By Monica Heger
One year into a three-year, 1,000-tumor sequencing project, researchers at the University of Chicago's Cancer Genome Project have sequenced the transcriptomes of over 100 tumors and cancer cell lines on the Illumina Genome Analyzer, and plan to move into exon capture and possibly collaborate with Complete Genomics for whole-genome sequencing, the Chicago researchers told In Sequence last week.
The team's plan is to sequence and compare clinically different tumors — those that respond to chemotherapy versus those that do not, for example — in order to improve treatment and diagnosis, said Thomas Stricker, clinical fellow in the lab. The $20 million project is a University of Chicago initiative that will be funded over five years from resources provided to the Institute for Genomics and Systems Biology in Chicago.
Stricker said the group began the project with libraries of 36 base pairs, but that number has been increasing — first to 75 base pairs, and now they are beginning to get read lengths approaching 100 base pairs. They are using a standard RNA-seq protocol and a paired-end sequencing strategy on the Illumina GA, and have sequenced RNA from over 100 tumors and cancer cell lines, primarily from breast and head and neck cancers. The team obtains its samples, which are all frozen, from collaborating surgeons and oncologists, as well as the University of Chicago's biospecimen bank.
They have sequenced several subtypes of breast cancer tumors in order to determine the differences between the types that respond well to treatment and those that do not, with the goal of finding new drug targets in the cancer subtypes that do not respond to current treatment options. The RNA-seq strategy allows them to "take the low-hanging fruit approach," said Stricker. "If you're looking for drug targets, those will be mutations that are expressed in the coding region."
However, he said, the RNA-seq strategy has some drawbacks, which is why they are now moving forward with an exon-capture strategy. He said they will begin running two lanes of exon sequencing and one lane of RNA-seq for each tumor or cancer cell line they sequence. "You get a more even distribution of reads in exon capture; you get more even coverage across the genome, which will make it easier to call SNPs," said Stricker.
The integrity of the RNA is also key for obtaining unbiased and sufficient coverage, and he said that the team had had some problems with RNA degradation, which limited the quality of the RNA-seq data.
Furthermore, biases tend to get introduced with the RNA-sequencing protocol, so some areas are sequenced many times, while others become lost, Stricker said. Nevertheless, he added, they've uncovered valuable data from the RNA-sequencing data to date. "We've found thousands of genetic variants, many of which appear to be deleterious and responsible for the cancer phenotype. Now, we're trying to narrow those down," he said.
They will also expand into other cancers such as leukemias and the major carcinomas. And, aside from exon capture, Stricker said the team is considering moving into whole- genome sequencing and is discussing the possibility of using Complete Genomics' service. The researchers are also considering purchasing an Illumina HiSeq 2000 when the machines become available.
Whole-genome sequencing would allow the researchers to more closely study family histories of cancer, said Stricker, because in those cases, the variants of interest are not always in the coding region. He said they are trying to decide whether it makes more sense to do the whole-genome sequencing in house or to contract it out to Complete Genomics. A number of factors, including costs, available resources, and throughput will drive their final decision.
Chicago's cancer sequencing project differs from the Cancer Genome Atlas project, which is composed of many different sequencing centers studying over 20 types of cancer from thousands of patients.
Kevin White, director of the Institute for Genomics and Systems Biology, who heads the cancer sequencing project at the University of Chicago, said his group is taking more of a treatment-focused approach, trying to identify clinically relevant differences between cancer subtypes.
He said the Chicago team will compare its data with that produced from the Cancer Genome Atlas, but will be focused more on comparing metastatic versus non-metastatic tumors or chemoresistant versus chemo-sensitive tumors. "Our overall goal is to accelerate the process of getting cancer genome discoveries translated into clinically useful tests," White said in an e-mail.