Gerhard Schöfl

2.7 million samples genotyped for HLA by next generation sequencing: lessons learned

BackgroundAt the DKMS Life Science Lab, Next Generation Sequencing (NGS) has been used for ultra-high-volume high-resolution genotyping of HLA loci for the last three and a half years. Here, we report on our experiences in genotyping the HLA, CCR5, ABO, RHD and KIR genes using a direct amplicon sequencing approach on Illumina MiSeq and HiSeq 2500 instruments.ResultsBetween January 2013 and June 2016, 2,714,110 samples largely from German, Polish and UK-based potential stem cell donors have been processed. 98.9% of all alleles for the targeted HLA loci (HLA-A, -B, -C, -DRB1, -DQB1 and -DPB1) were typed at high resolution or better. Initially a simple three-step workflow based on nanofluidic chips in conjunction with 4-primer amplicon tagging was used. Over time, we found that this setup results in PCR artefacts such as primer dimers and PCR-mediated recombination, which may necessitate repeat typing. Split workflows for low- and high-DNA-concentration samples helped alleviate these problems and reduced average per-locus repeat rates from 3.1 to 1.3%. Further optimisations of the workflow included the use of phosphorothioate oligos to reduce primer degradation and primer dimer formation, and employing statistical models to predict read yield from initial template DNA concentration to avoid intermediate quantification of PCR products. Finally, despite the populations typed at DKMS Life Science Lab being relatively homogenous genetically, an analysis of 1.4 million donors processed between January 2015 and May 2016 led to the discovery of 1,919 distinct novel HLA alleles.ConclusionsAmplicon-based NGS HLA genotyping workflows have become the workhorse in high-volume tissue typing of registry donors. The optimisation of workflow practices over multiple years has led to insights and solutions that improve the efficiency and robustness of short amplicon based genotyping workflows.

ABO allele-level frequency estimation based on population-scale genotyping by next generation sequencing

BackgroundThe characterization of the ABO blood group status is vital for blood transfusion and solid organ transplantation. Several methods for the molecular characterization of the ABO gene, which encodes the alleles that give rise to the different ABO blood groups, have been described. However, the application of those methods has so far been restricted to selected samples and not been applied to population-scale analysis.ResultsWe describe a cost-effective method for high-throughput genotyping of the ABO system by next generation sequencing. Sample specific barcodes and sequencing adaptors are introduced during PCR, rendering the products suitable for direct sequencing on Illumina MiSeq or HiSeq instruments. Complete sequence coverage of exons 6 and 7 enables molecular discrimination of the ABO subgroups and many alleles. The workflow was applied to ABO genotype more than a million samples. We report the allele group frequencies calculated on a subset of more than 110,000 sampled individuals of German origin. Further we discuss the potential of the workflow for high resolution genotyping taking the observed allele group frequencies into account. Finally, sequence analysis revealed 287 distinct so far not described alleles of which the most abundant one was identified in 174 samples.ConclusionsThe described workflow delivers high resolution ABO genotyping at low cost enabling population-scale molecular ABO characterization.

Prediction of spurious HLA class II typing results using probabilistic classification.

While modern high-throughput sequence-based HLA genotyping methods generally provide highly accurate typing results, artefacts may nonetheless arise for numerous reasons, such as sample contamination, sequencing errors, read misalignments, or PCR amplification biases. To help detecting spurious typing results, we tested the performance of two probabilistic classifiers (binary logistic regression and random forest models) based on population-specific genotype frequencies. We trained the model using high-resolution typing results for HLA-DRB1, DQB1, and DPB1 from large samples of German, Polish and UK-based donors. The high predictive capacity of the best models replicated both in 10-fold cross-validation for each gene and in using independent evaluation data (AUC 0.820-0.893). While genotype frequencies alone provide enough predictive power to render the model generally useful for highlighting potentially spurious typing results, the inclusion of workflow-specific predictors substantially increases prediction specificity. Low initial DNA concentrations in combination with low-volume PCR reactions form a major source of stochastic error specific to the Fluidigm chip-based workflow at DKMS Life Science Lab. The addition of DNA concentrations as a predictor variable thus substantially increased AUC (0.947-0.959) over purely frequency-based models.

P103 The promise of cost-efficient full-length HLA class I genotyping: Advances using nanopore sequencing

Aim Nanopore based sequencing has seen rapid advancement in the recent years with iterations in pore structure, sequencing chemistries and base callers. This has led to increasingly accurate sequencing of extremely long DNA molecules. Since current HLA genotyping algorithms are not optimized for nanopore data we developed a genotyping algorithm based on read grouping and subsequent mapping to a reduced reference database. Methods High raw sequencing errors present a challenge for genotyping based on direct mapping to a reference database. Our algorithm, Poretyper, obviates this initial mapping by first grouping the raw reads based on the distributions of k-mers within each read. A multiple sequence alignment derived from the groups of raw reads results in a set of consensus sequences which represent potential alleles. These consensus sequences are then used to create a culled reference database dramatically reducing the search space, thus reducing artefactual raw read mappings. Results HLA-A, HLA-B and HLA-C for a hundred samples drawn from the DKMS donor registry were sequenced using the MinION with R9.5 chemistry. All hundred samples were then genotyped using Poretyper and existing G-group pretypings could be recapitulated failing only for those alleles where no full length sequences were available. Conclusions Nanopore sequencing presents a viable and accurate platform for cost-efficient full-length HLA Class I genotyping. For those alleles where full length reference sequences are not available, an in silico extension of such allele sequences using the full-length sequence of the next closest allele presents a viable approach for full-length genotyping.

P061Allele-level genotyping of KIR2DL4 in large european population samples reveals highly significant heterozygote excess for 9A/10A allelic variants

Aim KIR2DL4 is an evolutionarily conserved framework member of the human killer-cell immunoglobulin-like receptor (KIR) gene family. It is unique amongst KIR genes in that it may elicit both activation and inhibition signals. Moreover, KIR2DL4 alleles are polymorphic for a frameshift mutation in the transmembrane domain that leads to a truncated cytoplasmatic tail. Alleles with a 10A homopolymer in exon 7 encode receptors that are expressed on the cell surface of NK cells. In contrast, alleles with a 9A frameshift mutation have been shown to produce soluble secreted KIR2DL4 receptors. Here, we investigate allele and genotype frequencies of 9A and 10A alleles in large European population samples. Methods In 2016, DKMS Life Science Lab has established an exon-based NGS workflow for KIR genotyping. Between 09/2017 and 03/2018, we have performed successful allelic-resolution genotyping for KIR2DL4 for approximately 380,000 potential bone-marrow donor samples originating from Germany (DE), Great Britain (GB), and Poland (PL). These samples where checked for presence of the 9A frameshift mutation. Results Overall, the defective 9A variant was slightly underrepresented in DE ( AF 9A  = 0.496, n  = 242,983) and GB ( AF 9A  = 0.475, n  = 31,100) but overrepresented in the Polish sample ( AF 9A  = 0.537, n  = 101,998). The fraction of homozygous individuals (10A/10A or 9A/9A) ranges from 44.4% (DE) to 44.9% (PL). Comparing expected and observed genotype frequencies indicates a highly significant deviation from Hardy–Weinberg Equilibrium ( P −9 ), and a consistent heterozygote excess across all three populations, with inbreeding coefficients ranging from FIS  =  −0.104 (GB) to −0.115 (PL). Conclusions The high frequency of 10A alleles in all three populations indicates no negative selection against a lack of cell surface expression of KIR2DL4. Rather, the strong signature of heterozygote excess across populations may best be explained as a result of overdominant selection (i.e., “heterozygous advantage”). This suggests that the presence of both, cell-surface-expressed KIR2DL4 receptors and soluble secreted KIR2DL4 receptors confers a selective advantage to humans.

P062Coding and non-coding variation in 60 novel full-length sequences of KIR2DL1

Aim KIR2DL1 is a member of the killer-cell immunoglobulin-like receptor (KIR) family, which is an important factor of the human immune system. KIR genes are key regulators of natural killer cell activity and partly bind to proteins of the human leukocyte antigen (HLA) family, e.g. HLA-B and -C in the case of KIR2DL1. Likely due to the complexity of the KIR locus with its extensive genetic variation, only little is known about the impact of allelic variation. Here, we identified novel KIR2DL1 alleles by routine high-throughput exon-based KIR genotyping and subsequently created full-length reference sequences. We analysed coding and non-coding variation to identify possible systematic differences between alleles. Methods We sequenced whole 16 kb amplicons of KIR2DL1 using shotgun sequencing (Illumina MiSeq) and single molecule real time (SMRT) sequencing (PacBio Sequel). Using the R package DR2S, we combined phase information from SMRT sequencing with the accuracy of shotgun sequencing to generate phased full-length sequences. Results We successfully generated 60 distinct KIR2DL1 allele sequences from 45 specifically selected samples. This includes 41 novel alleles and 17 distinct alleles that were previously only partially characterised. The sequence analysis revealed three deep allelic groups separated by 164 variable positions. Allelic variants of two groups harbour mutually exclusive nucleotides, while a third group of alleles may harbour nucleotides of both the other groups. The three groups are mostly separated by non-coding SNPs, indels, and a T homopolymer of fixed length in one and variable length in two groups. Some separating variants exist in most exons and alter the amino acid sequence of the leader peptide, the extracellular D1- and D2-domains containing the HLA binding sites as well as the cytoplasmatic domain. Conclusions Our sequencing efforts resulted in a 4-fold increase in known full-length sequences of KIR2DL1, enabling further research on this KIR gene. We gained insights into systematic differences at the sequence level which might be responsible for or indicative of medically relevant allelic differences. Especially variation in the D2-domain has been shown to be involved in binding to HLA proteins and may as such be clinically relevant.

OR38 Whole gene HLA class I genotyping with the minion nanopore sequencer

In 2014 Oxford Nanopores Technologies released the MinION, a portable miniaturized sequencer delivering multi-kb DNA reads based on nanopore sensing technology. Sequencing accuracy is improved by consecutive sequencing both strands of the DNA (2D reads). Being part of the MinION early access program we evaluated this new technology for whole gene HLA sequencing. As the sequencing yield exceeds the requirements for a targeted approach we developed a strategy for barcoding and demultiplexing of samples. Barcoded amplicons were generated by two subsequent long range PCRs covering the complete genes of HLA-A, -B and -C. Amplicons were pooled and library preparation was performed according to the Oxford Nanopore protocol (SQK-MAP006). Sequencing was performed on a MinION MKI with R7.3 flow cells. NGSengine from GenDX was used for sequence analysis and HLA genotyping. 96 randomly picked class I samples were pooled and sequenced. Over 170,000 reads were generated, approximately 65% of them as 2D reads. Two thirds of the 2D reads could be assigned to a unique barcode sequence. The HLA genotyping results were compared to allelic resolution genotypes obtained by full length sequencing the samples on both Illumina and PacBio instruments. The ONT typing results had a 95% concordance rate when considering 1st to 3rd field. Even when taking the fourth field in account concordance rate was above 90%. Given that the nanopore technology is rapidly evolving and the genotyping software has not yet been optimized for nanopore data, this study clearly underlines the great potential of the MinION for high resolution HLA genotyping.