[BioC] GOFunction error

[Wang [guest]]

When I use GOFunction package to analyse my data with 94 interestGenes and 265 refGenes, it report an error:`[<-.data.frame`(`*tmp*`, sigTerm[, 1] %in% noRelationTerm, 7,  : 
  replacement has 1 rows, data has 0
and a warining about in max(i)...
I look into the function, find the sigTerm has nothing:
> sigTerm
[1] goid        name        refnum      interestnum pvalue      adjustp    
<0 rows> (or 0-length row.names)
I use the exampledata to do it again, it runs OK. But I select part of genes in the interestGenes such as x <- interestGenes[1:2225] then use x as interestGenes to run
sigTerm <- GOFunction(x, refGenes, ontology="BP", filename="SigTerm"), it runs OK. The output file sigTerm.bmp has 3908KB while nothing in it.
I run x <- interestGenes[1:2228] and sigTerm <- GOFunction(x, refGenes, ontology="BP", filename="SigTerm") it also report errors like the above for another reason. I am confused why it is only perfect when use the exampledata while others are not. Can you figure out the error and help me to solve the problem?
system:win7
GOFunction:v2.9

 -- output of sessionInfo(): 

Finding statistically significant terms...
Treating for local redundant terms...
错误于`[<-.data.frame`(`*tmp*`, sigTerm[, 1] %in% noRelationTerm, 7,  : 
  replacement has 1 rows, data has 0
此外: 警告信息：
In max(i) : max里所有的参数都不存在；回覆-Inf


Finding statistically significant terms...
Treating for local redundant terms...
错误于if (pc > pcth) sigTerm[sigTerm[, 1] %in% activeChild, 8] <- sigTerm[sigTerm[,  : 
  需要TRUE/FALSE值的地方不可以用缺少值
此外: 警告信息：
In phyper(length(allcsiggenes) - 1, length(sgenes), length(genes) -  :
  产生了NaNs


--
Sent via the guest posting facility at bioconductor.org.