[BioC] gdata affects LumiBatch::combine method

Mark Cowley m.cowley at garvan.org.au
Thu Oct 11 07:56:16 CEST 2012 

Hi again,
Your workaround definitely works in an interactive session, but I was expecting to fix the problem by fixing it within lumi::combine… 

I rebuilt & installed lumi after removing the recursion within lumi's combine, however even when explicitly calling lumi::combine I still got the error. it turns out that its the recursion that's still present within the BiocGenerics::combine function that throws the error, despite the signature(c("LumiBatch", "LumiBatch")) not being defined there…

I've clearly got plenty of learning still to do with S4...

cheers,
Mark

I changed code from:
setMethod("combine", signature=c(x="LumiBatch", y="LumiBatch"), function(x, y, ...) 
{
   if (missing(y)) return(x)
   if (length(list(...)) > 0L) {
       callGeneric(x, do.call(callGeneric, list(y, ...)))
   …

to:
setMethod("combine", signature=c(x="LumiBatch", y="LumiBatch"), function(x, y, ...) 
{
	if (missing(y)) return(x)
	if (length(list(...)) > 0) {
		res <- combine(x, y)
		extras <- list(...)
		for(i in seq(along=extras)) {
			res <- combine(res, extras[[i]])
		}
		return( res )
	} 
# ….

build, install,
> library(lumi)
> data(example.lumi)
> lumi::combine(example.lumi[,1], example.lumi[,2], example.lumi[,3])
Error in callGeneric(x, do.call(callGeneric, list(y, ...))) : 
  'callGeneric' must be called from a generic function or method
> traceback()
3: stop("'callGeneric' must be called from a generic function or method")
2: callGeneric(x, do.call(callGeneric, list(y, ...)))
1: lumi::combine(example.lumi[, 1], example.lumi[, 2], example.lumi[, 
       3])
> lumi::combine
nonstandardGenericFunction for "combine" defined from package "BiocGenerics"

function (x, y, ...) 
{
    if (length(list(...)) > 0L) {
        callGeneric(x, do.call(callGeneric, list(y, ...)))
    }
    else {
        standardGeneric("combine")
    }
}
<environment: 0xc69b2c>
### ie, the one that's defined in BiocGenerics


On 11/10/2012, at 3:33 PM, Martin Morgan <mtmorgan at fhcrc.org> wrote:

> On 10/10/2012 09:25 PM, Mark Cowley wrote:
>> Thanks Martin,
>> Unfortunately, explicitly specifying lumi::combine seems to interfere with the S4 calling mechanism in an interactive session:
>> 
>> ### only lumi
>> library(lumi)
>> load("Rmisc/n525.RDa")
>> load("Rmisc/s217.RDa")
>> load("Rmisc/s295.RDa")
>> a <- lumi::combine(s217, n525, s295)
>> Error in callGeneric(x, do.call(callGeneric, list(y, ...))) :
>>   'callGeneric' must be called from a generic function or method
> 
> I can replicate that with example(LumiBatch) and then
> 
>  lumi::combine(example.lumi[,1], example.lumi[,2], example.lumi[,3])
> 
> I'd say this was an issue in methods, and the work-around (again!) would be to make the recursion here
> 
> > lumi::combine
> nonstandardGenericFunction for "combine" defined from package "BiocGenerics"
> 
> function (x, y, ...)
> {
>    if (length(list(...)) > 0L) {
>        callGeneric(x, do.call(callGeneric, list(y, ...)))
>    }
>    else {
>        standardGeneric("combine")
>    }
> }
> <environment: 0x1f53748>
> 
> into an iteration:
> 
>  res = example.lumi[,1]
>  for (i in 2:3)
>      res = combine(res, example.lumi[,i])
> 
> Martin
> 
>> 
>> a <- combine(s217, n525, s295)
>> class(a)
>> # [1] "LumiBatch"
>> 
>> ##### gdata then lumi
>> library(gdata)
>> library(lumi)
>> load("Rmisc/n525.RDa")
>> load("Rmisc/s217.RDa")
>> load("Rmisc/s295.RDa")
>> a <- lumi::combine(s217, n525, s295)
>> Error in callGeneric(x, do.call(callGeneric, list(y, ...))) :
>>   'callGeneric' must be called from a generic function or method
>> 
>> a <- combine(s217, n525, s295)
>> class(a)
>> # [1] "LumiBatch"
>> 
>> #### lumi then gdata
>> library(lumi)
>> library(gdata)
>> load("Rmisc/n525.RDa")
>> load("Rmisc/s217.RDa")
>> load("Rmisc/s295.RDa")
>> a <- lumi::combine(s217, n525, s295)
>> Error in callGeneric(x, do.call(callGeneric, list(y, ...))) :
>>   'callGeneric' must be called from a generic function or method
>> 
>> a <- combine(s217, n525, s295)
>> class(a)
>> # [1] "data.frame"
>> 
>>> sessionInfo()
>> R version 2.15.1 (2012-06-22)
>> Platform: x86_64-unknown-linux-gnu (64-bit)
>> 
>> locale:
>>  [1] LC_CTYPE=en_AU.UTF-8       LC_NUMERIC=C
>>  [3] LC_TIME=en_AU.UTF-8        LC_COLLATE=en_AU.UTF-8
>>  [5] LC_MONETARY=en_AU.UTF-8    LC_MESSAGES=en_AU.UTF-8
>>  [7] LC_PAPER=C                 LC_NAME=C
>>  [9] LC_ADDRESS=C               LC_TELEPHONE=C
>> [11] LC_MEASUREMENT=en_AU.UTF-8 LC_IDENTIFICATION=C
>> 
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods
>> [7] base
>> 
>> other attached packages:
>> [1] gdata_2.11.0       lumi_2.8.0         nleqslv_1.9.3
>> [4] methylumi_2.2.0    ggplot2_0.9.1      reshape2_1.2.1
>> [7] scales_0.2.1       Biobase_2.16.0     BiocGenerics_0.2.0
>> 
>> loaded via a namespace (and not attached):
>>  [1] affy_1.34.0           affyio_1.24.0
>>  [3] annotate_1.34.1       AnnotationDbi_1.18.1
>>  [5] bigmemory_4.2.11      BiocInstaller_1.4.7
>>  [7] Biostrings_2.24.1     bitops_1.0-4.1
>>  [9] BSgenome_1.24.0       colorspace_1.1-1
>> [11] DBI_0.2-5             dichromat_1.2-4
>> [13] digest_0.5.2          DNAcopy_1.30.0
>> [15] GenomicRanges_1.8.12  genoset_1.6.0
>> [17] grid_2.15.1           gtools_2.7.0
>> [19] hdrcde_2.16           IRanges_1.14.4
>> [21] KernSmooth_2.23-8     labeling_0.1
>> [23] lattice_0.20-10       MASS_7.3-20
>> [25] Matrix_1.0-6          memoise_0.1
>> [27] mgcv_1.7-20           munsell_0.3
>> [29] nlme_3.1-104          plyr_1.7.1
>> [31] preprocessCore_1.18.0 proto_0.3-9.2
>> [33] RColorBrewer_1.0-5    RCurl_1.91-1
>> [35] Rsamtools_1.8.6       RSQLite_0.11.1
>> [37] rtracklayer_1.16.3    stats4_2.15.1
>> [39] stringr_0.6.1         tools_2.15.1
>> [41] XML_3.9-4             xtable_1.7-0
>> [43] zlibbioc_1.2.0
>> 
>> 
>> Reordering my package loads so that lumi is last, and just using 'combine' works.
>> 
>> cheers,
>> Mark
>> 
>> On 10/10/2012, at 7:06 AM, Martin Morgan <mtmorgan at fhcrc.org> wrote:
>> 
>>> On 10/08/2012 09:40 PM, Mark Cowley wrote:
>>>> Dear list,
>>>> i'm trying to combine two LumiBatch objects, using the S4 combine("LumiBatch","LumiBatch") method, defined originally in BiocGenerics, and enhanced within lumi.
>>>> In a fresh session, with just lumi loaded, everything works fine, i.e. I get a LumiBatch object.
>>>> As soon as I load gdata, this breaks the combine method, instead returning a data.frame with very odd dimensions. This is true if I load gdata before, or after lumi. Note that showMethods("lumi") doesn't change after loading gdata. Any ideas?
>>> 
>>> Hi Mark -- unfortunately, this is not easily avoidable. gdata (a CRAN  package) defines its own function combine that masks the version used by lumi and other Bioconductor packages. The work-around is to invoke lumi's combine with, e.g.,
>>> 
>>> lumi::combine(x = example.lumi[, 1], y = example.lumi[, 3])
>>> 
>>> If gdata were in Bioconductor, we could coerce the author into writing a method for combine, as defined in BiocGenerics. If you were writing a package, then you could importMethodsFrom(lumi, combine) and use combine safely in your own package code whether or not gdata was on the search path.
>>> 
>>> Martin
>>> 
>>>> cheers,
>>>> Mark
>>>> 
>>>> ### lumi, then gdata
>>>>> suppressPackageStartupMessages(library(lumi))
>>>> Warning messages:
>>>> 1: found methods to import for function ‘eapply’ but not the generic itself
>>>> 2: replacing previous import ‘image’ when loading ‘graphics’
>>>>> load("Rmisc/n525.RDa")
>>>>> load("Rmisc/s295.RDa")
>>>>> showMethods("combine")
>>>> Function: combine (package BiocGenerics)
>>>> x="AnnotatedDataFrame", y="AnnotatedDataFrame"
>>>> x="ANY", y="missing"
>>>> x="AssayData", y="AssayData"
>>>> x="data.frame", y="data.frame"
>>>> x="eSet", y="eSet"
>>>> x="ExpressionSet", y="LumiBatch"
>>>> x="LumiBatch", y="ExpressionSet"
>>>> x="LumiBatch", y="LumiBatch"
>>>> x="matrix", y="matrix"
>>>> x="MethyLumiM", y="MethyLumiM"
>>>> x="MethyLumiQC", y="MethyLumiQC"
>>>> x="MethyLumiSet", y="MethyLumiSet"
>>>> x="MIAME", y="MIAME"
>>>> 
>>>>> a <- combine(n525, s295)
>>>>> class(a)
>>>> [1] "LumiBatch"
>>>> attr(,"package")
>>>> [1] "lumi"
>>>> ## the expected output
>>>> 
>>>>> suppressPackageStartupMessages(library(gdata))
>>>>> showMethods("combine")
>>>> Function: combine (package BiocGenerics)
>>>> x="AnnotatedDataFrame", y="AnnotatedDataFrame"
>>>> x="ANY", y="missing"
>>>> x="AssayData", y="AssayData"
>>>> x="data.frame", y="data.frame"
>>>> x="environment", y="environment"
>>>>     (inherited from: x="AssayData", y="AssayData")
>>>> x="eSet", y="eSet"
>>>> x="ExpressionSet", y="LumiBatch"
>>>> x="LumiBatch", y="ExpressionSet"
>>>> x="LumiBatch", y="LumiBatch"
>>>> x="matrix", y="matrix"
>>>> x="MethyLumiM", y="MethyLumiM"
>>>> x="MethyLumiQC", y="MethyLumiQC"
>>>> x="MethyLumiSet", y="MethyLumiSet"
>>>> x="MIAME", y="MIAME"
>>>> 
>>>>> a <- combine(n525, s295)
>>>>> class(a)
>>>> [1] "data.frame"
>>>>> 
>>>> 
>>>> 
>>>> #### gdata then lumi
>>>>> suppressPackageStartupMessages(library(gdata))
>>>>> suppressPackageStartupMessages(library(lumi))
>>>> Warning messages:
>>>> 1: found methods to import for function ‘eapply’ but not the generic itself
>>>> 2: replacing previous import ‘image’ when loading ‘graphics’
>>>>> load("Rmisc/n525.RDa")
>>>>> load("Rmisc/s295.RDa")
>>>>> showMethods("combine")
>>>> Function: combine (package BiocGenerics)
>>>> x="AnnotatedDataFrame", y="AnnotatedDataFrame"
>>>> x="ANY", y="missing"
>>>> x="AssayData", y="AssayData"
>>>> x="data.frame", y="data.frame"
>>>> x="eSet", y="eSet"
>>>> x="ExpressionSet", y="LumiBatch"
>>>> x="LumiBatch", y="ExpressionSet"
>>>> x="LumiBatch", y="LumiBatch"
>>>> x="matrix", y="matrix"
>>>> x="MethyLumiM", y="MethyLumiM"
>>>> x="MethyLumiQC", y="MethyLumiQC"
>>>> x="MethyLumiSet", y="MethyLumiSet"
>>>> x="MIAME", y="MIAME"
>>>> 
>>>>> a <- combine(n525, s295)
>>>>> class(a)
>>>> [1] "LumiBatch"
>>>> attr(,"package")
>>>> [1] "lumi"
>>>> 
>>>> 
>>>>> sessionInfo()
>>>> R version 2.15.1 (2012-06-22)
>>>> Platform: i386-apple-darwin9.8.0/i386 (32-bit)
>>>> 
>>>> locale:
>>>> [1] en_AU.UTF-8/en_AU.UTF-8/en_AU.UTF-8/C/en_AU.UTF-8/en_AU.UTF-8
>>>> 
>>>> attached base packages:
>>>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>>> 
>>>> other attached packages:
>>>> [1] lumi_2.10.0        nleqslv_1.9.4      Biobase_2.18.0     BiocGenerics_0.4.0
>>>> [5] gdata_2.12.0
>>>> 
>>>> loaded via a namespace (and not attached):
>>>>  [1] affy_1.36.0           affyio_1.26.0         annotate_1.36.0
>>>>  [4] AnnotationDbi_1.20.0  BiocInstaller_1.8.1   colorspace_1.1-1
>>>>  [7] DBI_0.2-5             grid_2.15.1           gtools_2.7.0
>>>> [10] IRanges_1.16.2        KernSmooth_2.23-8     lattice_0.20-10
>>>> [13] MASS_7.3-22           Matrix_1.0-9          methylumi_2.4.0
>>>> [16] mgcv_1.7-21           nlme_3.1-104          parallel_2.15.1
>>>> [19] preprocessCore_1.20.0 RSQLite_0.11.2        stats4_2.15.1
>>>> [22] XML_3.95-0.1          xtable_1.7-0          zlibbioc_1.4.0
>>>> 
>>>> _______________________________________________
>>>> Bioconductor mailing list
>>>> Bioconductor at r-project.org
>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>>> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>> 
>>> 
>>> 
>>> --
>>> Computational Biology / Fred Hutchinson Cancer Research Center
>>> 1100 Fairview Ave. N.
>>> PO Box 19024 Seattle, WA 98109
>>> 
>>> Location: Arnold Building M1 B861
>>> Phone: (206) 667-2793
>> 
> 
> 
> -- 
> Computational Biology / Fred Hutchinson Cancer Research Center
> 1100 Fairview Ave. N.
> PO Box 19024 Seattle, WA 98109
> 
> Location: Arnold Building M1 B861
> Phone: (206) 667-2793

More information about the Bioconductor mailing list