Prospecting for Gold in Genome Gulch

The human genome is much like the American West of the 1850s: Everyone wants a piece of the pie. Similar to gold prospectors of 150 years ago, biotech and pharmaceutical companies, and even universities, are frantically searching for the nuggets of gold that will help them find the mother lode—a gene whose function is sufficiently marketable to make all of the preliminary research worthwhile. Companies that do strike gold get to introduce new classes of drugs to the market. Others hope to make genetic tests that can identify people at risk for disease, or to help companies target their drugs to the genetic subpopulation that responds best. Because only time will tell which genes are useful and which are not, universities and companies are taking no chances, and are staking claims to as many of the recently revealed genetic properties as they can identify. This approach, however, is producing a bottleneck of sorts: Massive amounts of time and money must be spent sifting through those genes to find the ones that are implicated in disease. Scientists take three basic approaches to gene discovery: association, gene expression, and functional studies. Association studies—the most common gene identification approach— connect genetic markers with a disease population, whereas gene expression studies examine genes that are differentially expressed between diseased and normal tissue. Functional studies focus on genes whose proteins have a functional role in the diseased. Most of the effort in gene discovery—and therefore much of its sophisticated technology—revolves around association studies.1 Association Studies In association studies, researchers attempt to correlate genetic markers with disease in a population. The general idea is that if a genetic marker tends to be found more frequently in people with a given disease, then that marker must be near the disease gene. Once a marker is identified, a researcher can find the proximal genes using either public or private databases. Many research groups use SNPs (single nucleotide polymorphisms) as their genetic marker of choice. These single-base substitutions are the most common type of genetic difference between individuals, making them an effective way to survey the genome.1 What's more, databases exist that catalog thousands or even millions of SNPs. The SNP Consortium (TSC) has assembled the most notable collection (snp.cshl.org).2 Funded by 14 pharmaceutical and technology companies, TSC is a collaboration between research groups at Stanford University, Washington University School of Medicine in St. Louis, the Whitehead Institute for Biomedical Research in Cambridge, Mass., and the Wellcome Trust's Sanger Center, which identify and validate SNPs throughout the genome. TSC is collaborating with Princeton, NJ-based Orchid BioSciences to validate 60,000 of TSC's SNP candidates and to determine the frequency of those SNPs in healthy population and in three broad ethnic groups: African-Americans, Asians, and Caucasians. Yet most SNPs in the collection are neither validated nor mapped, a fact that some companies are trying to exploit. San Diego-based Sequenom, in collaboration with UK-based GlaxoSmithKline and Incyte Genomics of Palo Alto, Calif., has developed a comprehensive SNP assay portfolio that includes roughly 400,000 validated, gene-associated SNP candidates. Sequenom is currently studying the normal distribution of these SNPs in a healthy 15,000-member population and in samples from four ethnic groups. Foster City, Calif.-based Applied Biosystems (ABI), is also in the process of developing a comprehensive SNP collection. The company has launched an initiative to validate SNPs and create a set of SNP assays across every human gene to enable disease-association studies, says Dawn Madden, ABI's senior manager for genomics applications marketing. The company currently offers a custom service for providing SNP assays and reagents on request, but by this summer, ABI expects to offer the first of its 200,000 validated SNP sets as off-the-shelf assays for its one-step TaqMan PCR amplification technology. "You just add one reagent and your sample. It's very, very simple," explains Madden. Likewise, Orchid BioSciences intends to develop assays for some portion of the SNPs that the company has validated for TSC. Many companies will also develop assays for SNPs in public databases on request. SNP Genotyping With the intense competition to find disease-associated genes, companies are turning to faster, more efficient means of interrogating the genome. Traditionally, PCR amplification of SNP-containing fragments followed by sequencing was the preferred genotyping method. However, sequencing individual SNPs can be tedious work when surveying many thousands of samples and an equivalently large number of SNPs. Several companies, including Sequenom, have developed mass spectrometry-based methods for rapid and accurate genotyping, in which the spectrometer distinguishes SNP variants by their different masses. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry allows researchers to genotype a given SNP in individual samples or in an entire population at once.3,4 The system can also be multiplexed to genotype multiple SNPs, by designing the SNP-containing fragments to be in a different mass range. MALDI-TOF-based genotyping significantly cuts time and effort, says Toni Schuh, Sequenom CEO. An experiment that involves genotyping 100,000 SNPs in a 1,000-person normal population and a 1,000-person disease population would involve 200,000 individual reactions—one for each SNP in each population. "That is enormous high throughput," Schuh says. Only if a SNP shows an allelic frequency shift does Sequenom go back and sequence that one SNP in the entire sample population. "Our philosophy is one of ongoing compression," Schuh says. Hayward, Calif.-based Lynx Therapeutics offers another high-throughput genotyping technology, called MegaType™.5,6 "We have come up with a way to do these SNP allelic-frequency experiments essentially in one test tube, in one experiment," says Norman Russell, Lynx president and CEO. Lynx loads SNP-containing fragments onto beads using its Megaclone™ technology. The company labels the fragments from the control and disease samples with different fluorescent markers and combines those probes to hybridize with the beads. The beads will then contain a mixture of fragments from the two sample pools. If a particular SNP—and therefore fragment—is more common in one population than the other, the corresponding bead will contain more labeled probe from that sample. By running the beads through a fluorescence-activated cell sorter (FACS), Lynx can collect those beads, enabling the company to identify the SNP's sequence. Because MegaType doesn't rely on available SNP databases, the technology effectively eliminates concerns about the accuracy or relevance of known SNPs. Instead, it comprehensively surveys the entire genome. Only when the company finds a SNP of interest do Lynx researchers sequence the DNA fragment and locate nearby genes. The technology is also easily transferable to animal or plant models. "Because we don't need to know anything about the SNPs beforehand, Lynx has the unique opportunity to work in organisms beyond humans," Russell says. Lynx uses its own collection of DNA samples in its gene discovery efforts and will perform SNP analyses on customers' samples. Another genotyping methodology involves microarrays. Santa Clara-based Affymetrix, best known for its GeneChip® arrays for gene expression studies, offers a 1,500-SNP array that covers all 22 autosomes and the X chromosome. The company also offers a GenFlex™ Tag Array that allows researchers to genotype up to 2,000 SNPs from pooled DNA samples.7 Likewise, PolyGenyx in Worcester, Mass., offers its OmniScan™ and SNPScan™ microarray- based technologies for high-throughput and high-density SNP genotyping. Courtesy of Luminex Corp.Red Light, Green Light: Researchers can use Luminex Corp.'s multiple analyte profiling (MAP) system for both SNP genotyping and gene expression analyses. Because of the many SNP-detection options available, Orchid BioSciences has distanced itself from any one platform. "We have technology that will survive as platforms evolve—our biochemistry is applicable to several platforms and is able to evolve with them," says Michael Boyce-Jacino, Orchid's vice president of R&D and chief technology officer. The company's SNP-IT™ technology is compatible with ABI's SnaPshot™; Piscataway, NJ-based Amersham Biosciences' SnuPe™; Austin, Texas-based Luminex Corp.'s LabMAP® system; and Affymetrix's GenFlex Tag Arrays. Orchid BioSciences also offers a 96-well genotyping kit through Carlsbad, Calif.-based Invitrogen, and offers mid- to high-volume genotyping and validation services. Orchid will begin marketing the SNPstream® UHT this year, an ultra-high throughput system capable of scoring more than 140,000 SNPs in an eight-hour shift. Population Control With so many companies specializing in genotyping SNPs, it can be difficult to tell one from the other. One differentiating factor is the populations each company uses to identify genes. Companies such as Genomics Collaborative Inc. in Cambridge, Mass., provide access to DNA samples for a wide range of disease types. Other companies provide access to families with a high rate of a particular disease, or to twin samples. Many companies have turned to so-called founder populations for their DNA samples. These are populations that derive from a small number of founder individuals. Because they share their genetics, such populations are essentially an extremely large family and likely share the same set of genetic risks for disease. Reykjavik, Iceland-based DeCODE Genetics has taken this approach in Iceland, mapping genes in over 20 common diseases with a genealogy database linking the entire living population and going back 1,000 years, while Myriad Genetics in Salt Lake City, has had success identifying genes in Utah's Mormon population. DNA Sciences Inc. of Fremont, Calif., also uses the Utah population in its search for low frequency variants. "We use these very large families to see the footprints of genes that have variants that have low to moderate penetrance," says Ray White, DNA Sciences' chief scientific officer. "Not all the answers will be there, but we think that the majority of the interesting variants will occur among the lower frequency variants." Sequenom divides its collection of 15,000 samples from healthy individuals into pools based on age. The company then genotypes SNPs from its own SNP portfolio within each age group. Theoretically, SNPs that are in linkage disequilibrium with a fatal disease will become less and less common in the aging population. So far, Sequenom has found about 120 genes with alleles that disappear in older population groups and expects to find roughly 400. Not only does this approach quickly identify genes with a clear role in a fatal disease, but it also gives the company an idea of its eventual customer base for a possible therapy. "If you know that you are working on a gene that carries a polymorphism that removes 7% of all North Americans from the healthy pool then you know that you have about 30 million customers," Schuh says. Gene Expression and Functional Approaches Gene expression studies have made gene chips a hot commodity, and are a relatively simple way to identify disease-related genes. These studies compare the relative expression level of genes in normal and diseased tissue or in tissues at different developmental stages. If a particular gene is expressed at low levels in the normal sample and at extremely high levels in the disease sample, or vise versa, chances are it has something to do with the disease. Courtesy of Sequenom Inc.Mass Spec-Tacular: Scientists use Sequenom's MassARRAY system for high-throughput SNP genotyping. This MALDI-TOF mass spectrometry-based system can analyze up to 3,840 samples in a single, unattended run. DNA microarrays are convenient tools used to quickly screen the expression levels of several thousand genes; one microarray experiment can highlight virtually all of the gene expression differences between two tissues. ABI has also seen the value of expression profiling,8 and is developing a series of TaqMan-based assays to quantify gene expression. "We've been a leader in gene expression and we again are raising things up a notch by creating gene expression assays for every gene in the genome." Madden says. These assays should be available this year. Lynx's Massively Parallel Signature Sequencing (MPSS) technology is yet another method for precise, digital quantitation of gene expression across the entire genome.9 (see Tools & Technology). Association and gene expression studies highlight genes without regard for the protein's functional role in the cell—first, uncover the gene, and worry about what it does later. Yet both approaches ignore the volumes of data already at researchers' disposal. Years of research into disease biology have uncovered proteins with known roles in disease pathways. Not only might these genes be potential therapeutic candidates, but they may also open up new areas of research. Functional candidates could "be an entry into pathways that open up new genetic opportunities," says DNA Sciences' White. Companies with functional candidates can use high throughput sequencers to quickly search for poly-morphisms in that gene between diseased and healthy populations. Most companies examine functional candidates as part of an overall gene discovery strategy that includes association studies or gene expression studies. "There are 30- or 40,000 genes, but fewer than half of those do we know anything functional about, so we are automatically excluding half of the genes—maybe more—when we look at functional candidates," White says. "The functional candidates are an interesting set and deserve to be explored but they aren't going to be the whole answer," he adds. Rather, the complete answer is likely to involve a combination of all three approaches, access to the right DNA samples, and the ability to interrogate those samples quickly. And like the gold miners of long ago, companies will need one more commodity to strike it rich: plain, old-fashioned good luck. Amy Adams (amya@nasw.org) is a freelance writer in Mountain View, Calif. References 1. J.F. Taylor et al., "Using genetic variation to study disease," Trends in Molecular Medicine, 7: 507-12, November 2001. 2. The International SNP Map Working Group, "A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms," Nature, 409:928-33, Feb. 15, 2001. 3. K.H. Buetow et al., "High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight spectrometry," Proceedings of the National Academy of Sciences (PNAS), 98:581-84, Jan. 16, 2001. 4. M.S. Bray et al., "High-throughput multiplex SNP genotyping with MALDI-TOF mass spectrometry: Practice, problems, and promise," Human Mutation, 17:296-304, January 2001. 5. S. Brenner et al., "Gene expression analysis by massively parallel signature sequencing (MPSS) on microbead arrays," Nature Biotechnology, 18:630-4, 2000. 6. S. Brenner et al. "In vitro cloning of complex mixtures of DNA on microbeads: physical separation of differentially expressed cDNAs," PNAS, 97:1665-70, 2000. 7. Fan et al., "Parallel genotyping of human SNPs using generic oligonucleotide tag arrays," Genome Research, 10:853-60, 2000. 8. J.V. Snider et al., "Human disease characterization: Real-time quantitative PCR analysis of gene expression," Drug Discovery Today, 6:1062-7, Oct. 15, 2001. 9. J. Reinartz, et al. "Massively parallel signature sequencing (MPSS) as a tool for in-depth quantitative gene expression profiling in all organisms," Briefings in Functional Genomics, in press. Supplemental Materials Suppliers of Gene Discovery Tools and Services

Prospecting for Gold in Genome Gulch

Interested in reading more?

Become a Member of