Failure to detect mutations U2AF1 due to changes in the GRCh38 reference sequence

Abstract

The U2AF1 gene is a core part of mRNA splicing machinery and frequently contains somatic mutations that contribute to oncogenesis in MDS, AML, and other cancers. A change introduced in the GRCh38 version of the human reference build prevents mutations in this gene from being detected by many variant calling pipelines. We describe the problem in detail and show that a modified GRCh38 reference build with unchanged coordinates can be used to ameliorate the issue. This reference is available at https://zenodo.org/record/4684553 (doi:10.5281/zenodo.4684553)