|
Oxford, U.K., May 15, 2006 — Bioinformatics experts at the University of Oxford and tthe University of Erlengen-Nuremberg have created a series of new programs to meet the growing demand for computing tools to evaluate whole-genome association and linkage studies. The researchers believe that making experimental design and data analysis more
|
straightforward will result in less ambigious interpretations when defining genetic loci for complex diseases.
 Traditionally, complex traits are studied by seeking potential candidate genes, which can overlook the contributions of non-coding regions of the genome to the phenotype. The use of high-density microarrays allows researchers to scan entire genomes to define genetic loci or sets of loci associated with complex phenotypes.
“I think case control studies for whole genome association models are well suited for detecting novel loci,” Cardon said. “We now have the potential to do studies with very large sample sizes and the genotyping capacity for those large samples is becoming available. It’s still expensive, but it’s becoming realistic.”Even so, the ability to perform the statistical analysis and experimental interpretation is hampered by complicated bioinformatics tools, according to Cardon.
“Right now, we’re working very hard to develop simple tools for parsing this amount of data so that we can make some basic conclusions, said Lon Cardon, professor of bioinformatics at the University of Oxford and a principal research fellow at the Wellcome Trust for Human Genetics. “Many groups are writing scripts, but you still need computing expertise to use them. We need something that can handle the high dimensionality of the data easily.” Members of his team have created useful methods for analyzing haplotype data, among others
Dr. Tom Lindner, in the Department of Nephrology and Hypertension at the University of Erlangen-Nuremberg, has collaborated to develop a user-friendly program for linkage analysis.
|
Cardon and Lindner recently spoke to Wes Conard, editor-in-chief of the Affymetrix Microarray Bulletin, about the current state of bioinformatics for microarray analysis, as well as tools they are actively developing to address the increasing complexity of data from larger-scale arrays that are being used in whole-genome association studies for complex traits. The three discussed:
- Creating user-friendly tools for rapid interpretation of large amounts of expression data.
- Designing a stringent, reproducible SNP array experiment.
- Validating data in the face of increasing sample variability
|
New data analysis tools
Conard: What’s available and capable of analyzing 10K, 100K or 500K data for association and linkage studies?
Lindner: I differentiate between association and linkage - here I’m talking about linkage. The idea behind my software, easyLINKAGE, was to make the difficult and tedious set-ups and interpretations of the results easily available for a wide audience of scientists.
The scientists who generate genotypes want to analyze the data right away, without the difficulties of ftp-ing studies and setting up the input files. easyLINKAGE has no problem analyzing 10K data. I just received some 500K data, but I haven’t started analyzing it yet.
Cardon: We’re more focused on association studies, so we’re heading toward the 500K and more markers. My group has made some progress in developing software tools, but none that will scale to the problem at hand. At
|
