P110 Resolution potential of KIR typing based on high-throughput NGS genotyping data

Abstract

Aim Recently, we implemented an amplicon based next generation sequencing (NGS) workflow for routine KIR typing, which currently yields presence/absence calls for all KIR genes. While presence/absence data is a good starting point, donor selection might further benefit from allele level resolution. In the current setting, high resolution KIR typing results could, however, hardly be achieved as sequencing relies on exons 4, 5 and 7 only. We therefore analyzed whether expanding the sequenced regions by exons 3, 8 and 9 would significantly reduce allele level ambiguities. Methods Taking previously published KIR haplotype frequency data into account, we randomly generated 1000 artificial KIR genotypes, each consisting of two KIR haplotypes. Based on the IPD-KIR database for each genotype the exon specific sequence compositions were calculated and fed back into an algorithm to recalculate the genotype. We then compared the level of allelic resolution that could be achieved when using exons 4, 5, 7 alone or in combination with exons 3, 8, 9. Here the number of alleles that could not be resolved at 3-digit nomenclature determines the remaining level of ambiguity (LA), where an LA of one refers to an unambiguous typing. Results When performing the recalculation based on exon 4, 5 and 7 alone, we are able to achieve an overall low to medium allelic resolution, with an LA ranging from 1 to 16. We found that the majority of cases, with LA >1 are due to sequence coverage limitations rather than missing phase information. With LAs ranging from 1 to 4, the recalculation based on exons 3, 4, 5, 7, 8 and 9 showed a significantly higher overall allelic resolution. Particularly KIR2/3DL alleles were shown to benefit from additional exons. Interestingly, when analyzing the reasons for remaining ambiguities there is a balance between phasing and sequence coverage issues. Conclusion Expanding the sequenced regions by exons 3, 8 and 9 significantly reduces the KIR allele level ambiguity and thus enables high resolution KIR typings. In order to improve donor selection as well as transplantation outcome we will add exons 3, 8, 9 to our standard KIR typing workflow this summer.