[BioC] Trouble with gcrma background correction

[Herve Pages]

Hi Sucheta,

Sorry if you didn't get any response on this but I suspect
people won't be able to help you because the 10 lines
of code that you provide can't be run as is.
We should be able to reproduce your problem by simply copy/pasting
your code which is not the case (too many variables are defined somewhere
else e.g 'pms', 'mms', 'pm.affinities', 'mm.affinities', etc...).
This requires some extra work from _you_ (not always easy, I agree,
but necessary to have any chance that someone will look at your code).

H.

Sucheta Tripathy wrote:

> Hi Herve, Group,
> 
> Thanks for your response. As suggested by you, I have installed R version
> 2.4.0 with bioconductor 1.9. I recreated the cdf environment and computed
> affinity data.
> 
> I still get the same error:
> ============================
> Error in model.frame(formula, rownames, variables, varnames, extras,
> extranames,
>   :
>     invalid type (NULL) for variable 'affinities'
> Execution halted
> 
> Following is my sessionInfo():
> ==============================
> R version 2.4.0 (2006-10-03)
> i686-pc-linux-gnu
> 
> locale:
> LC_CTYPE=en_US;LC_NUMERIC=C;LC_TIME=en_US;LC_COLLATE=C;LC_MONETARY=en_US;LC_MESS
> AGES=en_US;LC_PAPER=en_US;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=e
> n_US;LC_IDENTIFICATION=C
> 
> attached base packages:
> [1] "methods"   "stats"     "graphics"  "grDevices" "utils"     "datasets"
> [7] "base"
> 
> And the error comes from this code fragment. Possibly I am passing wrong
> arguement for type fullmodel.
> 
> Code sample:
> ============
> 
> for(i in 1:ncol(pms)){
>  if(verbose) cat(".") if(type=="fullmodel"){
>   cat("type is fullmodel\n")  # ERROR HERE
> 
>     pms[,i] <- bg.adjust.fullmodel(pms[,i],mms[,i],
>                                   pm.affinities,mm.affinities,
>                                   index.affinities,k=k, rho=rho,fast=fast)
> 
>     cat("bg.adjust.fullmodel is done\n")
> 
>  pms[index.affinities,i] <- 2^(log2(pms[index.affinities,i])-
>  fit1$coef[2]*pm.affinities+mean(fit1$coef[2] *pm.affinities))
> 
>  }
> 
> ====
> 
> Any help in solving this will be greatly appreciated.
> 
> many thanks
> 
> Sucheta
> 
> 
> On Fri, November 17, 2006 1:21 pm, Herve Pages wrote:
>> Hi Sucheta,
>>
>>
>> Sucheta Tripathy wrote:
>>
>>> Dear Group,
>>>
>>>
>>> (Sorry for the repost- The earlier one could not get through)
>>> I am stuck with this problem for quite some time. I have been searching
>>> for an answer to this, but could not quite get a hold of it. Any help in
>>>  this will be greatly appreciated.
>>>
>>> I am using R version 2.3.1 and bioconductor version 1.9.
>>>
>> We strongly discourage you of doing this. Bioconductor 1.9 has been
>> designed and tested for R 2.4.z. It is not supported for R 2.3.z users so
>> you are on your own...
>>
>>> First I computed affinity using the following command:
>>>
>>>
>>> affinity.info <- compute.affinities("Soybean", verbose=TRUE)
>>> save(affinity.info,file = "soybean.RData");
>>>
>>> Then I load it using:
>>>
>>>
>>> library(gcrma) load("soybean.RData") source("gcrmaBG.R")
>>>
>>> when I run the script gcrmaBG.R inside another script, I get the
>>> following error:
>>>
>>>
>>> ERROR MESSAGES:
>>> ===============
>>> Adjusting for optical effect.Done.
>>> In first if and type is
>>> Adjusting for non-specific binding
>>> .type is fullmodel
>>> Error in model.frame(formula, rownames, variables, varnames, extras,
>>> extranames, :
>>> invalid variable type for 'affinities' Execution halted
>>>
>>
>> I suggest you upgrade to R 2.4.0 first and see if you can reproduce this
>> error. Also, don't forget to provide the output of sessionInfo() so, among
>>  other things, people know with version of gcrma you are using.
>>
>> One last advice: the long script you provided below doesn't help much
>> because:
>>
>>
>> (1) some lines have been cut in the middle of a variable name
>> if(type=="mm") pms[,i] <-
>> bg.adjust.mm(pms[,i],correction*mms[,i],k=k,fast=f ast) so we can't just
>> copy and paste it into a file and expect it to work without some extra
>> editing work
>>
>> (2) it's too long
>>
>>
>> Please try to reproduce the error in a few lines of code (no more than
>> 10)
>> and send only those 10 lines (with sessionInfo output).
>>
>> Thanks,
>>
>>
>> H.
>>
>>
>>
> 
>