International SNP Meetings

Some 100 invited international researchers from industry and academia attended the Second International Meeting on Single Nucleotide Polymorphism (SNP) and Complex Genome Analysis held in Munich, Germany, in September 1999. Below is a summary of meeting highlights with updates from the September 2000 meeting in Taos, New Mexico.

The overall tone of the 1999 meeting showed that SNP research still lacks a solid consensus about how best to proceedand this is at a time when vast sums of money are being spent on public and private SNP programs. This state of affairs is only partially improved today. SNP discovery alone will not determine how and whether SNPs should be used. Instead, this knowledge has to come from empirically determined guidelines and studies based upon best guesses and theory using the latest technologies.

SNP Discovery
The SNP Consortium (TSC) has established a well-structured program toward discovery and public cataloging of many hundreds of thousands of random genomic SNPs (at a 95% accuracy target) in 2 years, plus detailed mapping of over 170,000 of these. Original goals are being greatly surpassed. Other projects involve automated alignment and comparison of expressed sequence tag (EST) sequences for SNP discovery. Although limited by the depth and number of distinct EST contigs to perhaps a few tens of thousands of SNPs, the exercise is highly cost-effective. Furthermore, the intragenic location of these variants could make them individually more useful than TSC-discovered markers.

SNP Scoring and Detection
A plethora of competing SNP genotyping methods is being developed, but useful approaches must meet stringent requirements of both high throughput and accuracy. A fully ideal method still does not exist today. Extremely large sample materials may be required to achieve sufficient statistical power in association studies, and a genotyping error rate of as little as 1% could have a disastrous effect on statistical power. Claimed genotyping costs still range from a fraction of one dollar to many dollars per sample.

Two principal approaches to SNP scoring are in individual reactions and in a multiplexed fashion, usually achieved by using immobilized oligonucleotides on microarrays. The requirement for PCR before the detection reaction remains a major bottleneck. Simultaneous analysis of pooled DNA samples may be useful for increasing throughput and decreasing the cost of SNP scoring in association studies.

Diseases and Phenotypes
SNPs and association analysis are being used to (1) home in on disease-related mutations within large regions previously identified by linkage scans, (2) screen several variations surrounding a few or many prechosen candidate genes, or (3)follow extensive sequence studies of single candidate genes to determine associations between specific haplotypes and disease. These strategies can sometimes workalthough how best to maximize the rate of success is not known.

If there is no initial linkage to guide the search, however, the current answer seems to be to make educated guesses about which genes are likely to be important. Presenters summarized attempts to do this for Alzheimers disease, schizophrenia, dyslexia, and substance dependence. These studies began with a strong prior belief in the relevance of the tested candidate gene. Other presented data showed how intricate estimations of haplotype configurations, regression, and cladistic analyses can lead toward the precise intragenic location of the pathogenic allele and genotype combinations. This evolutionary perspective was a key take-home message.

The emphasis of most SNP research on tools and genotyping (finding the link between marker and pathogenic allele) was contrasted with the relative lack of attention to careful study design and sample ascertainment (finding the link between pathogenic allele and disease). Many case-control association studies may be futile because the lack of even rare families segregating the disease could indicate too-high genetic complexity.

Population Genetics
Speakers presented work on characterizing extant haplotypes and developing maps describing typical distances up to which allelic variants can be expected to be in linkage disequilibrium. These efforts can benefit disease-gene and population-genetics studies. Identifying potential targets of selective fixation via SNP analyses may be useful in revealing footprints of adaptive evolution. Revealing such dominant beneficial alleles could provide important targets for study.

Databases and Bioinformatics
Future high-throughput detection will require efficient systems for collecting and integrating voluminous amounts of data in high-quality databases. Representatives from the European Molecular Biology LaboratoryEuropean Biotechnology Institute presented software and database solutions to linking various locus-specific mutation databases with SNP databases allowing complex queries.

Many groups endeavor to mine SNPs from EST data, but measuring allele frequencies in silico is difficult, and many rare alleles might be among them. Prediction accuracy is estimated at only 60% to 80% in most cases, although this identifies gene candidates for further investigation. Others are attempting to map large numbers of SNPs onto human chromosomes and three-dimensional protein structures to understand phenotypic differences and human evolution using SNP data.

Intellectual Property, Commerce
A clear trend is toward granting patents on partial nucleic acid sequences or SNPs only when functions and commercial applications can be defined. Commercially, profits are expected to be generated in three areas: applying SNPs to pharmacogenomics by discovering functional implications, genotyping individuals for particular SNPs, and creating technology platforms for SNP discovery and use.

In the first situation, proprietary rights to the important SNPs are expected to generate profits via licensing. In the third instance, profits can be generated over shorter times by sales and licensing of patented technologies. But the greatest business success from SNP knowledge may be realized only if and when solid correlations between SNPs and gene function are determined.

A key question is whether the most obvious and rewarding SNPs (from a cost-benefit standpoint) already have been discovered and patented or if latecomers still have a good chance of finding valuable SNPs. [Reported by Anthony Brookes; Center for Genomics Research; Karolinska Institute; Stockholm, Sweden (anthony.brookes@cgr.ki.se)]

Detailed 1999 Meeting Report
Eur. J. Hum. Genet. 8(2), 154-56 (2000)

September 2000 Meeting

Click here, for more information on SNPs.

The electronic form of the newsletter may be cited in the following style:
Human Genome Program, U.S. Department of Energy, Human Genome News (v11n1-2).