By Monica Heger
Researchers at the University of California, Los Angeles have sequenced the genome of a brain cancer cell line on ABI's SOLiD sequencer and used the same cell line to also perform exome sequencing with Illumina's Genome Analyzer.
The UCLA scientists said the study, which was published last week in PLoS Genetics, illustrated that thorough sequence analyses of cancer genomes can be done quickly, efficiently, and at a low cost outside of the large sequencing centers.
The researchers analyzed the cancer cell line of a common and well-studied brain cancer known as U87MG. Glioblastoma is the most common and deadliest form of brain cancer, and U87MG is the most studied line, yet it has never been sequenced. So the researchers wanted to analyze that particular line in order to gain a better understanding of its mutations.
The team used the SOLiD sequencer from Life Technologies' Applied Biosystems group and a paired-end sequencing strategy. They generated an insert library of around 1.5 kilobases, obtained mate-pair read lengths of 50 base pairs and sequenced the genome to an average coverage of 30-fold.
"We got a very comprehensive analysis of the genome," said Stan Nelson, UCLA genetics professor and senior author of the study. Nelson's team identified over 2 million single nucleotide variations, 1,315 structural variations, nearly 191,743 small indels, and 35 interchromosomal translocations. "This genome is highly mutated," he added.
Nelson said that the long insert size helped in detecting interchromosomal translocations. "Because we had a 1.5-kilobase insert library, our chromosomal coverage was 15 times greater than our base coverage, and our base coverage was 30-fold," he said. "So, this gives us very good power to find every translocation."
While Nelson's group did detect significantly more translocations than some other cancer sequencing studies, he is not the first to detect translocations with next generation sequencing, and they can also be detected using cytogenetic methods. In a study published by a group at the Wellcome Trust Sanger Institute, researchers sequenced 24 breast cancer genomes, specifically looking for chromosomal rearrangements, detecting a total of 239 interchromosomal translocations across all 24 cancers (see In Sequence 1/5/2010). In two other cancer sequencing studies by the same group, the researchers detected three interchromosomal translocations in a melanoma sample, and seven from a small cell lung cancer sample (see In Sequence 12/22/2009).
After sequencing the cell line, the UCLA team also carried out exon capture with a custom-designed Agilent array, and sequenced the DNA on Illumina's Genome Analyzer. They targeted over 5,000 known cancer-causing genes. They did the exon capture in addition to the whole-genome sequence because they wanted to compare the two techniques to see if they found the same variants. Nelson said that the mutations detected by SOLiD were confirmed by the exome sequencing. "There were a few variants where the capture process didn't capture one allele," he said, but overall, the two were comparable.
He said that the sequence results showed both known mutations as well as novel ones. "There are some interesting mutations in cell adhesion genes, for instance, that are novel and not previously described," Nelson said. "These mutations may be the basis for the relative rapid cell migration that occurs as a phenotype in [glioblastoma] cells." He said his team is now studying those mutations to see if they do indeed affect cell migration.
Nelson said that one of his goals was to demonstrate that it is possible for a small team — his lab boasts five people — outside one of the large genome centers to efficiently and affordably sequence a cancer genome.
"The speed with which we can sequence and analyze the data [is] increasing very quickly, and the cost of doing this is plummeting," Nelson said, adding that his team performed the whole-genome sequence in about five weeks and at a reagent cost of around $30,000. In the future, he thinks it is likely that the genomes of cancer patients could be sequenced in order to personalize treatment. Although, before that can happen, the amount of DNA will have to be less than what was used in this study.
Michael Stratton, who heads the Wellcome Trust Sanger Institute's Cancer Genome Project, told In Sequence that the approach the UCLA team used was reasonable and they generated high-quality data and good coverage of the genome. However, he didn't agree that their method would necessarily be better at detecting translocations. He said while in theory, longer insert sizes would allow you to detect more translocations, he hasn't seen evidence of that in practice.
"For the genomes we published recently, we had a range of insert sizes, where we looked in detail for rearrangements. And we haven't found good evidence that says increasing the insert size finds more rearrangements," Stratton said, noting that his team has been able to detect rearrangements.
He also added that sequencing the cancer cell line as opposed to the primary tumor poses additional challenges in interpreting the data. The main problem, he said, is that most cancer cell lines do not have a normal cell line sample from the same individual, so it is not possible to characterize which of the mutations are somatically acquired.
While Nelson's group found several million mutations, Stratton said it is likely that only several thousand of those were somatically acquired. Knowing which mutations are somatically acquired is important because likely, only a subset of just five to ten of those are responsible for the development of cancer, he said.
However, he added, "these cancer cell lines are servicing most cancer researches around the world and most drug development, so it's reasonable to put in a considerable amount of effort, like sequencing the whole genome, to find all the variants," he said.