Improvement of HLA resolution with new LabType® SSO XR kits
G. Guidicelli (Cedex, FR)
PCR-SSO is the most commonly used assay for HLA-typing in solid organ transplantation. Our laboratory used the OneLambda LabType® SSO HD (HD) kits until January 2016 and changed to the new LabType® SSO XR (XR) kits for HLA-A, -B and -DRB1 typing after purchasing a LABScan3DTM. The XR kits have an increased number of bead regions in comparison with HD kits (500 vs 100), which allows the number of sequence specific oligonucleotides and covered exons to be enlarged. Also, the XR kits are claimed to improve typing resolution by a diminished number of allele ambiguities, including null alleles that need to be resolved. The aim of this study was to measure the improvement in typing provided by XR kits. We retrospectively selected patients who had their first HLA-typing determination performed with the HD kit from September 2015 to January 2016, and their second determination with the XR kit from January to August 2016. Results were analyzed against the IPD-IMGT/HLA Database v3.21.1 (2015-8). Forty-two patients were selected for HLA-A typing comparison, 50 for HLA-B and 41 for HLA-DRB1. For the 3 loci, the number of ambiguities generated by the two kits that concerned exons 1, 2, 3, 4 and 5 were compared. XR kits significantly decreased ambiguities for HLA-A (65% for those concerning exons 2 and 3, 95% exon 4 and none for exon 5) and -B (48%, 51%, 42%, 90%, 95% for exon 1, 2, 3, 4, 5, respectively), but just moderately for HLA-DRB1 locus (14% for those concerning exon 2). No ambiguities were resolved for other DRB1 exons as only exon 2 was amplified. The XR kits allowed eliminaion of 88%, 61% and 26% of potential null alleles for HLA-A, -B and -DRB1 loci, respectively. Therefore, the XR kit allows a significant improvement of HLA-typing resolution for HLA-A and -B loci in comparison with HD kits. In contrast, an improvement of XR HLA-DRB1 kit is required through at least studying exon 3. It could also include oligonucleotides that allow HLA-DRB3,4 and 5 typing
Implementing ABO genotyping into HLA sequencing workflow
A. Hansen (Birkenfeld, DE)
The ABO system is the most important blood group system in terms of transfusion medicine lately gaining more relevance in donor/recipient matching in the transplantation context. For decades, serological phenotyping has bee the most common and cost efficient method to determine the ABO blood group phenotype. ABO genotyping is even nowadays restricted to serologically ambiguous results or cases with limited availability of red blood cells. The frequently observed discrepancies between serological and molecular genetic typing results can be ascribed to the minimalistic way of ABO genotyping usually based on a limited number of SNPs residing in exon 6 and 7. In order to minimize these inconsistencies we developed a NGS based method to sequence the entire length of the ABO gene. Starting with the design of a long range PCR assay we finally succeeded generating large fragments covering the complete ABO gene including the regulatory enhancer region located 3.7 kb upfront of exon 1. The overlapping amplicons in the size range from 8-12 kb were basically sequenced according to standard NGS protocols on a MiSeq platform (Illumina). Developing and integrating an ABO analysis tool (HLA Twin, Omixon) turned out to be the major hurdle mainly hampered by the lack of useful reference databases and scarce sequence information especially in the remaining exons beside 6 and 7 and the non-coding regions. To prove the routine functionality of the established long range PCR assay and ABO software module integration we processed 300 serologically pre typed samples. With the described workflow we are able to sequence the ABO gene in its full length delivering ABO genotypes with so far unrivaled resolution not at last to fill the gaps in the reference databases. The most striking advantage for a high throughput registry typing laboratory is the avoidance of a serological ABO typing workflow in parallel.
Nanopore sequencing is coming of age: a new analysis platform for HLA allele assignment based on full-length de novo assembly.
B. Matern (Maastricht, NL, NL)
In the years past, MinION has often been dismissed as a useful platform for clinical diagnostics, citing high per-base error rates and lack of analysis tools as the main apprehensions. Recently, the technology has advanced rapidly, and we regularly observe reads with an average Phred score of Q=20, corresponding to 99.0% per-base accuracy. Oxford Nanopore releases regular updates on the sequencing and basecalling platforms, and members of the MinION community continually develop improvements on analysis and assembly tools. De novo assembly of nanopore reads remains a challenge, but data analysis has reached a point where reads are useable in analysis of complicated polymorphic regions, such as HLA. Our goal is to develop a high-throughput, highly accurate open-source analysis tool for HLA allele calling based on initial de novo assembly, followed by alignment to sequences from the IPD-IMGT/HLA Database. This analysis pipeline uses both community and in-house software in a structured way to sort, assemble, and classify high-quality Nanopore reads for applications in research, and ultimately routine diagnostics. As a proof of concept, we sequenced the full-length classical class I HLA genes (A,B,C) of three individuals with potentially novel HLA types. The MinION reads were analyzed, filtered for quality and length, and assembled into consensus sequences for each of the three class I loci. The sequences were aligned to allele references available in the IPD-IMGT/HLA Database for comparison with known alleles. For automated submission of novel full-length HLA alleles to the EMBL-ENA Database, an in-house allele submission tool was generated, and is available for general public use. Overall, the developed analysis tool represents a promising method for analysis of MinION data and assignment of HLA alleles based on full-gene Nanopore sequences.
What’s new genotyping KIR2DL5?
L. Krammes (Ellenberg, DE)
Killer-cell immunoglobulin-like receptors (KIR) expressed on NK cells are specialized in recognition of MHC lacking cells especially tumor and virus infected cells. In the last few years there is growing evidence that distinct KIR donor haplotypes lead to a better outcome of stem cell transplantations. Therefore extending donor profiles with KIR genotypes is desirable in order to improve donor/recipient matching. The inhibitory KIR2DL5 is a key gene to differentiate between the so far classified KIR haplotypes A and B. Most actual KIR genotyping methods unravel the presence or absence of the respective gene disregarding the variations of the present alleles. To investigate the impact of higher genotype resolution we established a NGS based typing strategy. Designing a long range PCR allowing a full-length gene amplification was the first step to implement KIR2DL5 genotyping into our NGS typing routine based on the Illumina MiSeq platform. During the initial testing of this KIR2DL5 NGS assay on 48 randomly selected samples the data analysis (HLA Twin software, Omixon) revealed several new variations in non-coding regions as well as an unknown polymorphism in exon 5. The C to A substitution at position 2860 leads to an amino acid change in codon 143 from proline to threonine. Validation of this so far undescribed variant was performed Sanger sequencing of exon 5 on a 3730XL DNA Analyzer (Applied Biosystems) with subsequent analysis using GeneStudio Professional software. The Sanger sequencing results confirmed the observed nucleotide exchange of this novel KIR2DL5 sequence variant. This observation highlights the importance of full-length, sequenced based KIR typing methods to improve knowledge about the variability of the KIR genes.
An evaluation of PAKLx™ – a Luminex®-based assay for the detection of anti-human platelet antibodies.
J. Bestwick (Pontyclun, GB)
The ‘monoclonal antibody immobilisation of platelet antigens assay’ (MAIPA) is considered the ‘gold standard’ for the identification of alloantibodies to human platelet antigens (HPA) in patients with alloimmune platelet disorders. However, it is technically demanding, requires typed reference platelet panels and has a long turnaround time. New technologies, i.e. the Luminex-based PAKLx assay (Immucor, Lifecodes) are now available for the detection of IgG antibodies to HPA-1, 2, 3, 4, 5 epitopes and ‘broad’ platelet glycoproteins (GP). We tested 48 sera from cases of suspected FNAIT (n=26), MPTR (n=5), PTP (n=1), heredity platelet function disorder (n=1), non-specified immune thrombocytopenia (n=12) and healthy controls (n=3), by the MAIPA and PAKLx assays in parallel. Doubling dilution titrations were also performed, to determine the comparative sensitivity of PAKLx, using three NIBSC reference reagents (05/106 - anti-HPA-1a, 03/190 - HPA-3a and 99/666 - HPA-5b). The MAIPA assay identified HPA and platelet GP antibodies in 32 of the 48 (66.7%) sera: HPA-1a (n=18), HPA-1b (n=4), HPA-2b (n=1), HPA-3a (n=1), HPA-5b (n=7), GPIIb/IIIa (n=3). PAKLx detected all antibodies found by the MAIPA and a further three, not MAIPA detected, in the 32 sera, viz., two broad GP-reactive and one HPA-2b. This gave the PAKLx assay a sensitivity of 100% and specificity of 98.8%. Titration studies showed that the PAKLx assay was more sensitive for the detection of HPA-1a antibodies, while the MAIPA showed increased sensitivity for HPA-3a and HPA-5b antibodies. In conclusion, the PAKLx assay is fast and easy to perform. Its superior sensitivity to detect anti-HPA-1a, implicated in 70-80% of FNAIT cases in the Caucasoid population, supports its routine use in initial laboratory investigations. However, due to its limitations of reduced sensitivity for HPA-3a and HPA-5b antibodies and its inability to detect HPA-15 antibodies, the PAKLx assay should not be used as a standalone method.
Exon phasing permits identification of new alleles by NGS not detectable by Sanger sequence-based typing
X. Lafarge (Bordeaux, FR)
A complementary typing on a bone marrow donor was firstly performed by bi-allelic sequence based typing (SBT, ALLELE SEQR kits, CELERA). The ASSIGN software (version 3.6, CONEXIO and IPD-IMGT/HLA Database 184.108.40.206, 2016-01-19) proposed interpretable results with cis-trans ambiguities: A*02:01:01:06, *11:172 / A*02:01:02, *11:01:34. PCR SSP technique (Olerup HLA-A*11 kits) excluded the allele A*11:172, resulting in the following typing: A*02:01:02, *11:01:34. Because the presence of these alleles seemed unlikely, we verified this result by NGS with HOLOTYPE HLA kits (OMIXON), and found a discordant result: A*02:NEW, *11:01:01:01. This new A*02 allele was characterized by a substitution of C>T in position 228.2 compared to A*02:01:01, accompanied by a amino acid modification (threonine > isoleucine). This substitution is what distinguishes an A*11:172 from an A*11:01:01:01. During sequence interpretation, exon phasing permits us to link distant polymorphic positions present in different exons, which is not possible with most of the classical SBT reagents. This example illustrates the SBT limitations because of the cis/trans ambiguities. For this individual, these ambiguities led to a wrong result because if two alleles already described can be interpreted, the new allele is undetectable. This limitation has been eliminated with the use of NGS provided that the introns are characterized and permitted to distinguish both alleles. Thus, combinations of two rare alleles found in an individual by SBT should be verified by an alternative technique.
Whole gene sequence determination of a rare human leukocyte antigen DRB1 allele by combining long range polymerase chain reaction and next generation sequencing.
T. Binder (Hamburg, DE)
Some years ago, we found a new HLA-DRB1*13 allele in a German family. It was found in a female potential stem cell donor, in two of her three children and in her mother. This allele, later called DRB1*13:67, seemed to be a hybrid allele between DRB1*13 and DRB3, at least in exon 2. To clarify the degree of recombination inside this allele, we developed a workflow based on long range PCR (LR-PCR) and next generation sequencing (NGS) on a MiSeq platform (Illumina) to provide the complete genomic sequence of this allele. Therefore, we designed different HLA locus and/or allele specific LR-PCRs. After amplicon generation by PCR and NGS data evaluation with two different HLA software tools (Omixon Twin, Omixon and NGSengine, GenDx) we ended up with allele-specific contigs based on phased sequence alignment according to the individual single nucleotide variants (SNVs) pattern present. The final alignment of these contigs was performed with the AliView (Muscle) and BioEdit (ClustalW) software along with published genomic sequences (IPD-IMGT/HLA Database). With our approach, applying LR-PCR and NGS including phased sequence analysis, we were able to determine the complete gene sequence of HLA-DRB1*13:67 (from 5'-UTR to 3'-UTR). The recombination between the DRB1*13 and the DRB3 allele, seems to be restricted to the first part of exon 2. Our approach, to combine these methods, was very successful to get the complete gene sequence of this rare HLA allele.
Comparison of three commercial kits (STR PP16, STR Mentype Chimera and InDel Mentype DIPScreen) usable for chimerism determination by fragments analysis
X. Lafarge (Bordeaux, FR)
New multpiplex PCR kits (Biotype diagnostic) CE-marked for chimerism determination using fragments analysis, performed by capillary electrophoresis, are commercially available. They must be evaluated and compared to the techniques routinely used. Three multiplex PCR were compared: STR Promega PP16, and Biotype diagnostic STR Mentype Chimera° and InDel Mentype DIPScreen kits using insertion/deletion polymorphism. PCRs were performed according to manufacturer’s instructions. Amplicons were denaturated by formamide and migrated in polyacrylamide gel simultaneously with a fluorescent ladder. Results generated by Biotype diagnostic reagents were analyzed with ChimerisMonitor 2.0. This dedicated software proposes informative loci for % recipient calculation, and uses areas under peaks for the selected loci. These 3 kits were preliminary compared with 3 donor/recipient couples with different ADN mixes at 25, 5 and 1% recipient.
Our work showed that all kits permitted manipulation with the same practicability. In addition, the results obtained are similar between the 3 kits on the tested samples. The commercial kits by Biotype diagnostic seems to be as appropriate as PP16 Promega kits for chimerism determination. A more extended evaluation should be performed, in particular concerning the informativity of the loci included in the kits and the Z scores obtained on external proficiency testing samples.
How can NGS technology help you in genotyping of "tricky samples"?
L. Kolesar (Budapest, HU)
Sequencing based typing (SBT) has become a gold standard for high resolution HLA genotyping. With an increasing number of known HLA alleles, the number of ambiguous results has continued to increase – a major challenge for the SBT method. Next generation sequencing (NGS) methods provide a new sequencing approach to HLA laboratories for ultra-high resolution typing. Holotype HLA and its associated software, HLA Twin, gives tissue typers a new tool for HLA genotyping on an allelic level with almost unambiguous results. Here we show, with several examples, the power of NGS for HLA typing by identifying a novel stop codon and a new splice site. By displaying “noisy reads” that were removed during the analysis of the two alleles in the “noise track” it was possible to detect a triploid sample and separate a contamination in the sample. Using the amino acid display track we could immediately identify if our novel allele creates aberrant protein. NGS techniques continue to evolve rapidly and the trends are clear that they will soon become a routine method in almost every tissue typing laboratory. Holotype HLA and HLA Twin facilitate this transition to a new era of high-resolution, ambiguity-free, accurate HLA genotyping.