Benchmark of tools for in silico prediction of MHC class I and class II genotypes from NGS data

The Human Leukocyte Antigen (HLA) genes are a group of highly polymorphic genes that are located in the Major Histocompatibility Complex (MHC) region on chromosome 6. The HLA genotype affects the presentability of tumour antigens to the immune system. While knowledge of these genotypes is of utmost importance to study differences in immune responses between cancer patients, gold standard, PCR-derived genotypes are rarely available in large Next Generation Sequencing (NGS) datasets. Therefore, a variety of methods for in silico NGS-based HLA genotyping have been developed, bypassing the need to determine these genotypes with separate experiments. However, there is currently no consensus on the best performing tool. We evaluated 13 MHC class I and/or class II HLA callers that are currently available for free academic use and run on either Whole Exome Sequencing (WES) or RNA sequencing data. Computational resource requirements were highly variable between these tools. Three orthogonal approaches were used to evaluate the accuracy on several large publicly available datasets: a direct benchmark using PCR-derived gold standard HLA calls, a correlation analysis with population-based allele frequencies and an analysis of the concordance between the different tools. The highest MHC-I calling accuracies were found for Optitype (98.0%) and arcasHLA (99.4%) on WES and RNA sequencing data respectively, while for MHC-II HLA-HD was the most accurate tool for both data types (96.2% and 99.4% on WES and RNA data respectively). The optimal strategy for HLA genotyping from NGS data depends on the availability of either WES or RNA data, the size of the dataset and the available computational resources. If sufficient resources are available, we recommend Optitype and HLA-HD for MHC-I and MHC-II genotype calling respectively.