Question: Problem in SPIA: de must be a vector of log2 fold changes. The names of de should be included in the refference array!

0

Moritz Kebschull •

**100**wrote:Dear Adi, dear list,
I am trying to run SPIA on some data generated using limma. I am,
however,
running into an error message I donÂ´t understand - my data (at least
to me)
look very similar to the "Colorectal" example data set used in the
vignette.
For the ALL list, I did the following. As ALL_Colorectal, all is a
character vector of Entrez IDs.
> entrezid = mget(ids, illuminaMousev2ENTREZID, ifnotfound = NA)
> all <- as.vector(as.numeric(entrezid))
> all <- na.omit(all) > all <- as.character(all)
> is.vector(all)[1] TRUE
> length(all)[1] 20573
> all[1][1] "66836"
> is.vector(ALL_Colorectal)[1] TRUE> length(ALL_Colorectal)[1] 18045
> data(colorectalcancer)
> ALL_Colorectal[1][1] "3491"
For the DE list, I did.
> pg=topTable(ebFit, coef=1, number=1000, p.value=0.05, lfc=1)
> de_pg <- pg[,5]> names(de_pg)<-pg[,3]
de_pg is a numerical vector of log2 fold changes, with EntrezIDs as
names,
as DE_Colorectal, the example dataset.
> is.vector(de_pg)[1] TRUE
> de_pg[1] 14825
5.189596
> is.vector(DE_Colorectal)[1] TRUE> DE_Colorectal[1] 3491
5.960206
The names of de_pg are in all, similar to the names of DE_Colorectal
being in ALL_Colorectal
> names(de_pg) %in% all [1] TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE
TRUE TRUE etc ...
> names(DE_Colorectal) %in% ALL_Colorectal [1] TRUE TRUE TRUE TRUE
TRUE TRUE TRUE TRUE TRUE TRUE etc ...
However, when I run spia, I get an error I cannot understand
> res<-spia(de=de_pg, all=all,
organism="mmu",beta=NULL,nB=2000,plots=FALSE,
verbose=TRUE,combine="fisher")
Error in spia(de = de_pg, all = all, organism = "mmu", beta = NULL, nB
= 2000, :
de must be a vector of log2 fold changes. The names of de should be
included in the refference array!
What am I doing wrong?
Many thanks for your help,
Moritz
[[alternative HTML version deleted]]

ADD COMMENT
• link
•
modified 6.2 years ago
by
Tarca, Adi •

**560**• written 6.2 years ago by Moritz Kebschull •**100**