dChip on a laptop & problems with R

89 views
Skip to first unread message

Dee Cronin Fenton

unread,
Jul 12, 2012, 7:27:27 AM7/12/12
to dchip-s...@googlegroups.com
Hi Cheng & dChip users!
 
I am using dChip for miRNA profiling analyses.  I first use ANOVA, which runs fine.  Then when I try to do clustering I get absolutely no heatmap.  My colleague has run this on his PC and gets a heatmap from the same data.  My "heatmap" is completely blank!  Please help?  I am also using dChip for linear discriminant analysis, which does not run at all.  I get a series of error messages as shown below.  (The LDA also will not run on my colleague's PC).  Any insights you can offer would be really appreciated.
 
Sincerely,
 
Dee.
 

{Sample classification by Linear Discriminant Analysis

Treat all the 103 arrays as 103 samples and 1 sample groups

Found 2 classes in file C:\Users\Dee\Documents\dChip\dChip_lda_result_p76

Reading gene or SNP list file 'C:\Users\Dee\Documents\dChip\dChip_anova_P76_001'...

Found 3 genes or SNPs

Obtaining classification data for 3 genes and 103 samples...

Using R for LDA...

R> group.id <- group.id + 1

R> train <- as.logical(training)

R> library(MASS)

R> dchip.lda <- lda(dchip, group.id, subset=train)

Function call failed: evaluation stopped because of an error

R> dchip.pred <- predict(dchip.lda, newdata=dchip, dimen=1)

Function call failed: evaluation stopped because of an error

R> dchip.ld <- dchip.pred$x

Function call failed: evaluation stopped because of an error

R> plot(dchip.lda) # not run; you can select this line and run at 'View/Run R codes'

Error in sprintf(ngettext(length(const), "variable %d appears to be constant within groups", :

use format %s for character objects

Error in predict(dchip.lda, newdata = dchip, dimen = 1) :

object "dchip.lda" not found

Error: object "dchip.pred" not found

R> dchip.result <- cbind(dchip.ld, order=1:length(dchip.ld), predict=as.numeric(dchip.pred$class), prob=apply(dchip.pred$posterior, 1, max))

Function call failed: evaluation stopped because of an error

Reply all
Reply to author
Forward
0 new messages