GenePeeks, a startup planning to offer a preconception genetic risk testing service for sperm donations, raised $3 million from private investors and the equity firms Selway and Columbus Nova last month in a series A funding round. The company will use the money to support its commercial launch, initially with a single sperm bank.
Co-founded by Princeton University molecular biologist Lee Silver and his partner Anne Morriss, GenePeeks will use genomic data from donated sperm and potential mothers to create "virtual gametes." It will then combine these computationally into thousands of virtual genotypes, which it will analyze for likelihood of disease. Based on this analysis, the service will provide a sperm bank client with a filtered pool of potential donors with whom she is least likely to have a diseased child.
According to Silver, who announced the company at last month’s Consumer Genetics Conference, GenePeeks is hoping to open shop within the next six months starting in partnership with a single sperm bank.
He told Clinical Sequencing News that the company is currently working with an Affymetrix chip focused on about 100,000 genetic variants linked to rare recessive and other single-locus diseases, but the plan is to transition to exome sequencing as soon as it becomes affordable.
“We are going to keep a very, very close eye on exome sequencing,” he said. “Price point is more important for us because we are offering this to people going to sperm banks, so they are going to be paying out of pocket … So it just depends on when the tech comes down to a good price for a good set of sequences.”
According to Silver, GenePeeks’ process is based on the creation and analysis of virtual genomes using Monte Carlo simulation and a proprietary genotype-phenotype database developed to help calculate a score representing the likelihood that a certain variant combination will disable a gene’s products.
“So whether it's microarray genotypes right now or then whole-exome sequencing and eventually whole-genome sequencing, it doesn’t matter,” Silver said. "We can be agnostic on what technology we use to get the data because ultimately we put all the data into a VCF file" for analysis in the same pipeline.
Silver said that GenePeeks’ computational strategy, at least for single-locus diseases, assesses genes independently of each other, meaning that regardless of whether the company is inputting genotype data or whole genome sequences, the process still works the same way.
“We create a virtual gamete based on knowledge of meiosis and recombination,” he explained at the Consumer Genetics Conference. “Each sperm and egg is one potential outcome that we bring together into a virtual genotype. This can be used to interrogate databases to predict health. You do this a million times, make a million virtual progeny, each of which has a discrete, non-probabilistic genome.”
While the company waits for exome sequencing costs to drop sufficiently it is starting with a chip that has “every known nonsense and missense mutation in every gene connected to a recessive disease.”
“These [diseases] are often called simple, but they are not really simple diseases because of allelic heterogeneity,” he said.
“The basic premise we are working with is that disease is a function of genotype, not a function of alleles,” Silver said. “We see variant combinations in individuals that don’t cause disease in those individuals because most of the genes we are looking at are genes that code for enzymes where maybe you can have only 50 percent of and it doesn’t affect your health.
“So, for example, for a gene that creates a metabolic enzyme we’ll look at an estimate of what would be produced. We can then compare to the data that exists that can tell us where disease is likely or not likely and how severe it would be.”
The company’s mission hits close to home for Silver’s partner Morriss, who conceived a child through sperm donation. Morriss’s son was born with a rare recessive disorder called medium chain acyl-CoA dehydrogenase deficiency, or MCADD.
In MCADD various mutations cause different levels of deficiency in production of an enzyme that breaks down fats.
Lee said that Morriss carries one of the most common disease alleles for MCAD deficiency on one chromosome, while on the other she has a fully functional allele. The donor sperm Morriss used to conceive came from a man who was also a heterozygous carrier of a mutation causing lack of MCAD enzyme function.
Current protocols for genetic screening of sperm donations vary, Silver said. But the first companies and academic centers offering sequencing-base preconception screening to couples have already brought tests on the market, most notably Good Start Genetics, which earlier this year launched a targeted sequencing panel for couples working with in vitro fertilization clinics (CSN 10/24/2012).
According to Silver, simple carrier screening isn’t sufficient to fully assess conception risk. For example, he said, many of the deleterious mutations that cause MCADD and other similar diseases result in mild disease caused by reduced, but not completely absent, enzymes. “But there are other nonsense mutations which are more severe, that can eliminate functionality of the protein completely,” he said.
“That’s what we are basing this testing on — the fact that carrier testing doesn’t really help in these cases, because its not just a question of you carrying a mutation, it’s what the mutation is and what it will do in combination with another mutation.”
GenePeeks’ process could technically be applied to non-anonymous parental couples in addition to single clients assessing donated sperm, and the company may move in that direction in the future, according to Silver. But starting with sperm banks is “easier logistically,” he said, “because it’s a negative selection of anonymous partners.”
By interpreting only virtual genomes, and not the genome of donor sperm itself, Silver said GenePeeks also gets around a thorny issue for sperm banks — that measuring genetic risk directly in donors would require them to disqualify potentially larger numbers of donors from their current and future banks.
“If they had to test all these donors, they would have to eliminate a lot of donors because they [would have to] say that any applicants who test positive are eliminated from the pool. This is the way all sperm banks operate,” Silver said at the October conference.
“We are going to collect genome sequences for each donor, but they [just] sit in our database. That is not a genetic test, that’s just information,” he said.
Silver also said at the conference that by limiting analysis to virtual genomes, GenePeeks’ service remains outside the purview of medical device regulation by the US Food and Drug Administration. “We create genomes that don’t exist but might exist, so we are not actually genetic testing anyone,” he said. However, these virtual genomes are going to be created using information from real genomes. Silver has said the company would make customers’ raw genetic data available to them if they want it.
As GenePeeks moves to exome and eventually whole-genome sequencing, Silver said the company will also expand from single-locus diseases to more complex diseases “in a stepwise fashion.”
“We will be developing our computational software based on information flowing out of all the research being done by other people. So the data has to be there for us to analyze,” he said. “But most likely we will work out from single-locus diseases to diseases affected by just a couple of genes, so looking at modifying genes.”
Ultimately, "we are assuming that the knowledge base broadens so that we can go beyond that to even more complex diseases,” Silver said. “I see no reason why we can’t get there.”