by Greg Yap

The genome sequencing revolution of the 1990’s propelled research into the modern genetic era. Inexpensive, reliable, and automated DNA sequencing methods allowed scientists to sequence the complete genomes of organisms ranging from lowly bacteria and viruses to higher plants, animals and humans. In the wake of this flood of information, we are now faced with the far more daunting task of determining how knowledge of billions of nucleotide bases can be put to practical use to improve human health and treat disease.

High-density microarray technology, invented by Stephen P.A. Fodor and colleagues,^(1-3) allows scientists to sift through enormous genome sequences and identify important biological information. Microarrays have now afforded scientists the ability to analyze gene expression for the complete coding content of the human genome in a single experiment. This objective analysis method has helped scientists discover the genetic pathways disrupted in diseases ranging from cancer to multiple sclerosis, and has enabled them to more accurately stratify disease, predict patient outcome, and make better therapeutic choices.

The most recent generation of microarrays allows scientists to quickly genotype 10,000, 100,000 or even 500,000 SNPs distributed across the human genome, enabling genetic association studies of complex disease and drug response. Researchers are also using high-density microarrays for targeted genotyping studies of over 10,000 SNPs. These tools for disease mapping studies,^(4-8) deliver high numbers of markers and high resolution, and have helped pinpoint genes linked to diseases such as sudden infant death syndrome,⁽⁶⁾ neonatal diabetes,⁽⁷⁾ and macular degeneration.⁽⁹⁾

Figure 1. Millions of Perfect Match/Mismatch probe pairs enable scientists to quickly genotype hundreds of thousands of SNPs and determine if a genome contains two copies of the A allele (AA), two copies of the B allele (BB) or a copy of each (AB).

Genotyping 500,000 SNPs across the genome

The data capacity afforded by microarrays from Affymetrix Inc. (Santa Clara, CA) facilitates the study of up to 500,000 SNP genotypes per sample analyzed. For any given SNP of two possible genotypes — A or B — probes are synthesized on the microarray corresponding to the sequence of both alleles. Following hybridization of the target to the array, scientists can then determine whether a SNP is an AA, AB, or BB genotype by simply analyzing if the A allele probes have detected their complementary sequence, if the B allele probes have detected their complementary sequence, or if both have detected complementary sequences.

Every SNP genotype represented on an Affymetrix microarray is measured through a perfect match probe, as well as a mismatch probe. The mismatch probe serves as an internal control, accounting for spurious signals and cross-hybridization. Each probe-pair is the basic unit used to call an SNP genotype. However, to ensure highly accurate genotype calls, GeneChip arrays routinely use multiple probe pairs to call the genotype for each SNP represented on the array.

The first probe pair contains the SNP precisely in the center of the 25mer sequence. The remaining probe pairs are positioned, or tiled, to the right and the left of this central position. This strategy is used to genotype the A and B allele of every SNP from both sense and antisense strands of DNA. The need for high-density manufacturing technique quickly makes itself obvious when genotyping large number of SNPs — over 10 million probes are used to genotype the 500,000 SNPs represented on the Human Mapping 500K Array Set.

500K: assay

The key to array-based SNP genotyping demands an assay that does not require allele specific amplifications. Affymetrix developed a whole genome sampling assay (WGSA) that uses only one primer to genotype hundreds of thousands of SNPs distributed throughout the genome.⁽¹⁰⁾ Previous SNP mapping efforts have been hampered by the need for locus specific amplification and the need for many tens of thousands of PCR amplifications — an expensive and cumbersome undertaking.

     The WGSA method uses a restriction enzyme to digest genomic DNA, creating various sizes of DNA fragments, each containing their respective SNPs. However, only certain sized fragments are applied to the array, so it’s critical to design microarray probes against those SNPs that are present on the DNA fragments. For example, the Mapping 100K set uses two separate restriction enzyme reactions, each of which creates a pool of DNA fragments containing over 50,000 SNPs to be genotyped. The same strategy is now being used on the 500K to genotype up to 500,000 SNPs.

Genotyping up to 10,000 custom SNPs

     Researchers are using newly developed MegAllele assays and GeneChip microarrays to genotype up to 10,000 targeted SNPs in a single experiment, enabling them to perform focused mapping studies and candidate-gene association studies quickly and in great detail.

     Array-based custom SNP genotyping required developing a method to simultaneously amplify thousands of selected SNPs and an equally scalable way to determine each genotype. Multiplexed assays and high-density microarrays now provide the individual scientist with a practical solution for large-scale genotyping efforts that were previously limited due to cost and ease-of-use hurdles.

Assay

     Molecular Inversion Probe (MIP) technology enables scientists to amplify up to 10,000 targeted SNPs in one highly multiplexed experiment that combines 10,000 different PCR reactions in a single tube. When the SNP sequence is amplified during PCR, a fluorescent base is incorporated at the variable SNP position, and depending on the genotype — either an A, T, C or G — the DNA will contain either a green, blue, purple or red fluorescent molecule.

Probe strategy

     Researchers then use microarrays to detect each SNP sequence, image the associated fluorescent color, and ultimately determine the final genotype. Every probe on the microarray surface is designed to detect a different SNP by hybridizing to a DNA sequence that acts as a SNP identifier; each of the 10,000 PCR primers in the initial MIP assay contain one of 10,000 unique DNA sequences that serve as a unique tag for each of the 10,000 SNPs.

     The scalable targeted genotyping method uses a four color high-resolution scanner to image the fluorescence associated with each probe; red corresponds to a thymine genotype, green to an adenine genotype, blue to a guanine genotype, and purple to a cytosine genotype. If a SNP sequence is imaged as a pure red square, scientists will know that both alleles of the SNP contain a “T” nucleotide — a “T/T” genotype — because only the thymine nucleotides that were incorporated during the MIP assay had red fluorescence. Likewise, if the probe lights up as pure green, scientists know both alleles have an “A” nucleotide — or an “A/A” genotype — because green fluorescence corresponds to adenine. If a probe fluoresces yellow however, scientists know that the SNP contains one red “T” allele and one green “A” allele that combine to make a yellow “T/A” heterozygote genotype. By examining these combinations of colors for as many as 10,000 SNPs on a single array, researchers can determine the exact genotype for each targeted SNP detected and imaged on the microarray.

Figure 2. Molecular Inversion Probe (MIP) technology and high-density microarrays enable scientists to select and genotype 10,000 targeted SNPs in a single experiment.

Discovering the genetics of complex disease

High-density genotyping microarrays enable scientists to conduct whole-genome association studies today to understand the genetics of complex disease or drug response. Josephine Hoh of Yale University, for example, used 100K SNP microarrays to identify a key mutation associated with age related macular degeneration. Hoh’s study, published in the April 2005 issue of Science, scanned 100,000 SNPs from just 146 people. Previous technologies that looked at far fewer SNPs would have required the study of thousands of people to generate results with the same scientific significance.

     Larger association studies can also be designed to help identify the comprehensive catalog of genes, pathways, and predictive biomarkers associated with disease. For instance, the Serono Institute used the Mapping 100K Set to identify 80 genes related to multiple sclerosis (MS). Serono scientists plan to use the Mapping 500K Set for a higher-resolution study, and they expect they will find many more MS-associated genes that will likely fall into five to ten different disease pathways.

     Often the ideal experiment tests a pre-existing hypothesis about particular genes, pathways, or regions of the genome. Dr. John Todd of Cambridge University's Juvenile Diabetes Research Foundation/Wellcome Trust Diabetes and Inflammation Laboratory (DIL) is among the first to use MegAllele panels of 10,000 non-synonymous SNPs — SNPs that change the sequences of proteins — for a whole-genome association study to find genes contributing to type 1 diabetes. Todd’s research group is comparing the SNP profiles between 1,000 control samples and 1,000 diabetic samples. He plans to analyze over 20,000 more DNA samples already collected from diabetes patients and their relatives.

Studies to identify genes associated with drug response, efficacy and toxicity may become one of the most promising applications for whole genome DNA analysis. Microarrays able to genotype more than 500,000 SNPs distributed across the genome now allow researchers to readily genotype large populations of responders vs. non-responders to a given drug for phenotypes including efficacy and toxicity. With these kinds of genetic studies, scientists hope to elucidate the genes contributing to variable drug response.

In late-stage clinical trials for example, microarray genotype analysis could be used to stratify patient populations to eliminate poor or toxic responders from key Phase III trials. Such stratification would help ensure maximum effectiveness through clearer statistical differentiation between drug and placebo, while also reducing size and cost of trials and improving the odds of drug approval. In addition, once a drug is on the market, patient stratification could be used to accelerate drug expansion into new indications through faster, smaller, more definitive Phase IV trials, or to establish medical superiority of a late-to-market drug relative to entrenched competitors in an important class of patients. Genome-wide genotype information will also fuel future research. By better understanding genetic mechanisms of drug response in patients, researchers will have made significant progress on finding the next generation drug.

Already, leading pharmaceutical, biotech, and university researchers have embraced microarray genotyping technology. Researchers in a pharmacogenomic study at the Mayo Clinic are using the 100K array set to investigate the genetic basis for differential responses to anti-hypertensive drugs in different patients and populations. These scientists hope to identify genes influencing drug response and ultimately tailor anti-hypertensive therapy for individual patients.

Whole-genome genotyping microarrays provide a way of examining the underlying genetics of responders and non-responders without any of the assumptions or limitations used in a candidate-gene approach. For most drugs with variable responses, little is known about why they work in some patients and not in others. Genotyping microarrays enable scientists to explore the whole-genome and identify predictive markers of disease and drug-response, that may ultimately provide more tailored, effective and safer courses of treatment and help avoid the over 100,000 annual fatalities from adverse drug reactions in the U.S. alone.⁽¹¹⁾

About the author

Greg Yap is Vice President, DNA Products with Affymetrix Inc. More information about the technology discussed in this article is available from Affymetrix. Affymetrix, Inc. 888-362-2447 www.affymetrix.com

1. Fodor, S. P. et al. Light-directed, spatially addressable parallel chemical synthesis. Science 251:767-73 (1991).

2. Fodor, S. P. et al. Multiplexed biochemical assays with biological chips. Nature 364:555-6 (1993).

3. Pease, A. C. et al. Light-generated oligonucleotide arrays for rapid DNA sequence analysis. Proc Natl Acad Sci USA 91:5022-6 (1994).

4. Gissen, P. et al. Mutations in VPS33B, encoding a regulator of SNARE-dependent membrane fusion, cause arthrogryposis-renal dysfunction-cholestasis (ARC) syndrome. Nat Genet 36:400-4 (2004).

5. Middleton, F. A. et al. Genomewide linkage analysis of bipolar disorder by use of a high-density single-nucleotide-polymorphism (SNP) genotyping assay: a comparison with microsatellite marker assays and finding of significant linkage to chromosome 6q22. Am J Hum Genet 74:886-97 (2004).

6. Puffenberger, E. G. et al. Mapping of sudden infant death with dysgenesis of the testes syndrome (SIDDT) by a SNP genome scan and identification of TSPYL loss of function. Proc Natl Acad Sci USA (2004).

7. Sellick, G. S., Garrett, C. & Houlston, R. S. A novel gene for neonatal diabetes maps to chromosome 10p12.1-p13. Diabetes 52:2636-8 (2003).

8. Uhlenberg, B. et al. Mutations in the Gene Encoding Gap Junction Protein alpha 12 (Connexin 46.6) Cause Pelizaeus-Merzbacher-Like Disease. Am J Hum Genet 75 (2004).

9. Klein, R. J. et al. Complement factor H polymorphism in age-related macular degeneration. Science 308:385-9 (2005).

10. Kennedy, G. C. et al. Large-scale genotyping of complex DNA. Nat Biotechnol 21, 1233-7 (2003).

11. Lazarou, J., Pomeranz, B. H. & Corey, P. N. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 279:1200-5 (1998).