A useful strategy in determining the relationship between protein structure and function is to examine proteins that are distantly related but within the same superfamily. This practice provides a conceptual framework for correlating protein function with the structural scaffold, and as such, provides a way to predict function from sequence in cases where it cannot be inferred by simple analogy from the function of a closely related homolog. The utility and practicality of this methodology increases with the growth of protein sequence databases, provided the distant members of a protein superfamily are identifiable. Unfortunately, distant relations are often difficult to find using current database screening methods. Sequence similarity search tools often fail to report distant relationships at statistically significant scores, and motif searches are typically unable to find the most distant relatives of a superfamily. Through the use of simple congruence analysis, however, it is possible to identify the distant relations in the """"""""noise"""""""" region (the region of statistically insignifican scores) of sequence similarity searches. We are currently developing programs to implement this analysis in collaboration with Conrad Huang, Tom Ferrin, Chris Botka and Teri Klein of the Computer Graphics Laboratory, that will be used it to identify new distant protein superfamily members.
Showing the most recent 10 out of 508 publications