Uniparental disomy (UPD) refers to an entire or partial region of homologous chromosomes originating from a single parent rather than from both parents. This concept was first described by Engel in 1980 [1] and subsequently demonstrated as a mechanism for human genetic disease in 1988 [2]. Mechanisms for UPD formation include trisomy rescue, gamete complementation, postfertilization error, and monosomic rescue [3–7]. UPD can be classified as uniparental isodisomy (isoUPD, from a single chromosome of a single parent), uniparental heterodisomy (hetUPD, from both chromosomes of a single parent), and mixed UPD (both isodisomy and heterodisomy on the same chromosome), and based on the origin, it can be distinguished as maternal (matUPD) or paternal UPD (patUPD) [3–7]. In a study involving 4 million healthy individuals, the overall incidence of UPD was 1 in 2 000 births, with matUPD incidence higher than patUPD [4, 6, 8, 9]. Most UPDs are not diseasecausing, but when pathology does occur, it is primarily due to disruption of genomic imprinting (https://www.geneimprint.com/site/genes-by-species) or by revealing harmful recessive alleles in substantial homozygous regions of the affected chromosome [8]. In addition, the American College of Medical Genetics and Genomics (ACMG) has established a technical standard for UPD diagnosis [10], and its interpretation has also been published [11]. Further resources on UPD can be found at the website https://cstl.de/DB/CA/UPD/0-Start.html (Liehr T. 2023. Cases with uniparental disomy).

Paternity testing is based on the law of Mendelian inheritance. When there are genotypic Mendelian errors in short tandem repeat (STR) markers that do not conform to inheritance laws, we always interpret them as slippage mutations or silent alleles rather than UPD, especially when only one marker on a chromosome is tested. In this study, we describe two cases of Mendelian errors in paternity testing. Using multiple kits on both the capillary electrophoresis (CE) and massively parallel sequencing (MPS) platforms, parent–child relationships were found between the child and the alleged parent, and UPD can explain the observed genetic anomalies in both cases.

Blood samples from the two paternity testing cases were collected with written informed consent. Case 1 was the standard trio (mother, child, and alleged father), while Case 2 was a duo (alleged mother and child). In both cases, the children had normal phenotypes. We used the DNA IQ™ System (Promega, Madison, WI, USA) to extract DNA following the manufacturer's protocol. DNA was quantified using the Qubit^® Quantitation System (Invitrogen, Carlsbad, CA, USA).

CE platform: for Case 1, PCR amplification of DNA was performed using the Goldeneye 20A (Goldeneye, Beijing, China) and AGCU21+1 (AGCU, Wuxi, China) kits. For Case 2, PCR amplification was performed using the PowerPlex^®21 (Promega), Microreader™ 23sp-B (Microread, Beijing, China), and Goldeneye 17X (Goldeneye) kits. Amplification products were subjected to CE using an ABI PRISM 3130XL genetic analyzer (Applied Biosystems, Waltham, MA, USA). MPS platform: in both cases, the MGIEasy Signature Identification Library Prep Kit (MGI Tech, Shenzhen, China) was used for library construction, which included 129 STRs (54 A-STRs, 27 X-STRs, and 48 Y-STRs), 215 single nucleotide polymorphisms (SNPs), and three mitochondrial hypervariable regions (HVR-I, II, and III). Sequencing was performed on the MGISEQ-2000RS sequencer (MGI Tech). All procedures were carried out according to the manufacturers' instructions.

For CE platform, allele typing was performed using GeneMapper ID-X software (https://www.thermofisher.cn/cn/zh/home/technical-resources/software-downloads/genemapper-id-x-software.html). For the MPS platform, the sequencing data were analysed using the sequencer's built-in typing software. Default thresholds were used in the analyses.

Using the CE and MPS platforms, a total of 54 autosomal STRs (A-STRs) and 215 autosomal SNPs were genotyped in two cases. Genotyping results for shared loci were consistent between the two platforms. The genotypes of all loci on chromosomes 6 and 3 in Cases 1 and 2, respectively, were homozygous, with Mendelian errors observed at certain loci (highlighted in bold in Table 1), and the loci on other autosomal and X-chromosomes followed the inheritance law (data not shown). In Case 1, the child's mtDNA and Y-STR genotypes matched the mother and alleged father, respectively; in Case 2, the child's mtDNA genotype matched the alleged mother. Furthermore, the combined paternity index (CPI) was calculated based on the STR stepwise mutation model (for non-Mendelian loci), and the results for both cases [Case 1: CPI = 7.3×10²¹ (39 A-STRs); Case 2: CPI = 6.11×10⁶ (39 A-STRs)] supported the parent–child relationship between the child and the alleged parent. In general, this approach is quite conservative, as Cavalheiro et al. [3] suggest that it reduces the strength of the evidence. They propose considering the genetic markers on the UPD chromosome as a whole when calculating the PI value for similar cases. The specific method is to use the prevalence of UPD on the chromosome in the population as the numerator, and the product of the obligate paternal allele (OPA) frequencies for UPD chromosome markers as the denominator [3]. This is a valuable suggestion. Unfortunately, the prevalence of UPD on chromosomes 6 and 3 was not found, and we did not calculate the CPI referred to this method.

In detail, a total of 18 autosomal genetic markers (4 STRs and 14 SNPs) located on chromosomes 6 and 3 were analysed in Case 1 and Case 2, respectively (Table 1). These markers are distributed over almost entire chromosomes (Figure 1). In Case 1, the Mendelian errors were observed in three STRs and four SNPs between the child and the mother; whereas in Case 2, there were four STRs and five SNPs (highlighted in bold in Table 1). Generally, when these allelic Mendelian errors are observed, STR slippage mutations, allele silencing, or point mutations are considered first. However, it is evident that the Mendelian errors we observed cannot be explained by the above reasons. As shown in Table 1 and Figure 1, the Mendelian errors occurred only on one specific chromosome, and all genetic markers along the entire chromosome show a homozygous state. Mendelian errors were not found on other chromosomes; in other words, genotyping errors or parental consanguinity were ruled out. Overall, this strongly suggests a complete isodisomy UPD (isoUPD) event, the UPD chromosome being of paternal origin. We find that certain markers on the affected chromosome show expected inheritance patterns (Table 1, including 11 markers on chromosome 6 in Case 1 and 9 markers on chromosome 3 in Case 2), especially most of the SNP markers (biallelic markers with lower discrimination power). Although this does not explicitly support the occurrence of UPD, it is compatible with the UPD hypothesis. In conclusion, UPD can explain the observed genetic anomalies in both cases, with the most likely mechanism being the failure of sister chromatid separation during the meiosis II of the paternal sperm, followed by trisomy rescue resulting in the loss of the maternal chromosome [4].

UPD is observed on all chromosomes [8]. Cases of UPD in paternity testing are relatively rare, with only isolated reports to date, such as chromosome 2 [12–14], chromosome 4 [15], and chromosome 21 [3]. These reports typically involve individuals with a single UPD chromosome, and in rarer cases, multiple UPD chromosomes [16]. In practice, UPD may be underestimated in paternity testing due to scenarios in which parents and children share alleles, single-marker detection on the chromosome, consanguinity between parents (resulting in suspected “UPD” on multiple chromosomes), segmental UPD (affecting only a specific region on a chromosome) [17], etc. When such scenarios are possible, it is necessary to use more kits or perform whole genome sequencing for verification. The ACMG's guidelines [10] recommended using at least two STR loci for UPD diagnosis. However, in the two cases presented in this study, we used four STR loci for UPD diagnosis, exceeding the number required by the ACMG. It is worth noting that this standard has certain limitations; e.g. the number of specified detection loci may be insufficient, leading to potential diagnostic errors in cases involving somatic mosaicism or segmental UPD. In addition, the standard lacks specific criteria regarding the polymorphism (such as heterozygosity) and distribution across chromosomes for the STR loci used in UPD detection. UPD should attract the attention of forensic practitioners. It is conceivable that failure to recognize the allelic patterns indicative of UPD in practice could lead to various issues, such as erroneous exclusion in paternity testing, reduced success rates in missing person database searches, and interference with IBD analysis in forensic genetic genealogy (FGG). Therefore, the presence of UPD should be considered when encountering atypical genotyping in forensic practice.

Here, we report two cases of UPD events occurring on chromosomes 6 and 3 in paternity testing. By testing with multiple kits on both the CE and MPS platforms, the results showed that the genetic markers from the affected chromosome exhibit a homozygous state. Considering the distribution of the detected markers across the entire chromosome and the parental single-chromosome origin, we conclude that both cases represent complete isodisomy UPD. Recent studies suggest that the prevalence of UPD may be comparable to the mutation rate of STRs [3]. In paternity testing, we should pay special attention to UPD, especially when the homozygous loci are all located on the same chromosome.

Thanks to Dr. Man Chen from the School of Forensic Medicine, Southern Medical University, for providing suggestions and revisions to this article. Thanks to Wei Peng, Bo Zhang, and Haofang Mu from the Center of Forensic Sciences, Beijing Genomics Institute for their assistance with massively parallel sequencing.

Jiangwei Yan conceived and designed this study. He Ren and Zhiyong Liu conceived the experiments and wrote the manuscript. Chong Chen and Yan Shi collected the samples. Jiarong Zhang, Ying Chen and Li Jia extracted DNA and helped to conduct the statistical analysis. Yacheng Liu and Jiangwei Yan revised the manuscript. All authors contributed to the final text and approved it.

This study was performed in accordance with the principles of the Declaration of Helsinki and approved by the Ethics Committees of Shanxi Medical University (No. 2020GLLO31). Written informed consent was obtained from all the participants.

None declared.

This work was supported by the grants from the Open-Fund of Shanghai Key Lab of Forensic Medicine, Key Lab of Forensic Science, Ministry of Justice, China [grant number KF202011] and National Natural Science Foundation of China [grant number 82030058].