As it is gearing up to become a “center of excellence” in cancer research, the Ontario Institute for Cancer Research is building a next-generation sequencing facility designed to analyze tumor DNA that may put it on par with other large-scale sequencing centers, In Sequence has learned.
Founded at the end of 2005, the Toronto-based non-profit institute is in the process of recruiting 20 internal faculty members and building new lab space. John McPherson, who joined OICR last July as director of cancer genomics, is currently setting up the institute’s cancer genomics and high-throughput screening platform.
The center has so far accumulated five Illumina Genome Analyzers and five Applied Biosystems SOLiD systems and plans to increase its next-generation sequencing capacity another three- to fourfold over the next two to three years, said McPherson.
Cancer genomics and high-throughput screening is one of four technology platforms housed at OICR. Besides next-gen sequencing, it will include genotyping on Sequenom’s MassArray system and Illumina’s BeadStation, laser capture microscopy, and qPCR. The other three platforms at the institute are imaging, biorepositories, and informatics.
The aim of the genomics platform is to analyze hundreds or even thousands of tumors, McPherson told In Sequence last week. “We feel that the old school of ’10 tumors, 10 normals’ doesn’t cut it anymore,” he said.
As whole-genome sequencing is still too expensive, even with the new technologies, he and his team plan to study the transcriptome, the methylome, the exome, structural variations, and copy number variations of tumor samples. “And the core of all that, right now, is next-generation sequencing,” he said.
Since he arrived at OICR in July from Baylor College of Medicine, where he was an associate professor in the department of molecular and human genetics and at the Human Genome Sequencing Center, McPherson has been setting up the sequencing equipment and building the bioinformatics back end. “We are really starting from scratch here,” he said. “We had empty labs and empty server rooms, building our clusters, our storage, et cetera.”
Right now, his team is still evaluating the sequencing systems as it collects fresh pancreatic cancer samples in order to generate xenografts for its first tumor-analysis project.
The reason he chose five Illumina and five ABI next-gen sequencing platforms is that they may serve different purposes. Though the two platforms are perceived by many as competing head-to-head, McPherson believes that is not necessarily the case. “I think that there is a lot of room for both in the arena,” he said. His group plans to identify the most appropriate applications for each system.
For example, one platform might be better suited for SNP detection, while the other might be stronger in mapping structural variations. Alternatively, it might be best to combine data from both systems for a given application.
“It remains to be seen which one is going to work better for which analysis,” McPherson said.
“We feel that the old school of ’10 tumors, 10 normals’ doesn’t cut it anymore.”
The two platforms might also have different biases, resulting, for example, from the nature of their sequencing chemistry or the way they amplify sample DNA. “At this point, we don’t know, we are still investigating that,” he said.
Over the next two to three years, the institute plans to increase its current sequencing capacity by three- to fourfold, though McPherson has not decided how many instruments of which type it will likely add as new systems from Helicos BioSciences, Pacific Biosciences, or others come to market.
Owing to the nature of the planned projects, the facility decided to forgo both 454’s sequencing system and Sanger sequencers, at least for now.
The reason he passed on 454’s Genome Sequencer is that “for what I want to do, at least initially, the longer reads are not as important to me as having more reads,” McPherson said.
He added that the center will outsource its Sanger sequencing needs, which will mostly involve sequence validation.
According to McPherson, the biggest challenge for the fledgling facility will be data analysis, especially since the Illumina and ABI platforms generate different data types.
To prepare for the onslaught of data, which he called “the biggest headache,” of the endeavor, the institute is currently expanding its computing equipment, which is “growing leaps and bounds at the moment,” McPherson said.
Storage capacity adds up to several hundred terabytes right now, he estimated, and will grow to “probably well over a petabyte.” The institute’s bioinformatics cluster has reached about 750 cores and will grow to around 4,000 cores eventually.
Also, about half of the institute’s 20 internal faculty will do research in the area of bioinformatics. Recently, OICR recruited Lincoln Stein from Cold Spring Harbor Laboratory to be its director of bioinformatics. He is expected to start in the fall.
Another challenge will be to keep up with the rapidly changing technology, both in terms of instruments and protocols. “It seems like every week things are changing,” McPherson said, adding that “it will settle down a little more by the summer, perhaps.”
In addition, because few the existing samples in OICR’s tumor collections have the appropriate consent for sequencing, the center will “have to start collecting anew or go back and re-consent,” he said.
Besides the internal research groups, housed at the so-called MaRS Center downtown, near the University of Toronto, OICR will also fund 30 faculty at other research institutions in Ontario.
OICR is not the first institute to embark on large-scale genomic analysis of tumors. For example, both the pilot phase of the National Institutes of Health’s Cancer Genome Atlas and the Wellcome Trust Sanger Institute’s Cancer Genome Project are currently ongoing.
McPherson said he and his US and UK colleagues who are working on these projects are working together closely to avoid duplicating efforts.