Development and evaluation of a transfusion medicine genome wide genotyping array

Abstract

Background: Many aspects of transfusion medicine are affected by genetics. Current single-nucleotide polymorphism (SNP) arrays are limited in the number of targets that can be interrogated and cannot detect all variation of interest. We designed a transfusion medicine array (TM-Array) for study of both common and rare transfusion-relevant variations in genetically diverse donor and recipient populations.

Study design and methods: The array was designed by conducting extensive bioinformatics mining and consulting experts to identify genes and genetic variation related to a wide range of transfusion medicine clinical relevant and research-related topics. Copy number polymorphisms were added in the alpha globin, beta globin, and Rh gene clusters.

Results: The final array contains approximately 879,000 SNP and copy number polymorphism markers. Over 99% of SNPs were called reliably. Technical replication showed the array to be robust and reproducible, with an error rate less than 0.03%. The array also had a very low Mendelian error rate (average parent-child trio accuracy of 0.9997). Blood group results were in concordance with serology testing results, and the array accurately identifies rare variants (minor allele frequency of 0.5%). The array achieved high genome-wide imputation coverage for African-American (97.5%), Hispanic (96.1%), East Asian (94.6%), and white (96.1%) genomes at a minor allele frequency of 5%.

Conclusions: A custom array for transfusion medicine research has been designed and evaluated. It gives wide coverage and accurate identification of rare SNPs in diverse populations. The TM-Array will be useful for future genetic studies in the diverse fields of transfusion medicine research.

Fig. 1.

Key content modules included in the TM-Array. Marker modules are divided into those in common with the UK Biobank Axiom Array, genome-wide booster for Asian and African American population that is customized for this array, Affymetrix commercially available content, and those transfusion medicine-related SNPs, CNPs, and genes along with module descriptions and counts in each category. AA, African American; Affy, Affymetrix; GWA, genome-wide associated; Mt, mitochondria; RBCs, red blood cells; TGF, transforming growth factor. [Color figure can be viewed at wileyonlinelibrary.com]

Fig. 2.

Evaluation of CNP calling through comparison with serological records of Rh antigens. D antigen and the principle component analysis (PCA) abstracted RHD gene CNP results. Forty-one D-negative and 13 D-positive subjects were classified differently comparing marker-based and serologic testing (concordance rate 99.6%).