Hi,

Wondering why is the pathway list of KEGG truncated in graphite package.

I see that KEGG database has 293 pathways (human), however in graphite the list contains only 242 pathways (to check in R; names(kegg))

Why is the discrepancy ? Or, why are some pathways not included ?

Can anyone help me understand.

thanks much in advance,

Ramya

There are a number of reasons why a KEGG pathway could be missing from graphite.

First of all, take as an example hsa00514 (Other types of O-glycan biosynthesis). The corresponding KGML file does not contain any "relation" or "reaction", so graphite cannot extract any meaningful edge. As a result, the pathway is ignored.

Pathway hsa00460 (Cyanoamino acid metabolism) shows the other two kinds of issues. Some of the edges present in the KGML file link a node to another pathway map (relations with type "maplink"); this is a type of information too weak to actually infer the presence of a specific interaction between two genes. There are then reactions, but these connect compounds to other compounds. graphite prunes them in order to provide a gene-centric view of the pathway. In this specific case no edge survives this transformation, so this pathway becomes empty and is ignored.

Finally, KEGG is free to modify its data at any time, but graphite is updated once every BioC release. As a result, what you are seeing today is actually six-months old data. It is totally possible that the KGML file you are downloading differs from the one we originally used to build the package.