SNPs, single-base mutations scattered throughout the genome, can allow geneticists to map traits with astonishing resolution and speed. An overview of some of the new technologies that companies are offering for SNP profiling, and the diverse applications for which researchers are using them, reveals a rapidly evolving field with a promising future.

By Alan Dove

Inclusion of companies in this article does not indicate endorsement by either AAAS or Science, nor is it meant to imply that their products or services are superior to those of other companies.

In 1910, Thomas Morgan observed something unusual in one of the many jars in his Columbia University laboratory: a white-eyed fruit fly. It was one of the best things to happen to biology since Charles Darwin landed a berth aboard HMS Beagle. Soon determining that the eye color trait was linked to the sex of the fly, Morgan embarked on a series of experiments that jump-started modern genetics. Biologists have been mapping genes with increasingly sophisticated tools ever since.

Later 20th-century researchers eventually discovered microsatellites, short DNA sequences that had been duplicated throughout the genomes of eukaryotes during evolution, providing more precise genetic signposts than Morgan’s phenotype-based maps. While many researchers still use this approach (see "The Decaying Orbit of Microsatellites"), the era of genome sequencing has brought another tool to the fore: single nucleotide polymorphisms, or SNPs.

The Numbers Game

SNP mapping has quickly become the standard tactic for two types of genetic studies in humans: linkage analysis and genomewide association. Linkage analysis is the direct descendant of Morgan’s original experiments. By tracking the inheritance of a trait and a set of genetic signposts, such as microsatellites or SNPs, through several generations of a family, researchers can map the trait to specific chromosomal regions.

Genomewide association (GWA), also called "genetics without families," treats the entire population as a single giant family, then tries to find associations between specific traits and sets of SNP markers. GWA eliminates the need to find large families with accurate, multigenerational medical records, and also allows geneticists to probe common, complex diseases. Unsurprisingly, it has become an extremely popular technique.

"Within the last seven months or so, we have had 14 studies come out doing [GWA] showing significant findings in various diseases, including prostate cancer, type 2 diabetes, [inflammatory bowel disease], and Crohn’s disease," says Todd Dickinson, director of product marketing at Illumina, based in San Diego, California.

Because they sample completely outbred populations, GWA studies require a different approach from traditional linkage analyses. In GWA, researchers generally screen a relatively small population using hundreds of thousands of SNPs, to highlight the regions of the genome most likely involved in a disease. In a second phase, the focus shifts to just those regions, using smaller numbers of SNPs across a much larger population, to provide statistical power.

In the first phase of a GWA study, screening more SNPs increases the chance of finding a meaningful association, so Illumina and their principal rival in the SNP market, Affymetrix, have been racing to boost the densities of its array-based assay systems. "We have evolved from 10,000 to 100,000 to 500,000, and our latest array actually has over 900,000 SNPs on it," says Jessica Tonani, a genotyping specialist at the Santa Clara, California, headquarters of Affymetrix.

Besides boosting the sheer number of SNPs on arrays, manufacturers are also trying to make them more informative. "We have intentionally selected to target what we call tag SNPs. These are SNPs that represent a group of SNPs that tend to be inherited together," says Dickinson. Illumina asserts that this effectively multiplies the screening power of its million-SNP chip.

The drive for more density may be reaching a practical plateau, though. "We’re hitting a point of kind of diminishing returns when you look at actually adding increased SNP content," says Tonani, who adds that most GWA studies would now benefit more from studying larger patient populations than from screening more SNPs. To enable that, she says Affymetrix is now focusing on lowering the prices of its high-density SNP arrays rather than packing more SNPs into them.

When Less Is More

Researchers working on the second phase of a GWA study, or those doing traditional family-based linkage analyses, have different SNP profiling needs. Rather than screening thousands of SNPs in dozens of individuals, they are screening dozens of SNPs in thousands of individuals.

"Typically for the whole genome analysis, people will identify a candidate region that they’re interested in, or several genes, and need to move to a lower number of SNPs but a higher number of samples," says Phoebe White, senior director of genotyping at Applied Biosystems (ABI) in Foster City, California.

To cater to that market, ABI now offers the SNPlex platform, which combines real-time PCR with the company’s ubiquitous capillary electrophoresis platform to track 12 to 48 SNPs per experiment. Because each SNPlex reaction is much cheaper than a high-density array, even small labs can screen SNPs across thousands of samples. "A large majority of labs these days do have a [capillary electrophoresis] platform, and so what we wanted to do was enable them to do these kinds of genotyping studies without bringing in a new type of platform," says White.

While using equipment on hand is obviously cheaper, labs that frequently plan to check specific genotypes may want to investigate the new LightCycler 480 from Roche Diagnostics in Basel, Switzerland. Besides functioning as a high throughput quantitative PCR platform, which can carry out a 40-cycle reaction on a 384-well plate in just 40 minutes, the LightCycler can also process SNP genotypes with a unique DNA melting point assay. Rather than designing separate primers for each allele of each SNP and carrying out complex enzymatic reactions, researchers simply create one set of primers per SNP and perform a single physical assay, in which the precise primer-template melting point reveals the sequences of the template’s SNP alleles. Any mismatches between a SNP allele and the primer sequence will change the melting point, and the LightCycler’s unbiased algorithm can even identify previously unknown SNP alleles.

For specific gene mapping applications, investigators may narrow their focus even further, looking at specific candidate genes associated with just a handful of SNPs. For example, patient responses to the anticoagulant warfarin associate tightly with a single gene. "It’s critical now to understand what your genotype is to know what kind of dosage you need, but the number of SNPs you need to screen for is less than 20," says White. At that scale, even the coverage provided by the SNPlex kit may be too broad, so researchers resort to profiling individual SNPs by standard real-time PCR.

While screening fewer SNPs is undeniably cheaper than screening many, groups working on GWA must be careful not to simplify the assay prematurely. "If you’re to move forward with a subset of SNPs, oftentimes you would miss certain things," says Affymetrix’s Tonani. Pointing to the recent GWA mapping of potential diabetes-related genes, Tonani adds that "in three separate studies, and of the nine genes that they actually identified, eight of the nine genes wouldn’t have been identified for this subset panel in at least one of the studies."

Into the Core

Like most new technologies, SNP profiling systems have become progressively cheaper in recent years. Indeed, according to some estimates, the cost per SNP genotype has dropped more than a thousandfold since 2000. Nonetheless, with GWA studies now profiling close to a million SNPs per patient, large-scale genotyping remains an expensive undertaking.

That’s why most labs now send their big SNP projects to a dedicated core facility. "I think nobody is buying the system and setting it up for a specific project. It’s always some kind of facility, either a company or within a place like ours," says Ann-Christine Syvanen, research group leader for the SNP Technology Platform at Uppsala University in Sweden.

Syvanen’s facility, like several others scattered around the world, functions as a national-level service. Scientists in Sweden and other Scandinavian countries can send samples to the Uppsala center, which charges only its marginal operating costs to genotype up to 1 million SNPs per sample.

For the high-density analyses, the center uses the Illumina Beadstation with Illumina’s GoldenGate and Infinium assays, while a GenomeLab SNPstream from Beckman Coulter setup handles projects with 12–384 SNPs per sample. "We have a facility that can analyze anything from a few SNPs or one SNP up to a million SNPs on chips, so we have the possibility to do any kind of SNP genotyping project really," says Tomas Axelsson, the Uppsala center’s manager. Axelsson adds that the center is now doing everything from candidate gene screening and family-based linkage studies to large-scale GWA work for researchers all over the country.

Besides the expensive equipment, a SNP screening core also provides trained technicians and, critically, the bioinformatics infrastructure to handle floods of genotyping data. While the makers of SNP-screening chips generally build database software into their chip readers, most big facilities also run their own databases for more complex analyses.

Storing and studying genome-scale datasets also creates a serious hardware problem. A typical GWA study may fill 2 terabytes of hard disk space, and more recent studies that also incorporate whole-genome sequence information may take up to 15 terabytes per study. Even the declining cost of disk storage hasn’t solved the problem, as the data must also be backed up and maintained in a permanent archive, a serious challenge for any type of computer data. "The technology for bioinformatics is evolving, but not the technology for storing the data," says Syvanen.

Chip makers are also aware of the informatics problems, and are starting to offer more sophisticated solutions. For example, in June, Illumina announced the launch of iControlDB, a database of controls for GWA studies. Regardless of the SNP profiling platform they’re using, researchers can tap into this trove to see SNP profiles for thousands of controls that don’t have a particular disease. The company hopes to have 20,000 control patients’ SNP profiles in the database by the end of 2007.

Making a Federal Case of It

Basic researchers have been avid consumers of SNP screening technology, but most experts have also been expecting pharmaceutical companies to adopt the technology, as part of the push for pharmacogenomic therapies. Pharmacogenomics, which seeks to customize therapies based on patients’ genotypes, could improve drug efficacy while simultaneously reducing the risk of side effects. Pharmaceutical companies are understandably circumspect about the details of their internal research programs, so until recently it was unclear how much SNP screening was really going on in industry.

"Over the past three years, we have gotten a number of voluntary genomic data submissions which were dealing with differential gene expression using microarrays. Over the past year or so, we have also started to get whole-genome association studies," says Federico Goodsaid, senior staff scientist in genomics at the US Food and Drug Administration’s Center for Drug Evaluation and Research.

Under current FDA regulations, companies can submit genomic data voluntarily as part of a drug approval application, if they think the results will improve the agency’s understanding of the product. Goodsaid reports seeing about "three to five" such submissions involving GWA so far. "But remember, we get to see something several months or maybe a year past the point at which people have started to do it," he adds, suggesting that a wave of GWA-based pharmacogenomic therapies may soon be headed for the clinic.

Indeed, equipment manufacturers are already anticipating a boom in clinical genotyping, with new systems that simplify the process and make it more robust. Luminex, in Austin, Texas, for example, now markets its adaptable xMAP bead-based flow cytometry system for a wide range of diagnostic genotyping applications, ranging from tissue typing to cardiac marker profiling. In the Luminex system, allele-specific oligonucleotides attached to microspheres serve as sensitive probes for targeted genotypes, which can be read quickly and reliably by flow cytometry.

Besides new gear, the next generation of clinical genotyping tests will also require clearer benchmarks. Because the standards for validating GWA-based biomarkers are still in flux, the FDA has now incorporated these studies into the second part of its Microarray Quality Control (MAQC) initiative. The MAQC’s first phase focused on the reproducibility of gene expression experiments, but the second phase aims to standardize microarray data analysis. "The questions we had in MAQC I went back to how you ran the hybridizations and stuff like that, but in MAQC II, we are worried primarily with what happens after the data [are] there," says Goodsaid.

Currently, each GWA study develops and uses its own data analysis algorithms, so the MAQC II team plans to construct a matrix of tests, cross-applying different algorithms to different data sets and comparing the results. The project, which should be completed in 12 to 18 months, should help set the tone for a burgeoning field. "My guess is that we’re going to be seeing a lot more [GWA] in the near future, because this is obviously a powerful technique for identifying different types of biomarkers," says Goodsaid.

Whether for developing pharmacogenomic therapies, searching for new disease-related genes in genomewide association studies, or performing traditional linkage analyses in families, SNPs are clearly revolutionizing genetics. For geneticists, the revolution has been a long time coming: after all, they’ve seen the whites of its eyes for nearly a century.

Alan Dove is a science writer and editor based in Massachusetts.

DOI: 10.1126/science.opms.p0700018

Personal Genomics Website

With more than 2,000 human DNA variations and their medical consequences now cataloged and available, the SNPedia website is expanding to include analyses of personal genomic sequences, with the first of many being the sequences from J. Craig Venter and James Watson. Built on a wiki-platform that encourages community annotation and updating, SNPedia provides a valuable online resource to scientists, physicians, and students studying genotyping and genome variation, and even to lay individuals interested in knowing more about their own genomes.
SNPedia
For information 240-465-0788
www.snpedia.com

Deepwell Plates

Eppendorf Plate Deepwell 96 and 384 set new standards for purity, stability, and centrifuge and automation capability. With innovative, optimized well geometry, an exclusive error-safe labeling system, and “g-safe” stability up to 6,000 x g, the plates deliver first-rate results. The plates incorporate the new RecoverMax design with conical well bottoms and smooth, rounded corners and transitions for square well plates. These features provide faster pipetting, resuspension, and mixing because the liquid collects in the center of the well bottom, while wicking in well corners is prevented to eliminate cross-contamination between wells. Another innovation introduced in these plates is the OptiTrack matrix, which combines a choice of five vibrant border colors and a high-contrast, ultraviolet-resistant and chemical-resistant alphanumeric labeling system. This ensures simple plate orientation, even in stacks, and quick, error-safe identification of plates and wells.
Eppendorf
For information +44(0)1223-200440
www.eppendorf.com

Genotyping BeadChip

The Infinium Human-Linkage-12 Genotyping BeadChip takes advantage of the 12-sample format for linkage analysis. It offers low cost per sample for linkage analysis plus industry-leading call rates, uniform marker distribution, and superior single nucleotide polymorphism (SNP) content. This new addition to the Linkage portfolio complements the GoldenGate Human Linkage V Panel, which is suitable for degraded samples such as formalin-fixed, paraffin-embedded specimens. It is also available for researchers interested in conducting linkage analysis studies using Illumina’s FastTrack Genotyping Service. Linkage analysis maps the location of disease-causing loci by identifying genetic markers that are co-inherited with the phenotype of interest. The 12-sample BeadChip and the flexibility of the Infinium Assay provide researchers with a cost-effective method for studying linkage with high information content.
Illumina
For information 858-332-4055
www.illumina.com

Cloning Ligation Kit

The Rapid DNA Dephos and Ligation Kit provides fast, convenient, and economical dephosphorylation and ligation of sticky-end or blunt-end DNA fragments, enabling ligation of DNA in 5 minutes and dephosphorylation in 10 minutes. The kit contains the new recombinant rAPid Alkaline Phosphatase, which is heat-inactivated in two minutes at 75°C, eliminating phenol-chloroform extraction steps and allowing dephosphorylation to be performed directly in restriction enzyme digests. Procedures are thus simplified for a broad array of applications, including cloning of fragments into plasmid or phage vectors, linker ligation, recircularization of linear vector DNA, and generation of libraries.
Roche Applied Science
For information +41-61-688-6983
www.roche-applied-science.com

Room Temperature Bacteria Storage

A new technology for storing and shipping E. coli at room temperature, CrudE is based on Biomatrica’s proprietary synthetic chemistry technology, which preserves unpurified plasmid and genomic DNA harbored in E. coli. Bacterial cultures can be directly dried into CrudE, allowing for quick and efficient sample collection. Users simply add water to recover intact DNA ready for subsequent downstream applications. A sample management database tracks and organizes samples using barcode and radio-frequency identification tags that are integrated into CrudE individual tube, 96-well plate, and 384-well plate formats.
Biomatrica
For information 866-379-6879
www.biomatrica.com

Note: Readers can find out more about the companies and organizations listed by accessing their sites on the World Wide Web (WWW). If the listed organization does not have a site on the WWW or if it is under construction, we have substituted its main telephone number. Every effort has been made to ensure the accuracy of this information. Inclusion of companies in this article does not indicate endorsement by either AAAS or Science, nor is it meant to imply that their products or services are superior to those of other companies.