NEW YORK (GenomeWeb) – Researchers at the University of Kiel in Germany have developed a next-gen sequencing-based HLA-typing and analysis pipeline that they say provides an alternative to commercially available solutions at a lower cost per sample.
The method, published earlier this month in Nucleic Acids Research, combines hybridization-based target capture with Illumina sequencing and open-source allele-calling software.
According to Andre Franke, director of the Institute of Clinical Molecular Biology at the University of Kiel and the senior author of the study, this team's method is unique in that all protocols are open access and the software is open source. "HLA typing is big business, and the procedures that have been published so far either lack good software for the analysis or the authors did not disclose details of the procedure, for example, the primer sequences," he told GenomeWeb.
Several companies have shown an interest in the NGS-based HLA-typing market. Illumina, for example, launched a TruSight HLA Sequencing Panel for the MiSeq earlier this month. Also, last year Pacific Biosciences and Dutch HLA-typing firm GenDx signed a co-marketing agreement for HLA sequencing, and Sirona Genomics said last fall that it is developing HLA-typing technology.
Franke's team set out to develop its method almost three years ago, initially for an EU-funded project that required them to sequence the HLA complex in patients with a chronic inflammatory disease.
At the time, he said, no commercial products for preparing the HLA genes for next-gen sequencing were available. His team obtained early access to panels from RainDance Technologies and Agilent, but neither captured the full complexity of the HLA region. "The designs worked perfectly when we worked with samples [containing HLA alleles from the] reference genome, but when we used real-life patient samples, we saw allelic dropout," he said.
Instead, they developed their own capture design, a panel of RNA baits that covers all known HLA alleles from the IMGT/HLA database, though their data analysis focuses on eight loci: HLA-A, HLA-B, HLA-C, HLA-DRB1, HLA-DQA1, HLA-DQB1, HLA-DPA1, and HLA-DPB1.
To validate their panel, they tested it on commercially available DNA from 357 cell lines with a variety of known HLA types. They automated the capture and library preparation using an Agilent SureSelectXT kit and sequenced the enriched DNA on an Illumina HiSeq 2000, pooling 48 samples per lane.
The analysis was performed using a fully automated calling algorithm, developed in house by Michael Wittig, which achieved allele call rates between 0.97 and 1 for the test samples, depending on the locus. The researchers were able to improve these rates to 100 percent concordance using visual inspection and manual correction.
The total cost is on the order of €100 ($109) per sample, most of it associated with sample preparation. According to Franke, this is about three times less than the list price per sample for Illumina's TruSight HLA Sequencing Panel. The turnaround time is more than a week, mostly due to the length of the HiSeq sequencing run, but it could be reduced to a few days by using the MiSeq sequencer instead, which would increase the cost per sample marginally.
Illumina's TruSight HLA Sequencing Panel covers a total of 11 HLA loci, which are amplified by long-range PCR and sequenced on the MiSeq system. Data are analyzed using the Conexio Assign software, and the entire process takes 3.5 days, according to the company.
Franke said the overall workflow of his team's method is simpler than that of Illumina's because it does not involve PCR, but the data analysis is slightly more complex because of the high off-target capture rate.
Another advantage of the capture-based method is that additional targets – for example, for blood groups or non-HLA genes of interest – can be included easily by adding baits, without increasing the overall cost by much.
The Conexio software that Illumina provides for its panel is the most popular software for traditional HLA typing, so many users are very familiar with it, but it is also expensive, Franke said, whereas his group's software is open source and runs on different operating systems.
While their software focuses on typing new alleles, Illumina's, which allows for mismatches in the alignment, can potentially discover novel alleles. However, "it's pretty rare to find something novel unless you go to different ethnicities," Franke said. "Our software can hint at these novel alleles, but you then have to follow this official workflow on how to describe novel HLA alleles."
Since submitting their method for publication, the researchers have made a number of improvements, for example, increasing the on-target rate to 17 percent on average and including quality markers and additional content.
In collaboration with industry designers, they are also making improvements to the software interface to assist with the workflow, which they plan to fully implement later this year.
Franke and colleagues are not planning to commercialize their method, though he said they are talking to Agilent about providing it as a kit.
At the University of Kiel, they are planning to use the method both for research and clinical applications. In an upcoming research project, they will HLA-type a large number of Asian samples in order to build a better reference panel for HLA imputation.
On the clinical side, they plan to use the method for validation work and for bone marrow typing. The lab has already typed several hundred samples for a local bone marrow registry at a cost that is lower than traditional HLA-typing with Sanger or Luminex technology, Franke said. Diagnostic applications will require two additional features to be implemented in the software, the ability to log and track manual changes and a user authentication feature.