Researchers from Stanford University have developed an accurate, high-throughput, and cost-effective method for HLA genotyping using the Illumina HiSeq and MiSeq platforms.
The researchers showed that the method was highly concordant with reference cell lines and was able to identify several unique alleles with mismatches, insertions, and deletions. In addition, they believe that the technique may be able to avoid typing ambiguities that can plague other methods.
As such, the group hopes to eventually commercialize an HLA typing test based on the method, but is first working to further clinically validate the approach and compare it more closely with standard methods.
The method, described in a paper published last week in the Proceedings of the National Academy of Sciences, is based on targeting a contiguous segment of polymorphic HLA genes and amplifying each in a single long-range PCR procedure covering most known polymorphic sites. In the study, the group focused on HLA-A, -B, -C, and DRB1, which together define the minimal content for HLA matching for allogenic stem cell transplantation.
According to Chunlin Wang, the paper's first author and a researcher at the Stanford Genome Technology Center, the group has since moved on to work on covering additional HLA genes including DQA and B, DPA and B, as well as DRB3, 4, and 5.
"All major HLA genes we are covering right now," Wang told Clinical Sequencing News.
In the study, the group reported experiments showing the method could provide both a high turnaround time on a small number of samples using the MiSeq, and extremely high throughput — an estimated 2,700 samples — in a single HiSeq 2000 run.
Additionally, the new strategy of using continuous amplified segments and paired-end sequencing allowed read lengths of 400-500 bases, effectively matching the read length of the Roche 454 platform while exceeding its throughput.
"When throughput goes up, the price can drop dramatically," Wang said, "so that is a major advantage."
Another advantage, he said, is increased resolution.
According to the Stanford team, most other next-gen HLA sequencing approaches have used amplification of separate polymorphic exons, rather than a longer contiguous gene segment.
"With HLA," Wang said, "the majority of humans are heterozygous. So when the two copies mix together, the combination sequence might be identical in two different people [who actually have] different parents. This is called combination ambiguity."
By including a greater area of the gene, the Stanford group's method offers the opportunity to identify variants in other regions that can resolve this ambiguity.
The same is true for identifying alleles that vary outside of commonly targeted regions, Wang said. Marcelo Fernandez-Vina, another author of the study and director of Stanford's HLA typing lab, explained that other methods' more narrow focus leaves clinicians essentially working on assumption. Even newer methods that use other next-gen sequencing technology like Roche's 454 focus on a select group of exons of HLA genes, leaving out intronic and other regions.
Roche's 454 Life Sciences launched kits for genotyping HLA genes at high and medium resolution in 2011. Each of the company's research HLA assays comprises a primer set to PCR-amplify between one and three exons of up to 10 HLA genes (CSN 4/5/2011).
Overall, Fernandez-Vina told Clinical Sequencing News, greater allelic resolution can help provide more accurate HLA typing, which is why the team's new method was exciting to him.
"We didn't do this in the paper," he said, "but one could even extend to promoters or to some regulatory segments as well. The potential is there that we could be looking at [everything]."
In the PNAS study, the Stanford researchers validated their approach by retyping 40 reference cell lines, which showed that the new method had over 99 percent concordance with the reference data. Additionally, the approach demonstrated the ability to identify unique alleles, an important benchmark for new typing methods, according to the authors.
They also used 59 clinical samples to demonstrate the potential for higher throughput, showing that 99.3 percent of alleles met a minimum coverage of 100 using a single lane of the HiSeq 2000. Projecting from this, the researchers estimated that in typing four HLA genes, they could pool about 180 samples per lane, or almost 3,000 samples in one instrument run.
"In the setting of testing many subjects simultaneously, the cost for high-resolution typing by this successful methodology is significantly lower than classical Sanger sequencing, and in the same range or lower than the cost of probe-based assays," the group wrote, adding that this could enable "comprehensive disease-association studies with large cohorts."
According to Wang, the group is already involved in several of these large efforts. He said the researchers have already started using the approach in a 2,000-sample study of how HLA molecules predispose patients to narcolepsy. Another project that may be funded soon would include looking at the distribution of HLA alleles in schizophrenia from 3,000 samples, he said.
The team is also planning a study that will compare their method with other next-gen sequencing platforms being used for HLA typing, specifically Roche's 454.
Wang said he currently doesn't see advantages to 454 over his group's method . He said his team initially started working with 454, but was driven to develop the new approach to increase throughput.
Life Technologies has also said it plans to develop research assays for human leukocyte antigen typing on the Ion Torrent PGM (CSN 4/18/2012).
Wang said the PGM's higher error rate could be a problem for the HLA typing application.
"In HLA genotyping there is much higher demand for accuracy because you use this for transplantation," he said. "Accuracy is [the] number one priority, so Ion Torrent might not be useful unless they increase their accuracy."
According to Wang, the group's ultimate goal is to demonstrate that their method of using contiguous segments can be used clinically.
Toward this end, he said the group will be working with Fernandez-Vina's lab to conduct a side-by-side comparison of the approach with the standard practice of resolving HLA ambiguities — typically sequence-specific primers and probes coupled with Sanger sequencing.
The group already showed in the PNAS paperthat it could type a few samples using the Illumina MiSeq with a turnaround time of about five days.
According to Wang, the plan is to evaluate about five samples a day over a three-month period. "If we provide a better result, a cheaper price, a shorter time, eventually we could use our technology in the clinical setting," he said.
Fernandez-Vina added that the comparison study could also serve to validate the method for commercial use in a clinical HLA-typing service or an LDT.
"Once the opportunity is here, we do want to commercialize," Wang said. "Not just for money, but because if this is just a lab experiment and doesn’t become a robust test, it's not a real achievement."
Fernandez-Vina said that technological development will also be necessary.
"As improved sequencing technologies are developed, we can adapt the typing method to suit any sequencing platform," the authors wrote in their report.