Deputy Director for Advanced Technologies and Strategic Partnerships
National Cancer Institute
Name: Anna Barker
Position: Deputy Director for Advanced Technologies and Strategic Partnerships, National Cancer Institute
Background: Senior executive at Battelle Memorial Institute; Cofounder of Oxis International and Bio-Nova
Education: : PhD, immunology and microbiology, Ohio State University
The proposed Cancer Genome Project, a joint effort of the US National Cancer Institute and the US National Human Genome Research Institute, faces an uphill task. There is a lot of sequencing to do, and no clear idea which tumor cells to concentrate on. In addition to the large variety of cancer types, many individual tumors consist of mixtures of cell types. And while sequencing technology continues to improve, no existing method will turn the project into an easy job.
So, in order to focus the Cancer Genome Project, the NHGRI and the NCI said last week that they would each contribute $50 million to a pilot project with the goal of guiding the larger effort, while examining the technology it might employ.
The pilot effort is expected to make use of the newest generation of sequencing and genomics technologies, some of which were awarded portions of the $32 million in research grants the NHGRI is distributing as part of its unrelated effort develop cheaper and faster sequencing (see Snippets, this issue).
We spoke to Anna Barker, the deputy director for advanced Technologies and Strategic Partnerships at the NCI, to learn more about the pilot project and about what sorts of technologies might interest the researchers.
What are the major technological obstacles that the project has to overcome?
I think there about three or four that we are focused on. One is basically the complexity of the tumors themselves. As you probably know, a surgically excised piece of tumor is not at all homogeneous, and knowing how to actually separate out, and ensuring that you have enough DNA content from cancer cells is a real challenge for us.
Are you going to study homogeneous tumor DNA content?
Some of the tumors that are more homogeneous, like hemotologic malignancies, we believe will be easier, but most of our major killers in cancer are actually the solid tumors, where we have a lot of collections, we have a lot of interest. I think the heterogeneity of the tumor itself, and even within the tumor, the heterogeneity that actually resides because of contamination from normal cells, stromal cells, all the things that actually contaminate it.
I think the second issue is, given that, how do we amplify the DNA from cancer? So, whole-genome amplification is an important technological challenge for us. It has been successfully done, but not routinely. If we are going to do a much more high-throughput project, we're going to have to amplify DNA from tumor cells reproducibly, and that's a real challenge.
From tumor cells or clonal lines?
Likely, it will be clonal, but nothing that we're going to be doing in the early days is going to be single-cell. What we've seen so far is that those folks who are doing whole-genome amplification with tumor are having some success, but it's an issue.
If I look at the two issues that probably will cause us the biggest challenges, at least in the short term I think in the longer term, it's going to be interrogating the data, and basically sifting out the signal-to-noise questions. And we're going to get a lot of data, there's no question we're going to find lots and lots of changes. But I think the issue is coming up with meaningful approaches to the data, and the algorithms that will be required to actually determine those that are most meaningful relative to cancer and the progression of cancer.
There are lots more technological challenges, but if I had to pick those off the top of my head, I think those are some of the major ones.
How many cancer types are you going to narrow this down to?
The short answer is, we just had a meeting the week before last of the cancer community and the sequencing community of a little over 140 people, I think, and one of the discussions was, 'How many cancers? How many types of cancer?'
And that breaks down into subtypes of cancer, because if you pick breast cancer, we know that there are [more than] 10 subtypes, and maybe as many as 20 subtypes of breast cancer.
The short answer is that we just haven't made a decision on that yet. We've got a lot of discussion going on with our advisory board on that, and the range is anywhere from one tumor, where you go very deep into the tumor, in terms of really understanding all the changes that you possibly can, up through several tumors, where you don't go as deep into the interrogation of the genome. And you just do more to look for patterns.
My guess is that we're going to settle on a few. And I'm guessing that it's going to be certainly less than five. And I'm thinking that that's the range that we'll be looking at for the pilot.
What sort of technologies is the project going to employ, and what sort of companies stand a chance of joining in on the project? Genomics? Sequencing?
Well, this meeting actually focused that question quite interestingly. It's certainly more than resequencing, and I think that we're not discounting any technologies. And we're certainly not going to be looking exclusively at any technologies.
I think it ranges from shotgun sequencing, to resequencing, to all the other arrays of technology that you can look at to interrogate the genome. Epigenomics is very much in play, and all of the technologies that could be used there. Certainly gene expression array analysis, back sequencing, there are numerous technologies, and I think our plans would be to really ask the community to tell us how to do this. Which technologies are going to be most meaningful? And I think our role is going to be to integrate the results from all those technologies, including sequencing, to get the most meaning out of these kinds of data.
And we really want to push the technology like everybody else, we really want the $1,000 genome.
This seems like a good project for the cheapo-genome approach. Have you been in discussion with any of those companies or researchers?
Yes. Dr. [Francis] Collins and the National Human Genome Research Institute has a fairly good-sized program in this area, and there are several small companies I know working directly with them on some very innovative approaches to gene sequencing.
It's a very interesting approach. There are several of those now, and I think they're all going to compete, and we're going to be the winners because of that competition. So, I'm hopeful that a lot of small companies will get engaged in really pushing the technology here.
That seems like something they would all be interested in competing over, too.
Yeah, it's certainly a bigger challenge than the [Human] Genome Project, in terms of just sequencing. But it's one that we'll have to have move forward fairly quickly in cancer, because it's a huge undertaking, obviously.
Who have you spoken to?
Everything's informal and everything's competitive, so we don't speak to anybody once we start our procurement processes, so that we don't actually preclude people from applying. There are lots of government rules about that.
Once we finish our meeting and get lots of feedback from the community, we're proceeding to establish the mechanisms for which people will apply to be part of the project.
When do you expect that to happen issuing Requests for Applications or Requests for Proposals?
There's a process for that, and we're hoping that that process will proceed so that the major portion of those would come out early in 2006. That's a very aggressive timeline for the government.
I've read criticisms of the project that had a lot to do with signal-to-noise. Is that a major concern?
Absolutely. I think that's a valid question, and that's why we're going with a pilot project to try to define the parameters that will be needed to make this a meaningful project, and also to establish its feasibility. Feasibility defined from our perspective as being technically feasible, also producing significant insights into cancer, as well as clinically meaningful data.
I think those criticisms are expected, and I think that's why we're doing it the way we're doing it. We have the same questions as the questions that are being asked by the folks that are actually making those comments, and I think we just have to go answer those questions.
Is there a technology that is key to overcoming that obstacle? Is it a bioinformatics question?
Yes. I think it's a bioinformatics question, and it's very complex, because there are lots and lots of different kinds of data that will be coming in here. I mean, everything from raw data to clinical data. There will be annotation on the specimens that are used, et cetera. There's an integrative issue, in terms of just integrating all the data, and there's also the interrogation of the data, or mining of the data. I think it's going to really challenge the bioinformatics community to really come up with some creative approaches here. So, I think this could be a real shot in the arm, as well, for medical bioinformatics, because we're pushing the envelope in terms of the state of the art here. But I think just like sequencing and other genomic-analysis technologies, we're going to be pushing for change and pushing for real new approaches. So again, I think it will be very, very important for small companies and new companies to get involved in this.