With a small grant from the US National Institutes of Health, DNAPrint Genetics and Senecio Software plan to develop ancestry-divining services and sell them to demographers and other social scientists.
Once it completes and perfects the software behind its technology, DNAPrint also hopes to use those assets and its “biogeographical origin” database to use admixture mapping to identify drug-response genes, along with potential drug responders.
In fact, it’s the post-drug-approval situation that may be most interesting from a pharmacogenomics standpoint. The company will “very shortly” acquire 51 percent of Biofrontera, a natural-compound drug developer, said Tony Frudakis, president and CEO of DNAPrint. “Not only will the drugs be targeted to [SNP-defined] segments of the population … but [also ethnic] subpopulations, however we can define them during our clinical trials.”
Frudakis envisions identifying ideal drug responders by screening for partial-penetrance SNPs, along with ancestry-related SNPs associated with response. The method would be used as a proxy with multigenic or incompletely penetrant traits lacking SNP information.
“There’s no reason to have an absence of that information,” said Paul Cusenza, a spokesperson for Perlegen, which genotyped several ethnicities to produce its SNP database in conjunction with the HapMap Consortium. “If someone believes that there is a genetic difference related to a response, then they should do a whole-genome study to find the genetic predictors of that response. And that’s the best way to make that drug useful in that perspective.”
Using ancestry-related SNPs as a proxy of response is simpler and cheaper than a whole-genome scan, and incomplete penetrance is “a very difficult problem to solve,” said Frudakis. He said the company had no plans to approach AstraZeneca about the early results of studies of its lung-cancer drug Iressa, which only showed a survival benefit for “Oriental” and non-smoking patients.
The kind of drug that would benefit from DNAPrint’s style of diagnostic would be similar to NitroMed’s heart-failure drug BiDil, said Frudakis. NitroMed resubmitted BiDil to the US Food and Drug Administration with an all-African-American patient population after it failed an initial drug trial that nevertheless showed significant survival benefits in African-Americans.
“If you do it by picking people who are black, that’s OK until you find something better,” said Robert Temple, associate director for medical policy at the FDA’s Center for Drug Evaluation and Research. For the moment, no one knows whether BiDils’ effects are genetically related at all, he said.
As far as CDER is concerned, there is no policy on ancestry-based theranostics. “We have not had discussions of this to my best knowledge, and it seems to me it sort of ‘de-precisions’ it,” said Temple. “If you can find a genetic characteristic, why wouldn’t you go after that genetic characteristic?”
Mapping Genes by Admixture
“In our mind, there’s a defect with the way research is done in pharmacogenomics today,” said Frudakis. “None of these companies or academic labs condition their statistical analyses on an appreciation of population structure,” he said.
Using knowledge about the commonalities of certain genetic regions within some ethnic groups, along with extensive analysis, admixture mapping effectively extends linkage-disequilibrium blocks. With ethnically admixed study participants, the decrease in the number of SNPs should simplify the process of finding genes and correct for spurious relationships related to ethnicity, Frudakis said.
“We could screen a genome with 2,000 markers, whereas a group using regular LD techniques would have to screen about one million markers per genome,” said Frudakis. In practice, the company would genotype a few thousand more, he said.
With admixture mapping, “nothing’s changed — there’s nothing new,” said David Altshuler, director of the Program in Medical and Population Genetics at the Broad Institute. The first research on admixture mapping appeared more than 15 years ago, he said. “Admixture mapping is a legitimate approach to mapping disease genes, but it is motivated in finding disease genes in certain communities, which counterbalances some of its concerning ethical ramifications,” he said.
The Tough Sell
When DNAPrint and Senecio begin shopping around their ancestral-origin service, they will face an inherently skeptical market. “Social science in general would not consider this a measure of race,” said Jerry Wicks, Senecio’s president.
The Census Bureau, for example, would never consider using genomic data, he said. “What this test does do is provide social scientists another facet of racial identities that could be used in studies” to augment the three accepted measures — self-reporting, observational, and parental reporting, he added.
Wouldn’t the test necessarily use genomics to define race? “You have to separate biogeographical ancestry from race, even though there is an obvious overlap,” said Wicks. Social scientists see “the dangers inherent” — that the possibility exists that “people will use this measure to reify a social construct” that is not real.
Wicks will informally introduce the service at the March 31 meeting of the Population Association of America, in Philadelphia.
"'What's the racial composition of this person?' - that might be more accurate than self-reporting," said Guang Guo, a University of North Carolina professor of sociology.
The Geneology Market
Extending ethnic-database information to tell people about their ancestry is misleading, said Altshuler. “It creates a false sense of certainty by putting a number on something, because, what does it tell you that you didn’t already know? ‘You have some fraction of ancestors from some broad continental origin, like Europe or Africa’ — so how exciting is that?” he added.
“I sent them my DNA,” said Esteban González Burchard, director of the University of California, San Francisco DNA Bank, and a researcher in admixture and disease. “They gave me my racial background.” Gonzalez was surprised when the company told him he had “some African” in his ancestry, he said. “The confidence intervals don’t cross zero, which is supposedly reliable,” he said.