Premium
Automated typing of red blood cell and platelet antigens from whole exome sequences
Author(s) -
Lane William J.,
Vege Sunitha,
Mah Helen H.,
LomasFrancis Christine,
Aguad Maria,
SmelandWagman Robin,
Koch Christopher,
Killian Jacqueline M.,
Gardner Cubby L.,
De Castro Mauricio,
Lebo Matthew S.,
Kaufman Richard M.,
Green Robert C.,
Westhoff Connie M.
Publication year - 2019
Publication title -
transfusion
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.045
H-Index - 132
eISSN - 1537-2995
pISSN - 0041-1132
DOI - 10.1111/trf.15473
Subject(s) - typing , genotyping , biology , antigen , polymerase chain reaction , zygosity , serology , exome sequencing , single nucleotide polymorphism , snp , histocompatibility testing , genetics , genotype , gene , antibody , human leukocyte antigen , phenotype
BACKGROUND Genotyping has expanded the number red blood cell (RBC) and platelet (PLT) antigens that can readily be typed, but often represents an additional testing cost. The analysis of existing genomic data offers a cost‐effective approach. We recently developed automated software (bloodTyper) for determination of RBC and PLT antigens from whole genome sequencing. Here we extend the algorithm to whole exome sequencing (WES). STUDY DESIGN AND METHODS Whole exome sequencing was performed on samples from 75 individuals. WES‐based bloodTyper RBC and PLT typing was compared to conventional polymerase chain reaction (PCR) RHD zygosity testing and serologic and single‐nucleotide polymorphism (SNP) typing for 38 RBC antigens in 12 systems (17 serologic and 35 SNPs) and 22 PLT antigens (22 SNPs). Samples from the first 20 individuals were used to modify bloodTyper to interpret WES followed by blinded typing of 55 samples. RESULTS Over the first 20 samples, discordances were noted for C, M, and N antigens, which were due to WES‐specific biases. After modification, bloodTyper was 100% accurate on blinded evaluation of the last 55 samples and outperformed both serologic (99.67% accurate) and SNP typing (99.97% accurate) reflected by two Fy b and one N serologic typing errors and one undetected SNP encoding a Jk null phenotype. RHD zygosity testing by bloodTyper was 100% concordant with a combination of hybrid Rhesus box PCR and PCR–restriction fragment length polymorphism for all samples. CONCLUSION The automated bloodTyper software was modified for WES biases to allow for accurate RBC and PLT antigen typing. Such analysis could become a routing part of future WES efforts.