The human leukocyte antigen (HLA) distribution in donor registry data is typically nonrandom as, mostly for economical reasons, typing additional loci or resolving ambiguities is selectively performed based on the previously known HLA type. Analyzing a sample of over 1 million German stem cell donors, we practically show the extent of the bias caused by the restriction of the input data for HLA haplotype frequency (HF) estimation to subsets selected according to their higher HLA typing resolution and, conversely, the correctness of estimates based on unselected data with a methodology suitable for heterogeneous resolution. We discuss algorithmic aspects of this approach and, also because of the sample size, provide some new insights into the distribution of HLA-DRB1 alleles in the German population and the application of HFs in unrelated donor search.

Europeans have often been considered a homogenous group in registry donor match predictions, but it is now evident that HLA haplotype frequencies vary across the European continent. Earlier studies have indicated that Finns in northeastern Europe have unique HLA characteristics, and the increasing availability of high-resolution registry donor data is now making more detailed comparisons possible. In the first phase of the present study, estimated HLA haplotype frequencies in stem cell donor registries of Finland and its neighbors Sweden and Russia were calculated using the algorithm of the German National Bone Marrow Donor Registry (ZKRD) and their frequencies were compared with one another and also with that of Germany. Virtual donor searches for 1492 high-resolution typed Finnish patients in the Finnish, Swedish and German registries were then performed, using individual match predictions for each registry. In the last phase, the impact of specifically Finnish-enriched HLA haplotypes on Finnish patients and the use of Finnish registry donors was assessed by analyzing 647 consecutive hematopoietic stem cell transplantation (HSCT) donor searches and 40 exported Finnish HSCTs. The Finnish HLA landscape was more homogenous than the 3 other studied populations, but also genetically distinct from them. The match predictions found a probable 10/10 match for 71%, 41%, and 31% of the Finnish patients in the German, Finnish, and Swedish registries, respectively. Thirty-four of Finland's 100 most frequent HLA haplotypes were represented with a frequency of <.0003 in Germany, and with an 8- to 3262-fold greater frequency in Finland than in Germany. Patients carrying these Finnish-enriched haplotypes were less likely to receive a matched HSCT but more likely to receive it from a domestic donor. Registry donors carrying them were more likely to donate stem cells, both nationally and internationally. The Finnish HLA isolate has a significant impact on both Finnish patients and registry donors, explaining the high use of national registry donors for Finnish patients. Haplotype frequency estimations are an important tool for small registries as well, to help optimize donor match predictions and the size of individual registries.

We present a catalog of common and well‐documented (CWD) alleles of the German population for the six HLA loci A, B, C, DRB1, DQB1, and DPB1. This study is based on a sample of over 5 million volunteer adult hematopoietic stem cell donors from the 26 German donor centers. To establish the catalog, allele and haplotype frequencies were estimated with a validated implementation of the expectation‐maximization algorithm. CWD criteria similar to existing CWD catalogs were applied in order to be able to put our findings into the context of relevant existing references. Overall, 2155 HLA‐A, ‐B, ‐C, ‐DRB1, ‐DQB1, and ‐DPB1 alleles were identified as CWD in the German donor population representing about 20% of the HLA alleles at two‐field resolution in the IPD‐IMGT/HLA Database release v3.25.0 from July 2016 for these six loci. We found a substantial concordance of CWD alleles between the three catalogs and showed the contribution of the German donor population to the CWD alleles domain. In conclusion, the definition of CWD criteria that allow interoperability, scalability, and flexibility will be crucial for the development of a worldwide CWD catalog.

The accuracy of human leukocyte antigen (HLA)‐matching algorithms is a prerequisite for the correct and efficient identification of optimal unrelated donors for patients requiring hematopoietic stem cell transplantation. The goal of this World Marrow Donor Association study was to validate established matching algorithms from different international donor registries by challenging them with simulated input data and subsequently comparing the output. This experiment addressed three specific aspects of HLA matching using different data sets for tasks of increasing complexity. The first two tasks targeted the traditional matching approach identifying discrepancies between patient and donor HLA genotypes by counting antigen and allele differences. Contemporary matching procedures predicting the probability for HLA identity using haplotype frequencies were addressed by the third task. In each task, the identified disparities between the results of the participating computer programs were analyzed, classified and quantified. This study led to a deep understanding of the algorithms participating and finally produced virtually identical results. The unresolved discrepancies total to less than 1%, 4% and 2% for the three tasks and are mostly because of individual decisions in the design of the programs. Based on these findings, reference results for the three input data sets were compiled that can be used to validate future matching algorithms and thus improve the quality of the global donor search process.

Estimation of human leukocyte antigen (HLA) haplotype frequencies from unrelated stem cell donor registries presents a challenge because of large sample sizes and heterogeneity of HLA typing data. For the 14th International HLA and Immunogenetics Workshop, five bioinformatics groups initiated the 'Registry Diversity Component' aiming to cross-validate and improve current haplotype estimation tools. Five datasets were derived from different donor registries and then used as input for five different computer programs for haplotype frequency estimation. Because of issues related to heterogeneity and complexity of HLA typing data identified in the initial phase, the same five implementations, and two new ones, were used on simulated datasets in a controlled experiment where the correct results were known a priori. These datasets contained various fractions of missing HLA-DR modeled after European haplotype frequencies. We measured the contribution of sampling fluctuation and estimation error to the deviation of the frequencies from their true values, finding equivalent contributions of each for the chosen samples. Because of patient-directed activities, selective prospective typing strategies and the variety and evolution of typing technology, some donors have more complete and better HLA data. In this setting, we show that restricting estimation to fully typed individuals introduces biases that could be overcome by including all donors in frequency estimation. Our study underlines the importance of critical review and validation of tools in registry-related activity and provides a sustainable framework for validating the computational tools used. Accurate frequencies are essential for match prediction to improve registry operations and to help more patients identify suitably matched donors.