Researchers at the Buck Institute for Age Research have developed a bioinformatics-based method to predict the molecular causes behind inherited genetic diseases. Using a combination of specially designed algorithms, the investigators interrogated available databases listing known sites of protein function to find other possible protein function sites. They then used the algorithms to look at proteins that are associated with disease--causing mutations and searched for statistical co-occurrences of mutations that were close to those functional sites. When they were finished, the team had analyzed some 40,000 amino acid substitutions, making their effort one of the most comprehensive mutation studies.
"There are a few reasons why this is difficult. Genetic data is hard to collect together, so it's a lot of effort to go out [and combine all] of these resources, pulling in all this mutation data, and doing this annotation analysis on them," says Sean Mooney from the Buck. "The second reason this is hard is that tools that predict functional sites or structural sites in proteins are usually developed by individual groups. They aren't inter-operable in any way, so we spent an enormous amount of effort pulling these tools in and making this annotation pipeline."
Mooney and his team have ported their algorithm-based approach into a Web tool, which is now freely available on his lab's website for anyone in the community to use.
The mutation profiling Web-based tool could have other applications, such as helping researchers who manage databases of clinically observed mutations develop hypotheses about what those mutations are doing on a molecular level. It could also help cancer researchers who are re-sequencing tumors to locate mutations that drive tumor growth.
"The functions that we looked at in the paper are relatively narrow in terms of the universe of molecular functions that proteins and genes participate in, and we would like to include those [other functions] in this approach," Mooney says. "Protein mutations make up a small percentage of the genetic variance. We want to expand this and look at other mutations out there."