At A Glance
Name: Arthur Holden
Position: chairman, SNP Consortium; chief founder, chairman, and CEO, First Genetic Trust; chairman, Flex Gene Consortium.
Background: After graduating magna cum laude from Union College with a bachelor’s in science, Holden received an MBA, with honors, from the JL Kellogg Graduate School of Management at Northwestern University. He has been a senior executive at Baxter International, and was CEO and a director of the biotech firm Celsis International.
The first phase of the SNP Consortium — to build a full genome-wide map of around 300,000 SNPs — is wrapping up, according to chairman Arthur Holden.
Right now, the consortium has identified around 1.9 million valid-ated SNPs. The group has mapped 90 percent of these, and has performed allele-frequency studies on close to 100,000.
This SNP dataset has been folded into dbSNP and integrated into all other existing public datasets, and the SNP Consortium continues to BLAST the dataset against the human-genome draft sequence.
The second portion of the $54 million endeavor, which will soon kick off, is to facilitate the HapMap Project. The consortium will underwrite and help build the HapMap’s data-coordinating center at Cold Spring Harbor Labs, which will be run by bioinformaticist Lincoln Stein. The consortium also may underwrite and help develop the full set of SNPs required to assemble the map, according to Holden.
Though the major goal of the consortium is nearly complete, Holden said he thinks the group has just begun. “We’ll be discovering SNPs for a long time,” he said in the interview below.
Holden also is optimistic about the future of SNP-genotyping tools within drug development and reckons they’re very close to rounding a critical technological corner. In fact, the traditional technology continuum — tools are born in academia, are tested in nonprofit labs, and sold to pharma — is not at play here.
“Right now the genome centers are very immature in their genotyping experience,” Holden explained. “Actually, most of the ... early-adoptive pharmaceutical companies are much more experienced.”
So where does that leave SNP-tool shops and their traditional first-line clientele? And what pharmas will recognize this technology and land on their feet when the market returns?
Holden spoke with SNPtech Reporter about these issues and more …
What’s the status of the SNP count in the public records, and how will the SNP Consortium eventually add to that?
If you look at it, there are more than four million SNPs in the public data set now. And I would very surprised if [the SNP Consortium] doesn’t expand that by one to two million additional SNPs.
But it’s not only the number of SNPs [in total], but the number of SNPs that are seen in multiple databases and that have been validated as true SNPs and mapped to lend themselves to laying down primers for genotyping. So that’s starting to be a very large number of SNPs.
We’ll be discovering SNPs for a long time. The issue is what profile of SNPs in a given population … will be required to draw meaningful associations. ...
We actually have precious little data on using genome-wide association studies. And that’s what the whole motivation of the SNP map was really about; moving to a new paradigm where we could put on microarrays enough SNPs across two populations — one that exhibits a certain response and another that doesn’t-and compare those to see what areas appear different.
The reality is there’s been precious few experiments that have been published on this. Some of the early data, which is unpublished and on which I cannot comment, indicate that we need quite a bit lower number of SNPs than you may have thought from some of the theoretical calculations that were done on linkage disequilibrium and [on] what sort of density of markers you’d need to draw an association.
So the fact is, until you do the experiments, the data aren’t there. But I would say right now that for broad association studies, looking to find initial linkage with a lot of phenomena, there are plenty of SNPs available. And well-characterized SNPs aren’t the real limiting step any more. It’s actually just [an issue of] beginning to develop the technique and process for [using] genotyping to characterize, in a reasonably cost-effective way, the appropriate number of individuals.
But I think the good news is that genotyping technology is moving forward, no questions about it, albeit slowly since some of the companies are hurting in this economy.
What’s your take on how pharma companies view technology available for identifying SNPs and applying them to drug discovery and development?
If you say, ‘Do pharma companies understand the importance of genomics and impact genomics can have?’ unequivocally, “yes.” How else would we have been able to get as broad a support in the consortium, to spend that much money, time, and effort developing this map? It wouldn’t have happened.
Are they all using genomics to its fullest? No. There are some aggressive leaders and some very passive followers. My guess is that those that get out front first and understand it will have a commanding competitive advantage over the long term. You can look out and ferret the early adopters.
What technologies are the academic labs using for the work in the SNP Consortium and the HapMap Project? We know, for example, that the Whitehead Institute for Biomedical Research is using Sequenom’s technology for its contribution to the SNP Consortium, and that the Baylor College of Medicine is using ParAllele’s chemistry in the HapMap.
Actually, I think a snapshot of any given period is pretty insignificant. I would say right now that genome centers are very immature in their genotyping experience. Actually, most of the … early-adoptive pharma companies are much more experienced.
It’s always the issue when you have an early technology and a lot of relatively small companies are looking to stabilize themselves. Some come in, some come out. You really see an exodus.
Just take the example of Orchid. We use Orchid’s technology extensively within the SNP Consortium to do validation activities, and their service provided excellent results. Problem was, they couldn’t coalesce as a viable business model in the genotyping space, so they had to get out of that.
It’s a very tough market to enter because everybody wants to do genotyping, and they feel like they have to do more genotyping in order to make the experiments feasible from a scientific point of view, but then the costs become prohibitive. So the technology platforms that can survive are the ones that allow a reasonable number of SNPs to be done at an acceptable level of cost, but it would also [require] a profit to keep a concern going. And that’s a tough thing to do.
I personally believe over time that the aggregation of capabilities in larger centers is inherently where we want to go with this, meaning that this basically becomes a service-business structure where you keep in a high-quality manufacturing-process environment someone that’s got both the technology and scale and operations to do this at the lowest cost and still make a profit. And that’s where I believe most people see this going. It will become an out-sourced activity.
What about the pharma and biotech companies that have their own in-house setup?
Actually, I think they’re the ones who will be the most rational about it. The fact of it is, even for the most aggressive pharma, that’s still not a massive amount of volume to justify all of this time and investment [of building in-house genotyping labs]. I think the problem is, it’s very expensive and the technologies are so early and there’s so much change in them that you invest in one platform and before you know it, it’s obsolete.
When do you think SNP-genotyping technologies will begin coming out of the background and begin playing a premiere role in drug discovery and development?
I think for drug discovery and development, clearly you’ve got to continue to improve the public access of SNPs that are characterized and validated. Also, you’ve got to get people to the point that they [have] budgets and resources to be able to do whole genome-association studies. And the early results there seem to show that is very promising.