what is wrong with the gage?

[wmseq]

Hi every one,
Wen I run the following command, I got the error information as follows:

> fc.kegg.p <- gage(exp.fc, gsets = kegg.gs, ref = NULL, samp = NULL)
Error in if (is.na(spval[i])) tmp[i] <- NA : argument is of length zero

Was it because I analyzed upregulated and downregulated genes at the same time, because the FCs for downregulated genes were negative?

Or, it is the problem of exp.fc file, because it is not a dataframe with the only column for FCs and the rows for GB number, but the first row is GB number, and the second is its corresponding FC value.

Thanks!!!

Richard

[bigmw]

please run following two lines in R, and post the output back here. I may figure out your problem.

lapply(kegg.gs[1:3], head)
head(exp.fc)

[wmseq]

4Hi bigmw,
Thank you very much for your kind help!
I am at home, I could not give you the first few rows of exp.fc.

I could not remember the file name containing the gene IDs and FCs extracted from et$table

I remember it looks as follows:

GB0398 GB4321 GB5432 GB5673
0.4563 -9.6782 3.6542 0.2378
GB0398 GB4321 GB5432 GB5673
0.5631 -1.6782 6.6542 0.8378
------------------------------------
------------------------------------

[bigmw]

What’s GB number used in your data? A type of Gene ID?
Make sure you use the same gene ID as those in kegg.gs. If you use the kegg.gs from gage or gageData package, the gene IDs are Entrez Gene IDs. You may use other common gene IDs like gene symbols etc, but keep the IDs of you gage input data (exp.fc) the same type.

Make sure you understand the basics and common use of gage. You can read through Section 6-7 (Get Started, Basic Analysis), especially page 4-8 of gage in main vignette:

http://bioconductor.org/packages/release/bioc/vignettes/gage/inst/doc/gage.pdf

You may also want to read about gage function (inputs and outputs etc). You can access it within R by:
?gage

[wmseq]

Hi bigmw,
The IDs are genbank accession numbers. Now the problem is that I am working on honey bee, a non model animal, and how can I convert these IDs to Entrez Gene IDs?
Thanks a lot!!

[bigmw]

gage and pathview do support honey bee. However, the primary KEGG gene ID is Enrez Gene. If you check korg data, you will see it:
> library(gage)
> data(korg)
> sel=grep("bee", korg[,3])
> korg[sel,]
kegg.code scientific.name common.name entrez.gnodes
[1,] "ame" "Apis mellifera" "honey bee" "1"
[2,] "tca" "Tribolium castaneum" "red flour beetle" "1"
kegg.geneid ncbi.geneid
[1,] "100576217" "100576217"
[2,] "100141543" "100141543"

Since honey bee is not a major species with Bioconductor gene annotation package as those listed in data(bods) under pathview package. You will have to map your GB numbers to Entrez Gene. I did some research online and found that the GB numbers are more likely to come from BEEBASE.

http://www.ncbi.nlm.nih.gov/nuccore/NC_007073.3?report=genbank&from=9255475&to=9257417

You can download the gene_info data file from NCBI ftp site:
ftp://ftp.ncbi.nlm.nih.gov/gene/DATA...s.gene_info.gz
under unix/linux shell, do:
gunzip All_Invertebrates.gene_info.gz
egrep '(^7460)' All_Invertebrates.gene_info >>am.gene_info.txt

Column 2-6 are (Entrez) GeneID, Symbol, LocusTag, Synonyms, dbXrefs. You see GB numbers in Column 5 and 6. After get the ID mapping, you can use mol.sum in pathview package to merge redundant IDs into a unified expression level.
?pathview::mol.sum
Then you are ready to go.

[wmseq]

Thank you very much, bigmw!
I will do it as you advised.