[BioC] Ringo package: error using computeRunningMedians function

[Joern Toedling]

Well, this looks all fine to me. I cannot tell you immediately where the
problem is, and since I have not run into such an error before, we'll have to
search for the source of the problem together. The fact that chromosomes 1 and
10 work but chromosome 11 does not strongly suggests that there is something
unusual with the elements of that chromosome 11 in the probeAnno object.

Further things we can check:
table(ProbeAnno["11.unique"])
table(ProbeAnno["11.index"] %in% featureNames(eSet))

Otherwise, consider setting
options(error=recover)
then when the error occurs, type the frame number of'computeRunningMedians'
and please tell me what the result of
str(chridx)
str(i)
str(chrrm)
are.
Alternatively, if you don't mind you could save the objects eSet and ProbeAnno
into a file (using command 'save') and provide them somewhere to me for
downloading. I could check these things myself then.

Regards,
Joern


On Mon, 31 Aug 2009 14:55:41 +0000, Diego Villar wrote
> Hello, Joern:
> 
>   Thanks a lot for your prompt reply.
> 
> *** How did you build the probeAnno object? ***
> 
>   I build the ProbeAnno object from a table resembling a pos file, using
>  the following:
> 
> ProbeAnno<- posToProbeAnno("alt_prom_array.pos", chrNameColumn =
> "CHROMOSOME",
>    probeColumn = "Name", chrPositionColumn = "POSITION",
>    lengthColumn = "LENGTH", sep="\t", genome="H.Sapiens (hg17)",
>    microarrayPlatform="Agilent alt prom array (custom)")
> 
> The pos file I used is a tab delimited text file which looks like:
> 
> Name	ID	CHROMOSOME	POSITION	COUNT	LENGTH
> chrX.71280574.71280634	FLJ44635_71280558_16	chrX	71280574	1	60
> chr10.116034763.116034823	VWA2_116034286_477	chr10	116034763	1	60
> chr11.383389.383449	PKP3_383245_144	chr11	383389	1	60
> 
> I built it this way since the array is an old Agilent custom array 
> (hg17) and the function that you propose in the Ringo script for 
> creating a probeAnno object from the genes section of an RG did not 
> work on my data.
> 
> The probeAnno generated in this way works fine with chipAlongChrom
> function on a region of chromosome 11, but after that I get the
> mentioned error when using computeRunningMedians.
> 
> chipAlongChrom(eSet, chrom="11", xlim=c(18371851,18373350),
>                ylim=c(-2,4), probeAnno=ProbeAnno, gff=gff)
> 
> smoothX <- computeRunningMedians(eSet, probeAnno=ProbeAnno,
> modColumn = "Cy5", winHalfSize = 400)
> 
> *** Assuming your probeAnno object is called P,what is the output of
> validObject(P), str(P["11.start"]) str(P["11.index"])? ***
> 
> > validObject(ProbeAnno)
> [1] TRUE
> > str(ProbeAnno["11.start"])
>  int [1:16314] 186565 186773 186844 186910 189356 189422 189471 
> 189516 189604 279308 ...
> > str(ProbeAnno["11.index"])
>  chr [1:16314] "chr11.186565.186625" "chr11.186773.186833" ...
> 
> *** And please always remember to add the result of typing
> "sessionInfo()" in
> your R to any question you ask on this list, as this allows us to see
> which
> software versions you are using ***
> 
> > sessionInfo()
> R version 2.9.0 (2009-04-17)
> i386-pc-mingw32
> 
> other attached packages:
> [1] Ringo_1.8.0        Matrix_0.999375-24 lattice_0.17-22    limma_2.18.2
> [5] RColorBrewer_1.0-2 Biobase_2.4.1
> 
> loaded via a namespace (and not attached):
>  [1] annotate_1.22.0     AnnotationDbi_1.6.0 DBI_0.2-4
>  [4] genefilter_1.24.0   grid_2.9.0          RSQLite_0.7-1
>  [7] splines_2.9.0       survival_2.35-4     tools_2.9.0
> [10] xtable_1.5-5
> 
> Thanks again for your help,
> 
>               Diego
> 
> Con fecha 31/8/2009, "Joern Toedling" <Joern.Toedling at curie.fr>
> escribió:
> 
> >Hello,
> >
> >I'm afraid I do not understand the error message. NAs in your ExpressionSet
> >should not be a problem. I suspect that the problem might be with your
> >probeAnno object, especially with the probes on chromosome 11 (since it seems
> >to work with chromosomes 1 and 10). How did you build the probeAnno object?
> >Assuming your probeAnno object is called P,
> >what is the output of
> >validObject(P)
> >str(P["11.start"])
> >str(P["11.index"])
> >? And please always remember to add the result of typing "sessionInfo()" in
> >your R to any question you ask on this list, as this allows us to see which
> >software versions you are using.
> >
> >Regards,
> >Joern
> >
> >On Mon, 31 Aug 2009 11:48:17 +0000, Diego Villar wrote
> >> Dear All,
> >>
> >>   I am trying to use Ringo for analysis of Agilent platform ChIP-
> >> chip. I get the following output/error when applying the
computeRunningMedians
> >> function to my expression set (2 samples, one normal and one dye-
> >> swap):
> >>
> >>   Chromosome 1 ...
> >> 170809ds ... 261108no ...
> >> Chromosome 10 ...
> >> 170809ds ... 261108no ...
> >> Chromosome 11 ...
> >> 170809ds ... Error en newExprs[chridx, i] <- chrrm :
> >>   NAs are not allowed in subscribed assignations
> >>
> >>    Does this mean missing values are not allowed at all for this
> >> function? How could I modify my input so that the function would
> >> work on my data? Would I need to substite NAs somehow in the ExpressionSet?
> >>
> >>    Thanks in advance for your time and help,
> >>
> >>                 Diego Villar

---
Joern Toedling
Institut Curie -- U900
26 rue d'Ulm, 75005 Paris, FRANCE
Tel. +33 (0)156246926