Nov - Paul Thomas¶
Speaker: Paul Thomas
November 20, 2003
Title of talk: Prediction of protein function and dysfunction on a genome-wide scale, and correlation with disease and other phenotypes
Affiliation: Celera Genomics
In the genomic era, one of the fundamental goals is to characterize the function of proteins, and their variants, on a large scale. The PANTHER Protein Classification was developed to associate proteins and their genes with protein families and biological ontology terms (molecular functions and biological processes), linking sequence with function. It is based on expert biologist curation of protein families, such that functional divergence events (subfamilies) are captured as statistical models. This allows not only classification of proteins of known function, but also accurate functional inferences for uncharacterized proteins (View PubMed abstracts ). The PANTHER project was initiated nearly five years ago, and I will discuss the philosophy and lessons that have driven the course of its development. I will also briefly discuss some examples of using the concise set of PANTHER functional categories to analyze large gene sets, such as the set of genes that show evidence of human lineage-specific selection since the last common ancestor with the chimpanzee.
Further, these statistical models can be used to predict the likely functional impact of a missense SNP. Analysis of existing genetic variation/mutation data suggests some general conclusions about the nature of causative cSNPs that underlie both Mendelian and complex disease. It also suggests that it may be feasible to consider a “candidate functional SNP” approach to finding molecular-level changes associated with disease or other phenotypes
*Not this is an industry guest speaker (10 mins)
James DeGreef, November 20, 2003
Proteomics Informatics: Scientific Discovery or IT Development?
This work is licensed under a Creative Commons Attribution-Share Alike 3.0 License.