, it may lead to inaccurate alignment in genomic regions with high sequence homology, such as repetitive elements or pseudogenes, potentially generating false-positive variant calls (58,59). Additionally, emerging evidence implicates that non-coding variations (including deep intronic variants, long non-coding RNA regulatory elements and mitochondrial DNA structural alterations) are also involved in the occurrence and development of SBDs, while WES is mainly restricted to exonic regions and adjacent splice sites, exhibiting markedly reduced sensitivity for detecting these non-coding variants (60).

WGS shares technical principles with WES but extends beyond WES's limitation to exonic regions. By enabling systematic identification of SNVs/indels and CNVs across the entire genome, WGS provides comprehensive genome-wide detection (49).

In fetuses with SBDs, WGS enables detection of subtle structural variations (such as <10 kb) and yields more precise genetic characterization (61,62). Beyond structural variants, WGS identifies regulatory mechanisms underlying gene expression dysregulation, including promoter/enhancer variants and deep intronic splice-altering mutations. Notably, WGS precisely maps complex genomic rearrangements such as intronic balanced translocation breakpoints that disrupt normal mRNA splicing (63,64). For fetuses exhibiting SBDs, WGS demonstrates superior diagnostic capability by identifying both variants detectable through CMA and WES, along with additional pathogenic variants (65). This comprehensive approach enhances diagnostic yield by 10-20% compared with conventional methods. Currently, WGS offers the most comprehensive diagnostic performance for SBDs, making it a potential 'one-stop' test that reduces the need for repeated testing (66).

Despite its superior sensitivity for structural variation detection, WGS presents technical challenges. First, genome-wide coverage generates extensive datasets with typically low per-base sequencing depth, increasing the risk of false-positive variant interpretation due to reduced confidence in mutation calling (67,68). Second, the substantial costs associated with per-sample sequencing and bioinformatic analysis create economic barriers to implementation WGS in routine prenatal screening programs (69). Thus, the application of WGS in prenatal diagnosis is restricted to supplementary roles in select cases such as fetuses with organ system abnormalities, a history of recurrent adverse pregnancies and negative CMA and WES results. However, with advancing sequencing technology and falling costs, WGS is expected to become a key tool in prenatal genetic diagnosis (70).

CNVs represent a class of genomic structural variations involving DNA segment duplications or deletions typically >50 kb. These variations can influence gene expression through dosage effects, gene structure disruption or regulatory element interference (71,72). To uncover such candidate genes related to SBDs, comprehensive bioinformatic analysis of CNV regions and their functionally associated gene networks provides a crucial approach (73,74) (Table II).

CNVs contribute to 10-15% of CHD cases. Recurrent CNVs in specific genomic regions, such as 1q21.1, 22q11.2 and 16p11.2, are frequently observed in patients with a CHD (75). These regions harbor dosage-sensitive genes critical for cardiac development, underscoring their role as a major genetic etiology of CHDs. TBX1 haploinsufficiency caused by 22q11.2 microdeletions represents one of the well-characterized core genetic mechanisms (76). A WES study by Zhao et al (77) further revealed that chromatin regulatory genes within the 22q11.2 region, including EP400, KAT6A, KMT2C, KMT2D, NST1, CHD7 and PHF21A, exhibit markedly higher mutation frequencies in patients with a CHD compared with controls, which may contribute to abnormal cardiac development through epigenetic regulatory mechanism. Additionally, a case-control study using CMA identified SLC2A3 duplications in 5.0% (18/346) of 22q11.2 deletion syndrome patients with CHDs and aortic arch abnormalities, a frequency markedly higher than in deletion-positive individuals with normal cardiac anatomy. This finding implicates SLC2A3 duplication as a potential genetic modifier of cardiac phenotypes in 22q11.2 deletion syndrome (78). CNV studies of other pathogenic regions have revealed additional candidate genes. For instance, Lin et al (79) performed CMA on a cohort of 1,118 Chinese fetuses with CHDs, identifying a notable association between proximal 16p11.2 deletions (BP4-BP5 region) and CHDs; moreover, this association may be mediated through TBX6 gene dysfunction. Another investigation of 78 non-22q11.2 deletion patients identified 15q21.1 and 2p22.3 duplications as novel pathogenic CNVs, with multiple implicated genes (IRX4, BMPR1A, SORBS2, ID2, ROCK2, E2F6, GATA4, SOX7, SEMAD6D, FBN1 and LTBP1) known to participate in cardiac development (80). Mak et al (81) performed a CMA analysis using a large control cohort comprising 3,987 Caucasian and 1,945 Singaporean Chinese subjects. The study identified 10 large rare CNVs, with further analysis revealing that nucleoredoxin in the 17p13.3 region, COL4A1 and COL4A2 in 13q33.3, and ZEB2 in the 2q22.3 region were strongly associated with CHDs. Dasouki et al (82) conducted CNV-seq analysis in 134 Saudi Arabian patients with CHDs, detecting 21 copy number gains and 11 losses. The genomic variants were primarily clustered in chromosomal regions 17q21.31, 8p11.21, 22q11.23 and 16p11.2. Functional and network analyses identified NPHP1, PLCB1, KANSL1 and NR3C1 as potential candidate genes associated with CHD pathogenesis.

Genetic investigations of OFCs have demonstrated the crucial role of CNVs and their associated genes in disease pathogenesis. Lansdon et al (83) performed CMA on cohorts from the Philippines (n=869) and Europe (n=233) with non-syndromic cleft lip/palate (nsCL/P), identifying recurrent 2q31.1 duplications (11 cases), 22q11.2 CNVs (4 cases) and 3q29 deletions (2 cases). Further analyses implicated COBLL1, RIC1 and ARHGEF38 as candidate genes, all involved in Rho/Rab GTPase signaling pathways associated with OFCs and craniofacial anomalies. Additionally, a separate multi-center study involving 270 orofacial clefts cases identified a rare duplication variant at 19p13.12, suggesting its potential as a pathogenic locus. The variant affects SYDE1, BRD4 and AKAP8, which exhibit spatiotemporally specific expression patterns in the branchial arches and frontonasal processes of E10.5 mouse embryos. These findings indicate that this gene cluster may represent a novel susceptibility factor for nsCL/P (84). Furthermore, in a CMA analysis of 467 OFC trios (comprising 1,375 subjects) and 391 control trios (902 subjects), Younkin et al (85) identified a genome-wide significant 62 kb non-coding region at 7p14.1 and proposed TARP as a candidate gene for the early prenatal diagnosis of OFCs, suggesting a potential role of T-cell receptors in OFC pathogenesis. A further study on this cohort identified two deletions: A 67 kb deletion in MGAM on chromosome 7q34 and a 206 kb deletion spanning ADAM3A and ADAM5 in the 8p11 region. The frequency of these deletions was markedly higher in affected families compared with in unaffected controls, further supporting their potential role as candidate genes (86).

CNV analysis has been used to identify associations among common and rare genetic variants and NTD risks and has identified several promising candidate genes. A WES study analyzed patients with meningomyelocele and identified 6 cases harboring a 22q11.2 deletion, which included an LCR22C-D deletion within the low-copy repeat region. Furthermore, the study pinpointed AIFM3, CRKL and PI4KA as key candidate genes within this locus (87). Additionally, a WGS-based analysis involving cohorts from the United States and Qatar (comprising 140 patients with spina bifida and 183 controls) expanded the gene network associated with NTD-related CNVs. Rare coding CNVs were identified in pathways critical to NTD pathogenesis, including cell cycle regulation (SH3GL3 and PARD3), mitochondrial metabolism (DMGDH and FOXRED1), transmembrane transport (SLC44A2 and SLC44A3) and signal transduction (VAV2 and DOCK10) (88). Tian et al (89) performed CNV analysis of all exonic regions in planar cell polarity (PCP) pathway-related genes (VANGL1, VANGL2, CELSR1, SCRIB, DVL2, DVL3 and PTK7) in 11 NTD probands, identifying 16 CNVs. The CNVs were predominantly located in DVL2, VANGL1 and VANGL2, suggesting these genes as strong candidates for NTD pathogenesis. These findings not only delineate the multi-pathway pathogenic framework of NTDs but also provide novel insights into the genetic association between NTDs and CNVs.

In summary, CNVs are notably associated with the incidence of SBDs. CNV analysis can elucidate the genetic effects of large-fragment deletion mutations, holding substantial relevance for identifying key pathogenic genes and exploring potential gene dosage-phenotype associations.

Small variants typically encompass SNVs, short insertions and short deletions (<50 bp) (16). These variants can impact the function of critical genes, thereby contributing to the development of SBDs (90,91). The detection and analysis of SNVs/indels have notably enhanced the understanding of these subtle genetic alterations in SBDs and have effectively expanded the candidate gene spectrum for these conditions (Table III).

Through SNV/indel analyses have served a key role in unraveling the complex genetic heterogeneity characteristic of CHDs. In a landmark study, Sevim et al (92) performed a systematic WES analysis on 559 families with CHDs, identifying 23 novel candidate pathogenic genes. Notably, HSP90AA1, IQGAP1 and TJP2 showed strong associations with isolated CHDs, whereas ROCK2, APBB1, KDM5A and CHD4 were primarily associated with non-isolated CHDs. Among these, HSP90AA1, ROCK2, IQGAP1 and CHD4 exhibited the highest relevance in pathogenicity prediction models, underscoring their central roles in disease pathogenesis. An additional WES study involving 52 Qatari families (178 individuals) identified four pathogenic or likely pathogenic variants: ROBO1, SLC2A10, SMAD6 and CHD7. Notably, ROBO1 was found to functionally interact with classical CHD-related genes such as TBX1, TBX5, NOTCH1 and NKX2-5 in interaction networks. Mechanistic studies using a chick heart development model demonstrated that SMAD6 contributes to CHD pathogenesis through synergistic regulation with NKX2-5 (93). Substantial progress has also been made in subtype-specific CHD research. Shi et al (94) analyzed WES data from 100 patients with pulmonary artery atresia and 100 healthy controls, integrating network analysis and gene expression validation to establish associations between DNAH10, DST, FAT1, HMCN1, HNRNPC, TEP1 and TYK2 with the pathological mechanisms of pulmonary artery atresia for the first time. In the research of total anomalous pulmonary venous connection, WES analysis of 78 sporadic cases and 100 controls identified seven candidate genes: CLTCL1, CST3, GXYLT1, HMGA2, SNAI1, VAV2 and ZDHHC8. Functional verification and zebrafish model interaction network analysis further validated SNAI1, HMGA2 and VAV2 as key candidate genes driving total anomalous pulmonary venous connection pathogenesis (95).

By examining protein-altering variations such as missense and frameshift mutations, SNV/indel analysis has revealed how coding sequence modifications induce functional changes (loss/gain-of-function), thereby offering novel perspectives on OFC candidate genes. WES of 84 affected individuals from 46 families with nsCL/P identified rare deleterious variants in TP63, TBX1, LRP6 and GRHL3, genes previously associated with syndromic forms (96). Mangold et al (24) performed targeted sequencing in a cohort of 576 European patients with nsCL/P and 96 non-syndromic cleft palate only cases, identifying four novel truncating GRHL3 mutations. Notably, all nine mutation carriers exclusively presented with non-syndromic cleft palate only, providing compelling evidence for the notable pathogenic contribution of GRHL3 to non-syndromic cleft palate only. A subsequent WES study of 132 non-syndromic cleft palate only cases and 623 multiethnic controls identified three novel candidate genes (ACACB, PTPRS and MIB1), along with de novo variants in GRHL3 and CREBBP, genes associated with Van der Woude syndrome and Rubinstein-Taybi syndrome, respectively (97). A trio-WES analysis of 130 African nsCL/P families identified pathogenic AFDN missense mutations (p.E485K and p.G1703R) that may disrupt connexin binding, potentially impairing cell polarity regulation and mesenchymal migration during facial morphogenesis (98). A WES study from Eastern Chinese cohorts demonstrated LAMA5 mutations in nsCL/P cases, with further investigations revealing the crucial role of LAMA5 in palatal development, establishing it as a strong candidate gene (99). Kumari et al (100) identified variants in TGFβ3, MSX1 and MMP3 through WES analysis of 245 Indian nsCL/P cases and validated these genes as strong candidates for OFCs.

The application of SNV/indel analysis has facilitated the identification of candidate genes within the underlying pathways of NTDs. WES performed on 23 myelomeningocele cases in France revealed that de novo variants in crucial genes of the vitamin B12 metabolism pathway (LRP2, MMAA and TCN2), genes related to folate metabolism (FPGS), the core regulator of choline metabolism (BHMT) and GLI3, a key regulatory element of the Sonic Hedgehog signaling pathway, were notably linked to genetic susceptibility to NTDs (101). A WGS study involving 140 isolated spina bifida samples further identified eight rare missense variants in the CIC gene, which were validated through functional analysis. These CIC missense variants found in NTD cases notably suppressed folate receptor 1 protein expression levels and disrupted the transduction of the PCP signaling pathway (102). Additionally, analysis of trio-WES in 43 cases of sporadic myelomeningocele/anencephaly and their parents identified seven core candidate genes: SHROOM3, PAX3, GRHL3, PTPRS, WBSCR28, MFAP1 and DDX3X. Among them, SHROOM3 is noted for its high degree of evolutionary conservation, the knockout of its double isoform resulted in 100% incidence of exencephaly and a 23% penetrance rate of spina bifida phenotypes in mouse embryos. As a key regulator of neural crest differentiation, rare variants c.218C>A in PAX3 can potentially lead to NTD phenotypes by disrupting neural crest cell migration during neural tube closure (103). Building on this understanding, Chen et al (104) employed targeted sequencing technology to perform a comprehensive analysis of the SHROOM gene family in a cohort of 343 individuals with NTDs and 206 control subjects in China. The findings revealed a markedly higher frequency of damaging missense variants in the SHROOM2 gene among the case group compared with the control group. These mutations lead to the failure of neural tube closure by disrupting the interaction between SHROOM2 and ROCK1 proteins, thereby interfering with the cytoskeletal remodeling mediated by the planar cell polarity pathway.

SNV/indel analysis systematically investigates subtle genetic alterations overlooked in CNV analysis, precisely mapping associated loci of SBDs and revealing the impact of nucleotide-level modifications on gene function. This approach notably establishes a crucial molecular foundation for optimizing precision prevention strategies of SBDs and exploring molecular pathways for targeted therapies.

The genetic basis of most SBDs stems from the cumulative effects of multiple minor-effect genes rather than single-gene mutations (105). Consequently, identifying candidate genes involved in polygenic synergistic pathogenesis is crucial for elucidating the genetic factors underlying these complex disorders. GWAS analysis represent a pivotal approach in contemporary genomics research, employing SNP arrays or NGS technologies to systematically examine genome-wide genetic variations. Through multivariate data analysis, GWAS analysis elucidates associations between these variations and phenotypic traits, thereby revealing the genetic architecture of complex traits (106,107). Advances in GWAS applications for SBDs have led to the discovery of numerous potential candidate genes, representing a notable breakthrough in understanding the molecular pathogenesis of these conditions (Table IV).

GWAS analysis has provided crucial insights into the complex genetic architecture of CHDs. A large-scale GWAS involving 40,000 UK Biobank samples identified 130 loci notably associated with right heart phenotypes, a number of which are located near known regulatory genes such as NKX2-4, TBX5/TBX3, WNT9B and GATA4. The pivotal role of these genes in cardiac morphogenesis has been further substantiated (108). Additionally, a single-trait GWAS analysis of 29,506 individuals with right ventricular structural abnormalities revealed 12 candidate genes shared across different right ventricular phenotypes, including TTN, ATXN2, PTPN11 and ACTN4. Notably, FHOD3, MYH6, MYL4 and TMEM43 exhibit functional overlap with Mendelian cardiomyopathy genes, with their encoded proteins primarily involved in myocardial contraction and cell adhesion processes (109). Lahm et al (110) identified MACROD2, GOSR2, WNT3 and MSX1 as critical genes in embryonic cardiac development, which were significantly associated with SNPs based on CMA and GWAS analyses of 4,034 Caucasian patients with CHDs and 8,486 healthy controls. Jin et al (111) expanded this research by performing WES combined with GWAS on 2,871 CHD probands. The findings demonstrated that MYH6 mutations were notably associated with an elevated risk of ventricular dysfunction, whereas dominant mutations in FLT4 were specifically linked to tetralogy of Fallot. Furthermore, rare inherited and de novo heterozygous variants were markedly enriched in genes such as CHD7, KMT2D, PTPN11, RBFOX2, FLT4, SMAD6 and NOTCH1 among patients with complex CHDs, highlighting novel candidate targets for further investigation.

GWAS analysis has identified multiple candidate genes potentially involved in OFCs. A GWAS of a Maya population (149 patients with nsCL/P and 303 controls) revealed notable associations between nsCL/P and genetic variants in IRF6, as well as loci at 8q24, 10q25 and 17q22. Single-marker analysis further confirmed these associations, particularly for IRF6, 8q24 and 10q25 (112). A pleiotropy-informed conditional false discovery rate analysis, incorporating 814 nsCL/P cases, 205 nsCPO cases and 2,159 controls from an African population GWAS, along with mouse facial development expression data, identified 19 potential candidate genes. Among these, MDN1, MAP3K7, KMT2A and ARCN1 emerged as top candidates, supported by evidence from mouse models or individuals harboring pathogenic variants in these genes (113). In a separate GWAS of oral clefts in a Polish cohort, eight single-nucleotide polymorphisms in PAX7 were strongly associated with nsCL/P, with the etiological role of the gene independently validated in a cohort of 247 patients and 445 controls (114). These findings not only expand the repertoire of candidate genes implicated in OFCs but also provide a molecular framework for understanding how genetic background influences craniofacial development across diverse populations. The application of GWAS has notably advanced the identification of candidate genes for SBDs, with large sample sizes enhancing result accuracy.

In genetic investigations of NTDs, a GWAS performed by Tindula et al (115) involving 321 Bangladeshi NTD cases and 177 controls failed to identify genome-wide significant variants. This negative finding may reflect insufficient statistical power due to the limited sample size. Notably, such challenges are pervasive in SBD research: Constrained by clinical sample availability, certain malformation subtypes lack large-scale GWAS implementations, while existing studies often fail to detect biologically meaningful susceptibility genes due to inadequate cohort sizes.

GWAS analyses enhance the understanding of SBDs with polygenic inheritance patterns by analyzing genome-wide associations between genetic variants and these defects while quantifying the cumulative effects of multiple variants. Furthermore, GWAS analysis provides a comprehensive assessment of gene-gene interactions throughout the genome, filling the research gap in the study of polygenic effects in CNV analysis or SNV/indel analysis, and greatly expanding the potential candidate gene pool associated with SBDs.