[BioC] lmFit Error: "comparison of these types is not implemented"

Gordon Smyth smyth at wehi.EDU.AU
Wed Oct 31 06:40:07 CET 2007


Dear Jose,

The error message you give does not originate from any limma 
function. In fact, I've never seen it before at all.

The problem isn't likely to be with quality of your data in the 
ordinary sense. It's more likely that the data are not want you think 
in some way.

To start finding the problem you might:

1. Use trace() to find in which function the error is occuring

2. Check you data, e.g.

      summary(subGeneIDMatrix[,c(2:3,16:17)])

If you write again, please give sessionInfo() etc.

Best wishes
Gordon

>Date: Mon, 29 Oct 2007 14:32:40 -0700
>From: Jose Rodriguez <jarod07 at ucla.edu>
>Subject: [BioC] lmFit Error: "comparison of these types is not
>         implemented"
>To: bioconductor at stat.math.ethz.ch
>
>I get the following error after executing the following command (using
>limma package for microarray analysis):
>
>  > fit3< lmFit(subGeneIDMatrix[,c(2:3,16:17)],design)
>Error in fit3 < lmFit(subGeneIDMatrix[, c(2:3, 16:17)], design) :
>   comparison of these types is not implemented
>
>The columns of the matrix I refer to are log transformed expression
>sets. The design matrix is simple:
>  > design
>        A      B
>1      1       0
>2      1       0
>3      0       1
>4      0       1
>attr(,"assign")
>[1] 1 1
>attr(,"contrasts")
>attr(,"contrasts")$`factor(c(1, 1, 2, 2))`
>[1] "contr.treatment"
>
>These two sets of data I am trying to compare are part of a larger
>experiment in which the error does not show up. Could the error be a bug
>or does it have something to do with the quality of this particular data
>subset?
>Thanks!



More information about the Bioconductor mailing list