Hello,

I used the kegga function to gather pathway enrichment for my dataset.  I would like to obtain a data frame (or any possible format) containing which genes are "up" and "down" for a given kegg pathway. So for example, I want to select all genes from the "N" column, "up" column" and "down" column for the row "path:dre04020".

I would also like to do this for a gene ontology enrichment table returned from goana.  For example, I would like to select/extract all genes from the N/up/down columns for row "GO:0022613".

I tried using the select function but haven't had much luck. Does anyone know if this can be done?

Thanks,

Matt

This is the sort of thing that we don't provide special functions for in limma, but which you can do reasonably easily for yourself if you are confident in using R.

Suppose you have done a limma linear model analysis resulting in a fit object. You might look at the topTable:

tab <- topTable(fit, n=Inf)

Now you want to a KEGG analysis. You could download the KEGG pathway annotation for zebra fish by:

GK <- getGeneKEGGLinks(species.KEGG = "dre")

Then you could do the KEGG analysis:

k <- kegga(fit, species.KEGG="dre", gene.pathway=GK)
topKEGG(k)

Now if you want to see the top-table genes that belong to a particular pathway (say path:dre01100), it is just a matter of using standard subseting operations:

i <- tab$GeneID %in% GK$GeneID[GK$PathwayID=="path:dre01100"]
tab[i,]

Edit:

The above code assumes your gene IDs are stored in the "GeneID" column. If you have stored them under a different column name, then you need to make the obvious change, such as:

i <- tab$ENTREZID %in% GK$GeneID[GK$PathwayID=="path:dre01100"]
tab[i,]

You might also need to use as.character(tab$ENTREZID) if you have stored the column as a number or as a factor.