Introduction

Short tandem repeats (STRs) are widely distributed throughout the human genome, highly polymorphic, and easy to genotype using the combined fluorescence multiplex PCR and capillary electrophoresis technique (also known as QF-PCR). Therefore, STRs are widely used in forensic individual identification, paternity testing, and population genetics research fields.^1–5 The STRs genotyping method based on QF-PCR is also used for quality control tasks such as sample identification, determination of the origin of fetal DNA, and detection of maternal cell contamination in prenatal diagnostics.^6,7 These quality control processes are necessary prerequisites to ensure the accuracy of prenatal diagnosis results. Additionally, based on the genotype and peak pattern of STRs, the molecular karyotype of the chromosome can be conveniently indicated.^8,9 However, when STRs are used for testing the molecular karyotype of target chromosomes or quality control of prenatal samples, the test performance is closely related to key system performance parameters, such as population allele frequencies and random match probability. In our routine sample quality control process for prenatal diagnosis, we utilized a multiplex genotyping system including 20 STRs located on chr13, chr18, and chr21 to test 2333 mother-fetus sample pairs. Based on the genotype dataset, we comprehensively evaluated the system performance of the 20 autosomal STRs in sample discrimination and kinship determination of the mother-fetus pair, which are critical aspects of sample quality control in prenatal diagnosis.

This is the first study to evaluate the application values of a genotyping system in prenatal diagnosis from a population genetics perspective among East Chinese Han population. The large-scale sample data significantly improved reliability in detecting maternal cell contamination and increased diagnostic confidence in rapid aneuploidy detection. Although the genotyping system currently covers only three autosomes (chr13, chr18, and chr21), it has demonstrated complete suitability for detecting common aneuploidies in prenatal diagnosis among East Chinese Han population.

Materials and Methods

Sample and Informed Consent

The study cohort comprised 2333 Chinese Han pregnant women who underwent prenatal amniocentesis or chorionic villus sampling at Ningbo Women and Children’s Hospital between November 2021 and April 2023. All subjects met the clinical indications required for prenatal diagnosis, peripheral blood samples were collected from the pregnant women using EDTA as an anticoagulant. All samples underwent genotyping analysis for quality control, followed by aneuploidy diagnosis for the prenatal samples. All participants provided informed consent for genetic analysis and future data use. The protocol was approved by the Institutional Review Board of The Affiliated Women and Children’s Hospital of Ningbo University (Approval No. EC2020-048). The study complies with the Declaration of Helsinki.

DNA Extraction

We collected amniotic fluid or chorionic villi from each pregnant women using the standard amniocentesis technique for QF-PCR. Following centrifugation, fetal cells were isolated. Genomic DNA was extracted from peripheral blood samples collected with EDTA anticoagulant using the QIAGEN QIAamp DNA Mini Kit (QIAGEN, Germany). The extracted genomic DNA samples underwent quality testing using the NanoDrop 2000 (Thermo Fisher Scientific, USA). DNA samples that passed the quality assessment were stored at −20°C for subsequent genotyping analysis.

Multiplex PCR Amplification

Multiplex PCR amplification was performed using the AneuFiler^TM kit (Approved by National Medical Products Administration of China with No.20233401830; Shanghai Cubicise Medical Co., Ltd., China), which was a clinical diagnosis assay for common aneuploidies of chr13, chr18, chr21 and sex chromosomes based QF-PCR. The PCR amplification system was prepared according to the product manual, and amplification was carried out using the Veriti^TM PCR System (Thermo Fisher Scientific, USA) under the provided PCR conditions. The multiplex system is a single-tube 30 plex PCR amplification system, including 6 STRs on chr13(D13S1822, D13S317, D13S892, D13S788, D13S1817, D13S325), 6 STRs on chr18(D18S862, D18S878, D18S391, D18S867, D18S865, D18S1371), 8 STRs on chr21(D21S1437, D21S2052, D21S1446, D21S1444, D21S1432, D21S1414, D21S1411, PentD), 1 sex locus (ZFXY) on chrX and chrY, 5 loci (IDXp, IDXq, DXS7132, DXS9902, GATA165B12) on chrX, 1 locus on chrY(SRY), 1 quality control marker (TAF9b) on chr3 and chrX, 1 quality control marker (CM11X) on chr11 and chrX, and 1 quality control marker (CM01Y) on chr1 and chrY. The complete marker specifications for the AneuFiler^TM kit are provided in the Supplementary Table 1.

Capillary Electrophoresis

Capillary electrophoresis was performed using the 3500Dx Genetic Analyzer with a 50 cm capillary and POP-7^TM polymer, all of which are products of Thermo Fisher Scientific (USA). According to the instructions of the AneuFiler^TM kit, 1.0 μL of PCR amplification products was taken to prepare the capillary electrophoresis system. CM-LIZ500 was used as the internal DNA fragment size standard in the capillary electrophoresis (Shanghai Cubicise Medical Co., Ltd., China). The capillary electrophoresis parameters were configured according to the instructions of the AneuFiler^TM kit.

Capillary Electrophoresis Data Analysis

Capillary electrophoresis raw data were analyzed using the GeneMapper v3.0 software package (Thermo Fisher Scientific, USA). The Panel file and Bin file of the AneuFiler^TM kit were provided by the supplier to establish the analysis methods. Following the instructions of the AneuFiler^TM kit, capillary electrophoresis profiles of each sample were obtained. The genotype data of the 20 autosomal STRs for each sample were exported into a Microsoft Excel file based on the instructions of the GeneMapper v3.0 software for subsequent statistical analysis. Peak height ratios between 0.80 and 1.50 were considered normal, whereas ratios between 0.45 and 0.65 or between 1.80 and 2.40 were interpreted as trisomy. Three alleles of equal peak area were also considered trisomy. The presence of a single peak was usually considered uninformative. The DNA profiling of normal diploid female sample, normal diploid male sample, and male trisomy21 are provided in the Supplementary Figures 1–3, respectively.

Population Genetic Analysis

The population allele frequencies (f) of the 20 STRs in the study group were obtained using the STRAF v2.1.5 online analysis package.^10,11 The Hardy-Weinberg equilibrium test was conducted, and the linkage disequilibrium between pairs of STRs on the same chromosome was also assessed using the STRAF v2.1.5. Additionally, a principal component analysis (PCA) was performed to investigate the presence of substructures within the study population.¹¹

System Performance Parameters

According to the guidelines for QF-PCR application in aneuploidies testing in prenatal diagnosis, the number of heterozygous loci (x) among m examined STRs (m ≥ 4) on the same chromosome is crucial for determining the ploidy of the autosomal chromosome.^8,9 If x equals zero, indicating that all of the m STRs are homozygous, it becomes difficult to determine the ploidy of the chromosome. The observed heterozygosity (H_obs) of the 20 autosomal STRs was calculated using a counting method with Microsoft Excel. The expected heterozygosity (H), random match probability (PM), individual discrimination power (DP), power of non-maternal or non-paternal exclusion in duos (PE_duos), as well as the cumulative random match probability (CPM), total discrimination power (TDP), and cumulative power of non-parental exclusion in duos (CPE_duos) were computed based on the population allele frequencies of STR using formulas from references.^10,12–14

Impact of Linkage Effects on the Detection Rate of Heterozygosity

The potential linkage effects of pairwise STRs on the same autosomal chromosome were evaluated using the Binomial distribution function. The actual sample proportion, P(x ≤ k) (k = 0, 1, 2,., m), in the study population was obtained with direct counting method. When the potential linkage effects of pairwise STRs on the same autosomal chromosome could be ignored, meaning that the genotyped STRs on the same autosome chromosome could be treated independently, the expected sample proportion, P(x ≤ k) (k = 0, 1, 2,., m), could be obtained using the Binomial distribution function with H as the population rate.¹⁵ Here, H represented the average expected heterozygosity of m STRs genotyped on the same autosomal chromosome. Nonlinear fitting for the actual sample proportion distribution of P(x ≤ k), denoted as P-true, and the expected sample proportion distribution of P(x ≤ k), denoted as P-Emulation, was performed using GraphPad Prism v8.0 software.

Result

Population Allele Frequencies of the 20 Autosomal STRs

The population allele frequencies of the 20 STRs genotyped on chr13, chr18 and chr21 are presented in Tables 1–3, respectively. And 20 autosomal STR genotypes of the 2333 mother-child sample pairs are shown in Supplementary Table 2. A total of 250 alleles were observed from the 20 autosomal STRs, and 8–23 alleles were observed for each locus with the corresponding allele frequencies spanned from 0.0002 to 0.5011. The loci with the highest number of alleles were D13S325 on chr13 (15 alleles), D18S1371 on chr18 (13 alleles), and D21S1444 on chr21 (23 alleles).

Table 1 Population Allele Frequencies of the 6 STRs on chr13 (2N = 4666)

Table 2 Population Allele Frequencies of the 6 STRs on chr18 (2N = 4666)

Table 3 Population Allele Frequencies of the 8 STRs on chr21 (2N = 4666)

Population Genetic Statistic

The observed heterozygosity and expected heterozygosity values spanned from 0.6438(D18S391) to 0.8757(D13S788) and from 0.6433(D18S391) to 0.8904(D13S788). Forensic statistical parameters were calculated and reported in Table 4, the PM, DP, and PE_duo were observed to range from 0.0218(D13S788) to 0.1852(D18S391), 0.8148(D18S391) to 0.9782(D13S788), and 0.2241(D18S391) to 0.6380(D13S788), respectively. Based on the Hardy-Weinberg genetic equilibrium test, except for D21S1414 and D21S1444, all other 18 STRs reached genetic equilibrium (Table 4, after Bonferroni adjusted, P > 0.0025). The linkage disequilibrium test for STRs on the same chromosome suggests that closely linkage was not observed (Table 4). PCA aims to reduce dimensionality by transforming multiple variables into a smaller set of composite indicators, extracting the most important characteristics from data. This analytical method projects the data onto a new coordinate system, where the greatest variance in the data lies on the first coordinate (called the first principal component, PC1), the second greatest variance on the second coordinate (PC2). In Figure 1, the X-axis represents the first principal component (PC1), and the Y-axis represents the second principal component (PC2). PCA analysis showed that the PC1 and PC2 interpreted 0.81% genetic variance and 0.8% genetic variance, respectively. Most of the Han populations were obviously cluster tightly within an elliptical strip in the PCA plot, indicating the study population does not show population stratification in the study samples. The system performance parameters of the multiplex genotyping system composed of the 20 autosomal STRs were presented in Table 5. The probability of at least one heterozygosity among STRs on chr13, chr18, and chr21 is 0.999953, 0.999743, 0.999994, respectively. The probability of all STRs being homozygous on chr13, chr18, and chr21 is 4.09E-05, 0.000195, 4.89E-06, respectively. On chr21, the CPM is the lowest, while the TDP is the highest. The CPE_duos for 20 STRs is 0.999982. These findings indicated that 20 autosomal STR loci are highly polymorphic and informative in the East Han population and can be used as a powerful tool in personal identification and parentage testing.

Table 4 Performance Parameters for Each of the 20 Autosomal STRs (N = 2333)

Table 5 Performance Parameters for the 20 STRs Genotyping System

Figure 1 Principal component analysis (PCA) results based on the 20 autosomal STRs genotype dataset. The X-axis represents the first principal component (PC1), and the Y-axis represents the second principal component (PC2). PCA analysis showed that the PC1 and PC2 interpreted 0.81% genetic variance and 0.8% genetic variance, respectively. Most of the Han populations were obviously cluster tightly within an elliptical strip in the PCA plot, indicating the study population does not show population stratification in the study samples.

Impact of Linkage Effects on the Detection Rate of Heterozygosity

In the study cohort, the actual and expected sample proportion with x heterozygotes, P(x = k) (k = 0, 1, 2,., m), on chr13, chr18, and chr21 were presented in Table 6. Based on the results in Table 6, P-true, the actual sample proportion with P(x ≤ k), and P-Emulation, the expected sample proportion with P(x ≤ k) could be obtained. And the nonlinear fitting curves of P-true and P-Emulation were depicted in Figure 2. For the three detected autosomal chromosomes, the two curves completely overlap, indicated that the linkage effect between STR loci on the same chromosome has a negligible impact on the probability of detecting heterozygotes.

Table 6 Sample Proportion with P(x = k) (k = 0, 1, 2,., m) on the 3 Chromosomes (N = 2333)

Figure 2 Nonlinear fitting curves of P-true and P-Emulation for the chromosome 13,18,21. The Y-axis stands for cumulative probability corresponding to the cumulative number of heterozygous markers, whereas the X-axis represents cumulative number of heterozygous markers. The blue curves represent for P-True which was the sample proportion with P(x = k) (k = 0, 1, 2,., m) obtained by directly counting method, the red curves represents P-Emulation which was the sample proportion with P(x = k) (k = 0, 1, 2,., m) obtained with the binomial distribution function Bx (H, m) and R2 stands for coefficient of determination for non-linear fitting. The two curves completely overlap, indicated that the linkage effect between STR loci on the same chromosome has a negligible impact on the probability of detecting heterozygotes, since the curve of P-Emulation was obtained under the hypothesis that all the detected STRs on the same chromosome were completely independent. With the cumulative number of heterozygous markers rising, cumulative probability went up: (a) Chromosome 13, (b) Chromosome 18, (c) Chromosome 21.

Discussion

The QF-PCR technology based on STRs has been widely used in quality control of prenatal diagnosis and rapid testing of common chromosomal aneuploidies.^8,9 In 2007, the Association for Clinical Genomic Science (ACGS) published a guidance document on the use of QF-PCR for rapid detection of common aneuploidies, which was updated in 2018 (referred to as international guidelines).⁹ In 2016, China also released the “Expert Consensus on the Application of Quantitative Fluorescent PCR Technology in Prenatal Diagnosis” (referred to as the Chinese Expert Consensus).¹⁶ Whether the long-term practice of STRs in forensic genetics,^1,5 or the international guidelines⁹ and Chinese expert consensus,¹⁶ all indicate that the number of STRs used and their heterozygosity are critical factors in determining the performance of the corresponding QF-PCR multiplex genotyping system. After obtaining the population allele frequencies of the STRs from a representative population, performance parameters of the STRs, such as the random match probability and power of non-maternal or non-paternal exclusion in duos, etc., can be calculated using the established method in the field of forensic genetics.^5,12,15 These system performance parameters are key indicators for assessing the quality control capability of STRs in prenatal diagnosis. Through these parameters, we can determine whether a multiplex genotyping system is able to differentiate different prenatal samples and accurately determine the kinship between the mother and the fetus. A multiplex genotyping system with a lower CPM and higher CPE_duos is important to eliminate the possibility of sample confusion in prenatal diagnosis procedure. Additionally, the allele frequency of STRs can generally be used to estimate the capability of a multiplex genotyping system for testing the chromosomal molecular karyotype. When multiple STRs on a chromosome are genotyped, this capability is closely related to the probability of at least one heterozygote being detected.

In this study, allele frequencies of 20 autosomal STRs on ch13, chr18, and chr21 were obtained from 2333 Chinese Han adult female individuals from East China that it is one of the first to systematically evaluate the performance of AneuFiler^TM multiplex genotyping system for prenatal diagnosis in a large East Chinese Han cohort. The Hardy-Weinberg equilibrium test showed that only two STRs, D21S1414 and D21S1444, did not meet genetic equilibrium in the study population. Further analysis revealed that this may be due to the presence of multiple rare alleles for the two STRs. After excluding genotype data containing the rare alleles for the two STRs from the dataset, no significant deviation from genetic equilibrium was observed in the distribution of their genotypes. Therefore, the obtained population allele frequencies can be used for estimating relevant parameters in subsequent analysis.¹² Based on the dataset of 20 autosomal STRs, PCA analysis and linkage disequilibrium tests showed no significant substructure or linkage disequilibrium in the study population, supporting the removal of rare genotypes from the dataset.¹²

Based on the analysis of population genetics, Tables 4 and 5 present the key parameters, such as observed heterozygosity and expected heterozygosity, for the 20 autosomal STRs. The 20 autosomal STRs exhibit high levels of polymorphism within the study population, with average expected heterozygosity values of 0.8097, 0.7478, and 0.7760 for chr13, chr18, and chr21, respectively, meeting the requirements of STR loci with high heterozygosity in international guidelines.⁹ The cumulative random match probability for these 20 STRs is 3.10E-23, indicating that their individual identification capability surpasses that of the 13 CODIS STRs and is comparable to that of the NDIS, which is an expanded panel of 20 core forensic STR loci.¹⁷ The CPE_duos for the 20 autosomal STRs also meet the level of commonly used parentage testing STR genotyping tools,¹⁵ meeting the requirements specified in the “Technical Specifications for Parentage Testing”¹⁸ These results also indicated that the 20 STRs have reached the same level as commonly used forensic STR genotyping tools in prenatal diagnosis quality control, including differentiating different prenatal samples and kinship determining for the mother-fetus pair. Additionally, prenatal samples such as cord blood, amniotic fluid, and chorionic villus sampling are prone to contamination by maternal cells (maternal cell contamination, MCC) or mixing with other individual samples, which can interfere with prenatal diagnostic results.¹⁶ Previous practices using forensic STR genotyping tools have demonstrated the capability of STRs in MCC testing. Theoretically, the 20 STRs utilized in this study also suitable for MCC testing. Moreover, since this multiplex genotyping system includes six or more STRs on chr13, chr18, and chr21, respectively, it enables rapid testing for these chromosomal aneuploidies while performing quality control in prenatal diagnostic procedures.

The profiles of STR, including genotype and peak-height patterns, are the main basis for determining the corresponding chromosome molecular karyotype.^8,9,16 It is evident that when all the tested STRs on a chromosome are homozygous (ie, with a heterozygote count of x = 0), it is difficult to determine the chromosome molecular karyotype. Based on the average expected heterozygosity of the STRs in this multiplex typing system, the probabilities of STRs on chr13, chr18, and chr21 being homozygous [P(x = 0)] is as low as 4.76E-05, 0.000258, and 6.33E-06, respectively. This indicates that the 20 autosomal STRs have ideal specificity in determining the corresponding chromosomes as disomy, achieving respective values of 0.999952, 0.999742, and 0.999994. However, the evaluation of this key parameter assumes that these STRs are independent of each other. This premise is the basis for the calculation with the multiplication principle of probabilities.^5,13 In traditional forensic multiplex STR genotyping systems, to ensure the independence of STR loci and facilitate the application of the calculation with the multiplication principle of probabilities, multiple STRs from the same chromosome are typically not included.¹ However, STR multiplex genotyping systems designed for prenatal aneuploidy testing typically require the inclusion of four or more genetic markers on the same chromosome.¹⁶ To evaluate whether the potential linkage effects between different genetic markers on the same chromosome have a negligible impact on critical parameters such as CPM and CPEduos, this study assessed the nonlinear fitting of cumulative probability distributions of samples with varying heterozygosity counts in the population. The actual cumulative probability distribution of different heterozygote counts on each chromosome shown in Figure 2 does not significantly differ from the simulated cumulative probability distribution based on the average expected heterozygosity, indicating that the potential linkage effects of these STRs on the same chromosome can be ignored, which is associated with uniform distribution of STRs on each chromosome in the AneuFiler™ multiplex genotyping system. This finding is consistent with the results of the linkage disequilibrium test. Additionally, the genetic distances between loci on each chromosome may introduce potential linkage effects. Based on the genetic distances of autosomal chromosomes provided in the UCSC Hg19 database, the average genetic distance of approximately 10.0 cM or more between them (average genetic distance on chr13 is 18.4 cM, on chr18 is 24.1 cM, and on chr21 is 9.9 cM).³ In the field of forensic genetics, extensive research has been conducted to assess the potential linkage effects between vWA and D12S391, two STRs located on the short arm of chr12 with a genetic distance of approximately 10.0 cM, on the calculation of the corresponding system performance parameters. Studies involving over 50,000 pedigrees have demonstrated that the linkage disequilibrium between vWA and D12S391 has a negligible impact on the calculation of critical forensic parameters, including CPM, CPEduos, and Kinship Index.³

One of the limitations of this study could be the lack of the STR allele frequencies in other ethnic populations, which constrains the generalizability of our findings. And the study can also expand on other potential applications such as distinguishing between monozygotic and dizygotic twins, distinguishing surrogacy cases, and risk mitigation in high-throughput prenatal labs. Additionally, our study was primarily based on allele frequencies in typical diploid samples and did not assess the diagnostic performance of this genotyping system in abnormal cases such as aneuploidy or maternal cell contamination, nor did it evaluate locus mutation rates. These aspects remain part of our ongoing research. Nonetheless, our research provided a framework for clinical implementation, and we plan to enlarge our cohort in other populations to confirm the applicability of these findings.

Conclusions

In summary, our study indicates that the 20 autosomal STRs included in the AneuFiler^TM multiplex genotyping system exhibit highly polymorphic in East Chinese Han population and are independent on chr13, chr18 and chr21. The comprehensive system facilitated the implementation of quality control measures in clinical prenatal diagnosis, including sample identification, identification of maternal contamination, and kinship verification of prenatal maternal-fetal sample pairs, thereby enabling accurate determination of chromosomal ploidy. Moreover, this system demonstrates extended clinical utility in zygosity testing for twin pregnancies, uniparental disomy (UPD) detection, pre-NGS sample quality control and so on. Nevertheless, further analysis and research involving more diverse populations and aneuploidy cases are still needed and may increase the diversity of applicable populations.