[BioC] Negative Values (?!) loaded by affy package from CEL file of MOE430A

[lgautier at altern.org]

well two ideas then:
- big endian / little endian problem (can you detail the architecture
your are running on ?)
- entanglements with the new features of R to read multibytes charsets.
(were your CEL files generated/edited on a machine with 'exotic' locales ?)




> Hello,
>
> I loaded a few CEL file from MOE430A experiments. Strangely, many PM
> values are negative! As a result, RMA and other routines did not work.
>
> Any ideas?? See scripts below. I am puzzled!
>
> I am running Bioconductor 1.6, R 2.1.0, with moe430acdf loaded.
>
> Thanks!
>
> Simon
> ################################################################
>
>  > result.batch<- ReadAffy(filenames=CEL.fileNames)
>
>  > result.batch
> AffyBatch object
> size of arrays=712x712 features (11886 kb)
> cdf=MOE430A (22690 affyids)
> number of samples=3
> number of genes=22690
> annotation=moe430a
>  >
>
>  > x<-pm (result.batch)
>  > dim (x)
> [1] 249958      3
>  > x[1:10,]
>         P3-real.CEL   P4-real.CEL  P5-real.CEL
>  [1,]  1.001874e-23 -2.117582e-22 4.567573e-08
>  [2,]  1.603000e+03  1.240000e+03 2.319708e-39
>  [3,] -8.589935e+09  1.454000e+02 2.759000e+03
>  [4,]  3.503246e-44  8.100000e+01 4.539706e-08
>  [5,]  5.238000e+02  4.010000e+02 2.319436e-39
>
>  > result.degradation<- AffyRNAdeg(result.batch)
> Error in var(x, na.rm = na.rm) : missing observations in cov/cor
> In addition: There were 50 or more warnings (use warnings() to see the
> first 50)
>  >
>  > MAplot (result.batch)
> Error in rowQ(imat, half) : cannot handle missing values  -- yet
> In addition: Warning message:
> NaNs produced in: log(x, base)
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
>