Novel or not? Reference alleles, genes, and genomes to unmask the true nature of the ABO*AW.10 allele associated with weak A phenotype

The main genetics underlying the four major phenotypes within the ABO blood group system were originally reported by Fumiichiro Yamamoto in 1990 and retrospectively commented on in 2021.1–3 Within these 30 years, in addition to the original discoveries, more than 200 allelic variants have been reported, many of which differ only by single nucleotide variant(s) (SNV) from the currently accepted genomic (NG_006669.1) and coding ABO*A1.01, (NM_020469.2) consensus sequences. These numerous alleles are of clinical significance as they can result in weakened antigen expression or abolish expression completely, which may in turn cause ABO typing discrepancies, failure to determine the ABO group, or assignment of an incorrect group. Such typing errors could ultimately lead to unnecessary overconsumption of group O cellular blood components in patients, transfusion of blood of the wrong group, or rejection of a potential blood donor. Therefore, for blood group molecular typing laboratories, electronic databases serve as an important resource to associate an allelic variant with the predicted blood group phenotype. Since two such databases, the Blood Group Antigen Gene Mutation Database (BGMUT) and the dbRBC of the National Center for Biotechnology Information (NCBI), were discontinued after 2016, the listing maintained by the Working Party for Red Cell Immunogenetics and Blood Group Terminology (RCI and BGT) of the International Society of Blood Transfusion (ISBT) is now the only internationally recognized database on ABO polymorphism. Recent developments in next-generation sequencing (NGS) will enable the sequencing of millions of human genomes, ushering in the transition from the era of genetics to the era of genomics. It is hardly prophetic to predict that this technological revolution will drastically increase the number of known allelic variants and not only of the ABO locus. It is therefore critical to reliably and accurately document the alleles that are already known, which in all probability represent the most frequent variants, to define a reliable starting point. However, reference alleles do not necessarily have to be those alleles that are the most common representatives in all regions of the world. This is exemplified in East Asian populations such as Korea and Japan where an ABO allele, denoted ABO*A1.02 is more common than the consensus sequence ABO*A1.01. Differing from the consensus sequence by only one SNV in the coding region, c.467C>T giving rise to p.Pro156Leu, ABO*A1.02 results in the same A1 phenotype according to currently used methods of detection. In this issue of Transfusion, a letter by Minjeong Nam et al. argues that a novel variant allele reported by Kim et al. in 2020 in Korean individuals was not actually novel. In 2020, Hye Ran Kim et al reported that they found a “novel” A allele carrying both the c.467C>T and the c.784G>A nucleotide substitutions. The c.467T and the c.784A variants had been listed separately on the ISBT database. The c.784G>A change results in weak A antigen expression, defined as ABO*AW.10. The c.467C>T substitution in combination with c.784G>A was claimed to be the novelty. Their research was conducted based on the reference allele ABO*A1.01, and their finding was directly submitted as an ABO*AW variant allele to GenBank (accession number KU961549.1). Minjeong Nam and colleagues now point out that the A allele with combined c.467C>T and c.784G>A nucleotide substitutions was not novel but had been reported earlier, primarily in Korean journals.7–11 Nam notes that in the Korean ethnic group, ABO*A1.02 is more prevalent than ABO*A1.01 (78.6% vs. 21.4%); therefore, Korean researchers have chosen ABO*A1.02 as a reference allele for the genetic analysis of A alleles.7–11 Kim et al. reply to Nam in this issue of Transfusion that the only internationally accessible publication not written in Korean does not clearly indicate the existence of c.467C>T in combination with a c.784G>A substitution for ABO*AW.10. In this editorial, we suggest that the confusion discussed seems to be caused by a combination of factors that all can learn from: (1) The use of different reference sequences may be one reason for the dispute. When screening GenBank using NCBI's blast algorithm, entering the search sequence c.784A alone and in combination with the Received: 16 February 2022 Accepted: 16 February 2022