[Bioc-sig-seq] Error while trying to xyplot. Perhaps a lattice problem.

[ig2ar-saf2 at yahoo.co.uk]

Hi Deepayan,

I had a partial victory with xyplot.

Changing chromosome == "chr19" to chromosome == "chr19.fa" indeed fixed the problem.

I can plot to my screen now doing

xyplot(log(count) ~ nread | sample,
        as(gdapply(E1, islandReadSummary), "data.frame"),
        subset = (chromosome == "chr19.fa" & nread <= 40),
        layout = c(1, 4), pch = 16, type = c("p", "g"))

Its a very nice 4-paneled figure for control 1, control 2, treatment 1 and treatment 2.

The not so successful story is when I try to plot to an eps file. I do

postscript(file = fileOutputGraph,
            horizontal=FALSE,
            onefile=FALSE, # needed for EPS output. It is counterintuitive.
            paper="letter",
            #width=8.5, height=8.5/3,
            width=6, height=8.5,
            colormodel="rgb")

xyplot(log(count) ~ nread | sample,
        as(gdapply(E1, islandReadSummary), "data.frame"),
        subset = (chromosome == "chr19.fa" & nread <= 40),
        layout = c(1, 4), pch = 16, type = c("p", "g"))

but the eps that is produced has a beheaded top panel and the figure becomes only black and white.

I can't report any error messages because there are none. I tried to plot to png or pdf but neither worked, no error messages from R. Only output to eps worked to some extent.

Can you give some pointer to solve it?

Thank you,

Ivan





----- Original Message ----
From: "ig2ar-saf2 at yahoo.co.uk" <ig2ar-saf2 at yahoo.co.uk>
To: Deepayan Sarkar <deepayan.sarkar at gmail.com>
Cc: bioc-sig-sequencing at r-project.org
Sent: Wednesday, 29 April, 2009 0:23:05
Subject: Re: [Bioc-sig-seq] Error while trying to xyplot. Perhaps a lattice problem.


Hello Deepayan,

Thanks for giving me a hand with xyplot.

The output of str(nread.islands) is

> nread.islands <- as(gdapply(E1, islandReadSummary), "data.frame")
> str(nread.islands)
'data.frame':    4886 obs. of  4 variables:
$ nread     : num  1 2 3 4 5 6 7 8 9 10 ...
$ count     : num  80871 12722 2548 666 176 ...
$ chromosome: Factor w/ 21 levels "chr1.fa","chr10.fa",..: 1 1 1 1 1 1 1 1 1 1 ...
$ sample    : Factor w/ 4 levels "C1","C2","T1",..: 1 1 1 1 1 1 1 1 1 1 ...
> 

Aha! I see what you mean. Instead of using

chromosome == "chr19"

I should be using

chromosome == "chr19.fa"

Now I am logged in remotely to my work computer. Bad connection. I'll try it tomorrow and let you know if it works.

Thank you.

Ivan





----- Original Message ----
From: Deepayan Sarkar <deepayan.sarkar at gmail.com>
To: ig2ar-saf2 at yahoo.co.uk
Cc: bioc-sig-sequencing at r-project.org
Sent: Tuesday, 28 April, 2009 19:01:16
Subject: Re: [Bioc-sig-seq] Error while trying to xyplot. Perhaps a lattice  problem.

On Tue, Apr 28, 2009 at 9:58 AM,  <ig2ar-saf2 at yahoo.co.uk> wrote:
>
> Hello everybody,
>
> I am working with the yet unreleased package chipseq.
>
> There is a very nice xyplot that I am trying to reproduce from the
>
> http://www.bioconductor.org/workshops/2009/SeattleJan09/ChIP-seq/ChipSeqWorkflow.pdf
>
> However, R complains saying
>
> "Error in limits..and.aspect(prepanel.default.xyplot, prepanel = prepanel,  :
> need at least one panel"
>
> What am I doing? This
>
>> class(E1)
> [1] "GenomeDataList"
> attr(,"package")
> [1] "BSgenome"
>> names(E1)
> [1] "C1" "C2" "T1" "T2"
>> islandReadSummary <- function(x)
> + {
> + g <- extendReads(x, seqLen = 200)
> + s <- slice(coverage(g, 1, max(end(g))), lower = 1)
> + tab <- table(viewSums(s) / 200)
> + ans <- data.frame(nread = as.numeric(names(tab)), count = as.numeric(tab))
> + ans
> + }
>> nread.islands <- as(gdApply(E1, islandReadSummary), "data.frame")
> There were 50 or more warnings (use warnings() to see the first 50)

Could you share the output of

str(nread.islands)

?

>> xyplot(log(count) ~ nread | sample,
> +        nread.islands,
> +        subset = (chromosome == "chr19" & nread <= 40),
> +        layout = c(1, 4), pch = 16, type = c("p", "g"))
> Error in limits..and.aspect(prepanel.default.xyplot, prepanel = prepanel,  :
>  need at least one panel
>>
>
> Can anybody shed some light?
>
> Clarification: the warning "There were 50 or more warnings (use warnings() to see the first 50)" comes from IRanges. chipseq is using the slightly older start/end convention rather than the newer shift/step convention.
>

Actually, that's happening inside your 'islandReadSummary' function, where

s <- slice(coverage(g, 1, max(end(g))), lower = 1)

should now be

s <- slice(coverage(g), lower = 1)

-Deepayan

> Thank you,
>
> Ivan
>
>
>
>
> _______________________________________________
> Bioc-sig-sequencing mailing list
> Bioc-sig-sequencing at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing
>





_______________________________________________
Bioc-sig-sequencing mailing list
Bioc-sig-sequencing at r-project.org
https://stat.ethz..ch/mailman/listinfo/bioc-sig-sequencing



      
-------------- next part --------------
A non-text attachment was scrubbed...
Name: eraseme.eps
Type: image/x-eps
Size: 12288 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/bioc-sig-sequencing/attachments/20090429/ec48ec5f/attachment-0001.bin>