Journal: medRxiv

Article Title: Calling for diversity: improving transfusion safety through high-throughput blood group microarray genotyping

doi: 10.1101/2023.12.15.23299980

Figure Lengend Snippet: ( A ), Dot plot of RHD (NG_007494.1, x-axis) and RHCE (NG_009208.3, y-axis) ISBT genomic reference sequences created with the YASS genomic similarity search tool. The diagonal line indicates the successful forward alignment and hence the sequence similarity of the two genes over almost their entire sequence. ( B ) Schematic approach of microarray genotyping. An immobilized probe on a bead chip designed against a specific genomic locus is hybridized with a target fragment. A polymerase extends the 3’ end of the probe by one base according to the template sequence that represents the desired SNP position. dNTPs are labeled with a specific hapten for subsequent fluorescent staining. Depending on the respective extended base and the wavelength of the fluorescent signal detected, the genotype for the desired SNP position can be determined for homozygous A or B, as well as heterozygous AB individuals. ( C ) Due to sequence paralogies, such as in RHD and RHCE , the 50 bp probes are not unique enough in the genomic context to exclusively bind their desired genomic fragments. As a result, signals detected by the camera constitute complex mixed signals from both paralogous loci as well as from different alleles present that cannot be easily interpreted.

Article Snippet: To decrease the complexity of our model and ensure convergence of the neural network given our sample size, we determined the allele discriminating SNPs out of the 1,500 SNPs typed on the array for the paralogous loci and evaluated if the B-allele frequency (BAF) in the Illumina raw genotyping data and genotypes of RhCDE- and MNS-SNPs correlated with the frequency of the reference antigen of our reference data.

Techniques: Sequencing, Microarray, Labeling, Staining

Journal: medRxiv

Article Title: Calling for diversity: improving transfusion safety through high-throughput blood group microarray genotyping

doi: 10.1101/2023.12.15.23299980

Figure Lengend Snippet: Concordance between blood group typing based on microarray genotyping data and respective gold standard reference methods. For ABO (striped background) a serological assay was used as a reference. All other antigens were reference typed with MALDI-TOF genotyping. Using our approach, MNS and some Rh antigens of paralogous loci (plaid background) were called using a deep-learning-based classifier, while the remaining antigens were called using static software logic. The concordance is shown as percentage in the center of the doughnut chart for 20 antigens and blood group systems (upper part of hexagons). Besides ABO and Rh, 12 other blood group systems were typed and benchmarked. For each antigen or system, the number of typing discordances and total benchmark tests is given in the lower part of the hexagons.

Article Snippet: To decrease the complexity of our model and ensure convergence of the neural network given our sample size, we determined the allele discriminating SNPs out of the 1,500 SNPs typed on the array for the paralogous loci and evaluated if the B-allele frequency (BAF) in the Illumina raw genotyping data and genotypes of RhCDE- and MNS-SNPs correlated with the frequency of the reference antigen of our reference data.

Techniques: Microarray, Serologic Assay, Software

Journal: European Journal of Human Genetics

Article Title: A comparison of genotyping arrays

doi: 10.1038/s41431-021-00917-7

Figure Lengend Snippet: Rotterdam Study HRC1.1 imputed were used for comparison with sequencing data and imputation quality. HapMap and 1KGPp3v5-imputed genotypes were used for imputation quality, clinically relevant genes, and GWAS catalog analyses. Theoretical analyses such as mtDNA and SNV density used the array manifest files as basis.

Article Snippet: To check for potential manifest bias, we additionally investigated the genotyping rate and imputation quality across genetic backgrounds, for which Illumina and Affymetrix kindly provided raw genotype data of HapMap samples ( N = 210) for a subset of the investigated arrays (Core, OmniExpress, GSAv1, PMRA, Global and Omni2.5).

Techniques: Sequencing