Monday, April 23, 2007
High probability pathways
Researchers at the Wellcome Trust Sanger Institute and colleagues ordered pathways by probability of having at least one driver mutation in a kinase. The researchers merged pathway data from three data sets: Reactome, Panther and INOH. This resulted in 527 unique combinations of genes whose encoded kinases perform documented reactions. By using a t-test and False Discovery Rate statistical analysis of the oncoprobability of the genes contained within each unique gene group, the likelihood that each group contained at least one driver mutation was calculated and the gene groups were ranked accordingly. The merged data were then mined to find each reaction or pathway that involved each exact group.
If a gene group comprised a pathway in its entirety, then it is listed in the "Pathway" column and the associated gene group is listed in the "Genes in pathway" column. If a gene group comprised a component of a pathway, then that group is listed in the "Pathway components" column. Both the pathways and their components are ranked [shown in brackets] by the probability that the gene group has at least one driver mutation. Only the top 47 of the 527 unique gene groups were examined for this table. And three of these 47