[Bioc-sig-seq] error with MEDIPS.CpGenrich -> resize +, segfault in MEDIPS.plotCalibrationPlot

[Lukas Chavez Wurm]

Hello,

as recommended by Steve, the best is to switch to R 2.12. and to install 
MEDIPS via biocLite().

However, your MEDIPS.CpGenrich(data=barr, extend=fraglen) induced error 
will remain. This is because the resize() function has no start 
parameter anymore. Therefore, we have updated our MEDIPS.CpGenrich() 
function. The updated version (MEDIPS v. 1.1.2) should be available the 
next days via biocLite(). Until then you can run the MEDIPS.CpGenrich() 
function without specifying the extend parameter. By this, the CpG 
density of your short reads will be tested without an additional 
smoothing. This should be fine also.

Regarding the error that appears in the context of the 
MEDIPS.plotCalibrationPlot() function, I have seen that you try to 
specify a main parameter that does not exist. I have also updated the 
MEDIPS.plotCalibrationPlot() function in order to allow defining custom 
headers for the plots  by a main parameter (also available from version 
1.1.2). Until then try to create the calibration plots without 
specifying a main parameter. I have also seen that you first call the 
png() function and afterwards calibration plots are plotted in a loop. 
Maybe you can try to create the calibration plots one after the other in 
the loop as:

for(bar in calib_list){  
png(paste("Calibration_plot_bar", counter, ".png", sep=""))
MEDIPS.plotCalibrationPlot(data=bar, linearFit=T, plot_chr="all", rpm=T, 
xrange=10)
dev.off()
counter=counter+1
}

All the best,
Lukas


bioc-sig-sequencing-request at r-project.org wrote:
> Date: Fri, 22 Oct 2010 09:48:53 -0400
> From: Steve Lianoglou <mailinglist.honeypot at gmail.com>
> To: Nora Rieber <n.rieber at dkfz-heidelberg.de>
> Cc: "bioc-sig-sequencing at r-project.org"
> 	<bioc-sig-sequencing at r-project.org>
> Subject: Re: [Bioc-sig-seq] error with MEDIPS.CpGenrich -> resize +
> 	segfault in MEDIPS.plotCalibrationPlot
> Message-ID:
> 	<AANLkTims2kJEdSqcuLW-m5YrkursPC0sv3PkwFuY=yY7 at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1
>
> Hi Nora,
>
> On Fri, Oct 22, 2010 at 9:17 AM, Nora Rieber
> <n.rieber at dkfz-heidelberg.de> wrote:
>   
>> Dear all,
>>
>> I am working with the MEDIPS package and get the following error when
>> calling the function MEDIPS.CpGenrich:
>>
>>     
>>> cpgEnrichment_barr<-MEDIPS.CpGenrich(data=barr,extend=fraglen)
>>>       
>> Preprocessing...
>> Error in .local(x, width, fix, use.names, ...) :
>>   unused argument(s) (start = FALSE)
>> Calls: MEDIPS.CpGenrich -> resize -> resize -> .local
>>
>> I checked the resize function which indeed does not require a "start"
>> argument, the function call in MEDIPS.CpGenrich however looks like this:
>>
>> ranges(pos2) <- resize(ranges(pos2), extend + 1, start = FALSE)
>>
>> This sounds like a version problem.
>> This is my sessionInfo:
>>
>> R version 2.11.1 (2010-05-31)
>> x86_64-unknown-linux-gnu
>>     
>
> You're correct in thinking that it's a version problem and I'm pretty
> sure that upgrading to the latest version of R (2.12) will solve your
> problem.
>
> It looks as if the MEDIP package was just released in bioconductor 2.7
> (which is only meant to run on R 2.12), so I guess you installed MEDIP
> by some "other" (not supported) means?
>
> Keep in mind that using biocLite to install your packages is probably
> the best way to ensure that you don't run into versioning problems, so
> go ahead and:
>
> (i) download/install the latest R;
> (ii) biocLite() your way into the base bioconductor install; and
> (iii) biocLite("MEDIP")
>
> You should be good to go after that.
>
> Hope that helps,
> -steve
>