Partial RHD genotypes in genome databases

BACKGROUND: Partial RHD genotypes permit anti-D formation after D+ RBC exposure. Better understanding of partial RHD genotype frequencies (GF) could improve ancestry-specific genotype testing and allele-specific anti-D risk estimation. Large genome databases are now easily accessible but quantification of multivariant RHD alleles in them has not been described. STUDY DESIGN AND METHODS: We obtained allele frequencies for single-nucleotide-variant (SNV) partial RHD alleles in the Genome Aggregation Database (gnomAD) and All of Us (AoU) databases and calculated hemizygous or homozygous GFs by Hardy-Weinberg formulas. For multivariant alleles, we obtained GFs from hemizygous or homozygous prevalences of informative variant pairs in gnomAD variant co-occurrence (VCO) queries. We compared database GFs to those derived from USA RHD gene surveys primarily in Black individuals and sickle cell disease (SCD) patients. RESULTS: In African ancestry, DAU3, weak D type 4.0, DIIIa, DAU5 and DIVa (in descending order) each had gnomAD GFs of 0.45%-1.02%. African gnomAD GFs were statistically similar to Black/SCD gene survey GFs. Admixed American (formerly Hispanic/Latino) alleles were similar but at 10% of African frequencies. For others, the most common genotypes in gnomAD and AoU were: DVII in European (gnomAD 0.15%), Middle Eastern (0.17%) and South Asian (0.03%); type 15 in East Asian (0.002%); type 4.0 (0.13%) in Ashkenazi Jewish. Extended RHCE gene conversions such as DVI were undetected by VCO and likely were misaligned into RHCE. DISCUSSION: SNVs and VCO in large genome databases provided partial-RHD GFs in diverse populations. This information will be useful in estimating allele propensities for anti-D formation.