Blood Group ABO Genotyping in Paternity Testing (2024)

Language: English | German

Abstract

Keywords: ABO genotyping, Paternity likelihood, Power of exclusion, Paternity trios

Abstract

Introduction

After the discovery of the Mendelian inheritance of the ABO blood group system by von Dungern and Hirschfeld in 1910 [1], blood group phenotyping including further systems has been introduced to paternity investigation [2]. For more than 60 years investigation of blood group systems and the HLA system represented the standard in forensic genetics. This serological era was then replaced by molecular genetics through the introduction of DNA fingerprints first on the basis of minisatellite (i.e. variable number of tandem repeat; VNTR) polymorphisms, later by the analysis of microsatellite (i.e. short tandem repeats; STR) polymorphisms [3]. Recent efforts led to the introduction of single nucleotide polymorphism (SNP) typing for human identification [4].

The ABO gene is located on the long arm of human chromosome 9 and consists of 7 exons. Exons 6 and 7 encode the major proportion including the catalytic domain of the glycosyltransferases that mediate the expression of A and B antigens [5, 6]. A number of studies have shown a clear genotype-phenotype correlation [7, 8, 9], and more than 240 gene variants are listed in the open access database dbRBC at the NCBI website [10]. Most of the gene variants are characterized by one or more SNPs leading to amino acid changes or stop codons. In addition, some variants result from nucleotide insertions or deletions (indels) mostly leading to a frame shift of the reading frame and subsequent alteration of the encoded amino acid sequence. The ABO blood groups A₁, A₂, and B result from the wild type allele ABO*A101 and the major variant alleles ABO*A201, and ABO*B101. The O phenotype is mostly caused by deletional O alleles such as ABO*O01 or ABO*O02 each characterized by the indel polymorphism 261G>del in exon 6. The most common non-deletional O allele is ABO*O03 with a point mutation in exon 7 also leading to the O phenotype (table ​(table11).

In the present study we determined the frequencies of the major ABO alleles in a representative German population by using PCR with sequence-specific primers (PCR-SSP). We then used serologic and genetic ABO blood grouping in paternity trios (child, mother, alleged father) in order to compare paternity likelihood (PL) and power of exclusion (PE) of both blood grouping approaches.

Material and Methods

Results

PCR-SSP-based genotyping for the ABO polymorphisms 261G>del, 802G>A, 803G>C and 1061C>del was performed on 1,335 samples of an anonymized blood donor DNA bank. Unambiguous results could be obtained for all samples (table ​(table2).2). Each of the 4 polymorphisms fulfilled the Hardy-Weinberg equilibrium. The 261del variant characteristic for the majority of O alleles was the most prevalent variant of the ABO gene (1,590 of 2,670 alleles; 59.6%). Absence of all four variants indicated the presence of the wild type A1 allele. However, mutations at other sites of the ABO gene cannot be excluded by the low-resolution genotyping.

In order to determine the specificity of the low-resolution genotyping approach we compared the ABO blood group phenotype (A, B, AB, or O) registered in the blood donor files with the determined genotype. Blood groups A, B, and AB revealed a 100% phenotype/genotype correlation with the different genotypes (table ​(table3).3). For 519 of the 523 samples (99.2%) with blood group O, we could determine a corresponding blood group O genotype. In the remaining 4 samples the low-resolution genotyping approach indicated blood group A (O01/A1). It is assumed that in these samples rare variants (Var), i.e. O alleles with mutations at other sites of the ABO gene, are present. We therefore classified the genotype of the 4 samples as O01/Var, and the frequency of Var was 0.15% (4 of 2,670 alleles) in our cohort. In summary, 1,331 of 1,335 (99.7%) blood groups were correctly determined from the low-resolution ABO genotype. Based on the observed allele frequencies (O01, 0.5955; A1, 0.2176; B, 0.1004; A2, 0.0637; O03, 0.0213; Var, 0.0015) we calculated statistical parameters: i) observed heterozygosity = 0.5775; ii) PIC = 0.5380; iii) PD = 0.7803; iv) PE = 0.2648.

To further evaluate the impact of the ABO genotype in kinship analysis, we performed the low-resolution ABO genotyping and ABO phenotyping (A1, A2, B, and O) in 95 paternity trio cases (child, mother, alleged father). Previously, paternity of the alleged father was determined on the basis of STR analysis including 12 markers (D3S1358, D7S820, D13S317, D16S539, D18S51, D21S11, CSF1PO, FGA, PentaE, PentaD, TH01, TPOX). Paternity was excluded in 35 cases and confirmed in 60 cases. In 23 of the 35 excluded paternities (65.7%) neither the ABO genotype nor the ABO phenotype enabled exclusion of the alleged father. In 7 cases (20%) both ABO genotype and phenotype confirmed exclusion of the alleged father. In the remaining 5 cases (14.3%) the ABO genotype, but not the ABO phenotype, enabled exclusion of the alleged father (table ​(table44).

In the 60 cases with confirmed paternity, both PL and PE of the ABO genotype was significantly higher than those of the ABO phenotype (table ​(table5).5). Only in three cases where the child's ABO phenotype was B and the mother's phenotype was A or O the PE of the ABO genotype was slightly lower compared to the ABO phenotype (80.9% vs. 84.8%). In summary, the ABO genotype was superior to the ABO phenotype in all 95 paternity trio cases.

Discussion

In our study we determined the prevalence of the major ABO gene variants in a population of German blood donors. Four diallelic markers were determined by PCR-SSP with optimized sensitivity in order to reduce DNA consumption [12]. The ABO genotype could be obtained from less than 50 ng genomic DNA. However, the impact of DNA degradation on ABO genotyping by PCR-SSP has not been investigated so far. The genotypes were used to deduce the corresponding ABO blood group phenotype, and we found 99.7% genotype-phenotype correlation for our low-resolution genotyping approach. The frequency of the different blood groups in our cohort corresponded to the expected frequencies (A 43%; B 9%; AB 4%; O 44%) for Caucasian populations [13]. Thus, we conclude that our study cohort is a representative sample of the German population. In 4 of 1,335 samples (0.3%) with O phenotype but O01/A1 genotype we proposed rare variant alleles. Sequence analysis of the entire ABO gene confirmed rare alleles such as Aw13 and O51 (data not shown). Most of the alleles stored in the dbRBC database represent rare ABO alleles characterized by point mutations in different exons of the ABO gene. Many variant alleles have a phenotypic equivalent such as weak expression of A or B antigens or blood group O. The genotype-phenotype correlation could be enhanced by investigating more of the rare variants.

We demonstrated that PL and PE of the ABO genotype is superior to the ABO phenotype. The inclusion of additional ABO alleles could even strengthen this finding; however, analysis of further SNPs would be necessary to identify such alleles. The O02 allele is a common variant of the O01 allele carrying the indel polymorphism 261G>del in exon 6 and nine additional nucleotide changes in exons 2–7. It was shown that the O02 allele accounts for approximately 40% of O alleles [14]. The differential typing for O01 and O02 alleles would further enhance the statistical power of ABO genotyping, but was not performed in this study. Using a theoretical allele frequency of 0.41 of the O01 and 0.18 of the O02 allele in our population and assuming Hardy-Weinberg equilibrium for all alleles we re-calculated the statistical parameters. Discrimination of O01 and O02 alleles further increased the observed heterozygosity to 0.6856 (before 0.5775), PIC to 0.7025 (before 0.5380), PD to 0.8925 (before 0.7803), and PE to 0.4064 (before 0.2648). These values are similar to those reported in ABO genotyping of a small (n = 60) Italian cohort [15]. We assume that discrimination between the O01 and O02 alleles is reasonable and could be performed by typing a further SNP such as 646T>A or 681G>A in exon 7 of the ABO gene.

Typing of highly polymorphic STR markers is the current worldwide standard in kinship analysis and forensic case work. The high PD and PE for each STR marker enable a reliable determination of paternity. Other marker systems such as blood groups, HLA, or RFLPs became uncomely due to the lower statistical power and higher labor input. However, the reasonable mutation rates of up to 1% of each STR marker could hamper result interpretation [16]. The mutation rate per meiosis of diallelic markers including SNPs is estimated to be 100,000 times lower compared to STRs. Therefore, the reliability of SNP-based kinship analysis is higher and could probably replace STR-based analysis in future. By using ABO blood grouping in paternity trio cases we could show that ABO genotyping is superior to phenotyping with regard to statistical parameter such as PL and PE. In 12 of 35 trios (34.3%) where paternity was excluded by at least 5 of 12 STR markers, the ABO genotype also excluded paternity. The ABO phenotype could exclude paternity in only 20% of the cases. However, we must point out that neither ABO genotyping nor ABO phenotyping is sufficient for valid paternity testing. It should be regarded as an supplementary and optional approach to add further value to paternity cases based on STR systems.

Since the discovery of SNPs as the genetic basis of the ABO and of many other blood groups, SNP typing could probably bring back blood grouping to paternity testing on a molecular genetic basis. For this purpose the genotype-phenotype correlation is an important prerequisite and has been demonstrated for many different blood group systems including ABO, Rh, Duffy, MNS, Kidd, and Kell. Identification of individuals using anonymous SNPs on different chromosomes such as the SNPforID markers [4] could be combined with SNPs of different blood group systems. In selected cases and in terms of control purposes, i.e. to match a blood sample with a DNA sample, the phenotypes of blood groups on different chromosomes could provide useful information. Phenotyping and genoytping of HLA antigens could be used as well.

The application of SNP typing in a routine laboratory such as clinical diagnostics or paternity testing requires proof of reliability of the method and standardization of the procedure. In our laboratory, ABO phenotyping and genotyping is validated for clinical diagnostics and paternity testing, and the procedure is defined in standard operation protocols. The same is true for other blood group systems such as Duffy, MNS, Kidd, and Kell. Low-resolution ABO genotyping by PCR-SSP method including agarose gel electrophoresis is a very time- (less than 2 h; many samples simultaneously) and cost-effective method (less than EUR 2.- per DNA sample). In paternity cases or any other forensic case work DNA samples for STR analysis can also be used for SNP typing. Using PCR-SSP with optimized sensitivity [12], genotyping requires less than 10 ng DNA per SNP and can be performed on DNA extracted from buccal swaps. Other SNP genotyping methods including multiplex PCR with capillary electrophoresis [17] or other procedures [18] would require even lower DNA amounts.

Finally, for the development of systems for genetic identity testing based on diallelic markers, we would recommend to combine anonymous SNPs and SNPs associated with an easy-to-determine phenotype such as blood groups or HLA. This could open a new field of clinical application of genetic identity testing, e.g. where the identity of a hematopoietic stem cell transplant shipped from the donation site to the transplantation center is rechecked.

Disclosure Statement

The authors declare no conflict of interest.

Acknowledgements

This work is dedicated to the Polish pioneer in serology and discovery of ABO blood group inheritance Ludwik Hirzsfeld.

References