P073 Next generation sequencing of all classical HLA-class I and II genes

Abstract

Introduction Unambiguous, high resolution HLA typing plays a central role in hematopoietic stem cell transplantation. Current methodologies, primarily Sanger SBT, often fails to resolve all ambiguities creating the need for additional testing and longer turnaround times. Next generation sequencing (NGS) has the potential to address these shortcomings by providing full length, phase resolved sequences of HLA alleles with a single test. The Immucor Mia Fora NGS TM is an HLA typing kit and analysis software package that allows for typing of multiple samples per run in multiplex using the Illumina platform. In this study, we evaluated the NGS system against Sanger SBT methods. Methods NGS was performed on 11 HLA genes (HLA -A, B, C, DQA1/B1, DPA1/B1, and DRB1/3/4/5) for 69 subjects (1221 alleles) in alpha test, 69 (1223 alleles) in beta, and 35 (618 alleles) in further test. Automated liquid handling instruments were used for long range PCR and library preparation. Size selection was performed using the Pippin Prep followed by real-time PCR quantification on the ABI 7500. The final library was sequenced on the Illumina using the 300-cycle v2 kit. FASTQ files were analyzed with the Mia Fora software. NGS alleles to the 3rd fields were compared to the typing obtained by SBT, SSO, and SSP. Results A panel of 173 subjects carrying 3062 HLA class I and II alleles (total of 369 unique HLA alleles including novel alleles) was selected and typed by NGS. The overall concordance between NGS (automatic allele assigned by the software without editing) and Sanger SBT typing ranged from 98.5–99.7% in class I and 96.7–100.0% in class II. After manual review using the software’s tools for flagged alleles, the overall concordance will reach even higher. Conclusions The current NGS HLA typing product is able to produce robust and reliable allelic level HLA typing at all 11 HLA class I and II loci with no or very few ambiguities. Discordance was primarily observed in class II, most notably certain DRB1 ∗ 04 and DQB1 ∗ 03 alleles which often suffered from low read coverage in the presence of other alleles, and the ability to assign novel alleles or ambiguous results (e.g. DPB1) by the software. Improvements on test condition to prevent potential preferential amplification of certain HLA alleles would improve even more the accurate assignment of alleles at all HLA loci.