Dear Martin and Kasper, Thank you very much for the detailed explanation. The real class (LeukoSet was just a test and learning object) seems now to do the magic I wished, and my understanding of s4 oop implementation has grown :). I try now to refine de Class by writing setValidity, show, ect... methods. I'm happy to ask again if I'm get stuck. T: Elmar -----Original Message----- From: Martin Morgan [mailto:mtmorgan@fhcrc.org] Sent: 09 December 2009 20:49 To: Kasper Daniel Hansen Cc: Bucher Elmar; bioconductor at stat.math.ethz.ch Subject: Re: [BioC] implement a additional slot to the eSet class (Formar: Error in .local(.Object, ...) : could not find function "checkClass") Kasper Daniel Hansen wrote: > You might want to detail the dimensions and type of this additional > slot. My guess - based on the name -- is that this can easily be > carried out within the eSet stuff. > > Kasper > > On Dec 9, 2009, at 9:42 AM, <ext-elmar.bucher at="" vtt.fi=""> <ext- elmar.bucher="" at="" vtt.fi=""> wrote: > >> Hi Martin, >> >> Thank you for the prompt "not exported function" hint. It was what >> I want to do :). >> >> To your question: " What specifically do you want to accomplish? " >> >> I like to implement a additional slot to the eSet class. So it >> differs from the code you suggested to study. I list here the >> essential of the code I have written till now. The code is not >> working as I wished. The main problem is, that the additional slot >> (leukoData) shows the correct values inside the initialize-routine >> (observer by browser()), but the output of the initialize-routine >> shows for leukoData just a empty AnnotatedDataFrame. Do you have a >> clue what I do wrong there? Hi Elmar -- The initialize,eSet-method handles your case 'out-of the box', so there is no need for an initialize,LeukoSet-method at all. Specifically ## define LeukoSet class, inhert eSet class setClass("LeukoSet", contains = "eSet", representation=representation(leukoData = "AnnotatedDataFrame"), prototype=prototype(new("VersionedBiobase", versions = c(classVersion("eSet"), classVersion("AnnotatedDataFrame"), LeukoSet = "0.1.0")))) ## generic accessor function for the leukoData slot - abstract data type ADT setGeneric("leukoData", function(object,...) standardGeneric("leukoData")) setMethod("leukoData", signature("LeukoSet"), function(object,...) object at leukoData) and you're ready to go: > a0 <- matrix(0:8, nc = 3, + dimnames=list(c("y0","y1","y2"),c("x0","x1","x2"))) > e1 <- assayDataNew(a0=a0) > pd.df <- data.frame(abc = c("phe_null","phe_eins","phe_zwei"), + numb = c("+0","+1.1","+2"), + row.names = c("x0","x1","x2")) > fd.df <- data.frame(abc = c("fea_null","fea_eins","fea_zwei"), + numb = c("0","1.1","2"), + row.names = c("y0","y1","y2")) > ld.df <- data.frame(abc = c("leo_null","leo_eins","leo_zwei"), + numb = c("-0","-1.1","-2"), + row.names = c("a0","a1","a2")) > mewe.df <- data.frame(labelDescription = c("alpha BETI c","nu MERI c"), + row.names = c("abc","numb")) > pd.s4 <- new("AnnotatedDataFrame", data = pd.df, varMetadata = mewe.df) > fd.s4 <- new("AnnotatedDataFrame", data = fd.df, varMetadata = mewe.df) > ld.s4 <- new("AnnotatedDataFrame", data = ld.df, varMetadata = mewe.df) > miame.s4 <- new('MIAME', name = "Elmar Bucher", + lab = "VTT Technical Research Center of Finalnd - Medical Biotechnology", + contact = "elmar.bucheratvtt.fi", title = "the LeukoSet", + abstract = "crash test", + url = "http://www.petzl.com/en/outdoor/petzl-tv", + other = list(notes = "check it out!")) > lset <- new("LeukoSet", assayData = e1, phenoData = pd.s4, + featureData = fd.s4, leukoData = ld.s4, + experimentData = miame.s4, annnotation = "htsdb") > lset LeukoSet (storageMode: lockedEnvironment) assayData: 3 features, 3 samples element names: a0 phenoData sampleNames: x0, x1, x2 varLabels and varMetadata description: abc: alpha BETI c numb: nu MERI c featureData featureNames: y0, y1, y2 fvarLabels and fvarMetadata description: abc: alpha BETI c numb: nu MERI c experimentData: use 'experimentData(object)' Annotation: > leukoData(lset) An object of class "AnnotatedDataFrame" rowNames: a0, a1, a2 varLabels and varMetadata description: abc: alpha BETI c numb: nu MERI c > In terms of your initialize method... >> >> Elmar >> >> >> /*** begin listing ****/ >> >> library(Biobase) # base library >> >> ### build class and methodes #### >> removeClass("LeukoSet") # only for development reason >> >> # define LeukoSet class, inhert eSet class >> setClass("LeukoSet", contains = "eSet", representation(leukoData = "AnnotatedDataFrame"), prototype(new("VersionedBiobase", versions = c(classVersion("eSet"), classVersion("AnnotatedDataFrame"), LeukoSet = "0.1.0")))) >> >> # manipulate generic functions >> # initialize to set the input to the slots. signature LeukoSet. >> setMethod("initialize", signature("LeukoSet"), >> function(.Object, assayData, leukoData = new("AnnotatedDataFrame"), ... ) { >> #function(.Object, ... ) { >> #.local <- function(.Object, assayData, leukoData = annotatedDataFrameFrom(assayData, byaxis= 'z'), ... ) { >> # ctrl assayData stolen from eSet Initialization method >> mySlots <- slotNames(.Object) >> dotArgs <- list(...) >> isSlot <- names(dotArgs) %in% mySlots >> if (missing(assayData)) assayData <- do.call(assayDataNew, dotArgs[!isSlot], envir = parent.frame()) >> else { >> Biobase:::checkClass(assayData, "AssayData", class(.Object)) >> cat("hallo assayData \n") >> nms <- if (storageMode(assayData) == "list") names(assayData) >> else ls(assayData) >> dupNames <- nms %in% names(dotArgs[!isSlot]) >> if (any(dupNames)) warning("initialize argument(s) '", paste(nms[dupNames], collapse = "' '"), "' also present in 'assayData'; argument(s) ignored") >> } >> # ctrl leukoData >> if (!missing(leukoData)) { >> Biobase:::checkClass(leukoData, "AnnotatedDataFrame", class(.Object)) >> cat("hallo leukoData \n") >> dimLabels(leukoData) <- c("leukoNames", "leukoColumns") >> } >> # pass ctrl to the next methode in the linearization that was computed by the generic function and return >> browser() >> callNextMethod(.Object, assayData = assayData, ...) at this point you pass ... to callNextMethod, but you've 'captured' leukoData in the method signature, so it doesn't get passed to callNextMethod. Consequently, the next method (initialize,eSet-method) returns an object with the default value for this slot -- an empty AnnotatedDataFrame. You could have if (!missing(leukoData)) { ## whatever manipulations you'd like, then... } callNextMethod(.Object, assayData=assayData, leukoData=leukoData, ...) (the assayData check is run anyway, in eSet,initialize-method). The other approach, assuming a leukoData<- method, is along the lines of .Object = callNextMethod(.Object, assayData=assayData, ...) leukoData(.Object) <- leukoData .Object but is likely to involve additional copying. I think a better way to do this is to provide a constructor that does whatever magic you'd like, and then to call new("LeukoSet") with the correct arguments -- it's a cleaner interface for the user, and makes the initialize method less tricky (in hind sight, that would have been a better strategy for ExpressionSet etc.) LeukoSet <- function(assay=matrix(), phenoData=annotatedDataFrameFrom(assay, byrow=FALSE), featureData=annotatedDataFrameFrom(assay, byrow=TRUE), leukoData=annotatedDataFrameFrom(assay, byrow=FALSE), ...) { if (!missing(leukoData)) { ## etc } new("LeukoSet", assayData=assayDataNew(assay=assay), phenoData=phenoData, featureData=featureData, leukoData=leukoData, ...) } LeukoSet() LeukoSet(matrix(rnorm(200), 50, dimnames=list(NULL, letters[1:4]))) Martin >> #.Object <- callNextMethod(.Object, assayData = assayData, ...) >> #do.call(callNextMethod, c(.Object, assayData = assayData, ...)) >> #} >> #.local(.Object, ...) >> }) >> >> # generic accessor function for the leukoData slot - abstract data type ADT >> setGeneric("leukoData", function(object,...) standardGeneric("leukoData")) >> setMethod("leukoData", signature("LeukoSet"), function(object,...) object at leukoData) >> >> # TEST >> # the environment >> e1 = new.env(hash = TRUE) >> e1$a0 <- matrix(0:8, nc = 3) >> rownames(e1$a0) <- c("y0","y1","y2") >> colnames(e1$a0) <- c("x0","x1","x2") >> # the list >> l1 = list(a0 = e1$a0) >> # the matrix >> m1 = e1$a0 >> # annotateddataframes >> pd.df <- data.frame(abc = c("phe_null","phe_eins","phe_zwei"), numb = c("+0","+1.1","+2")) >> fd.df <- data.frame(abc = c("fea_null","fea_eins","fea_zwei"), numb = c("0","1.1","2")) >> ld.df <- data.frame(abc = c("leo_null","leo_eins","leo_zwei"), numb = c("-0","-1.1","-2")) >> rownames(pd.df) <- c("x0","x1","x2") >> rownames(fd.df) <- c("y0","y1","y2") >> rownames(ld.df) <- c("a0","a1","a2") >> mewe.df <- data.frame(labelDescription = c("alpha BETI c","nu MERI c"), row.names = c("abc","numb")) >> pd.s4 <- new("AnnotatedDataFrame", data = pd.df, varMetadata = mewe.df) >> fd.s4 <- new("AnnotatedDataFrame", data = fd.df, varMetadata = mewe.df) >> ld.s4 <- new("AnnotatedDataFrame", data = ld.df, varMetadata = mewe.df) >> # vectors >> a.v <- "htsdb" >> # miame >> miame.s4 <- new('MIAME', name = "Elmar Bucher", lab = "VTT Technical Research Center of Finalnd - Medical Biotechnology", contact = "elmar.bucheratvtt.fi", title = "the LeukoSet", abstract = "crash test", url = "http://www.petzl.com/en/outdoor/petzl-tv", other = list(notes = "check it out!")) >> >> # run >> lset <- new("LeukoSet", assayData = e1, phenoData = pd.s4, featureData = fd.s4, leukoData = ld.s4, experimentData = miame.s4, annnotation = a.v) >> # browser! >> lset >> leukoData(lset) >> >> >> /*** end of listing ***/ >> >> >> >> -----Original Message----- >> From: Martin Morgan [mailto:mtmorgan at fhcrc.org] >> Sent: 08 December 2009 20:55 >> To: Bucher Elmar >> Cc: bioconductor at stat.math.ethz.ch >> Subject: Re: [BioC] Error in .local(.Object, ...) : could not find function "checkClass" >> >> Hi Elmar -- >> >> Ext-Elmar.Bucher at vtt.fi wrote: >>> Dear List, >>> >>> I'm writing on extension of the eSet class. >>> >>> getMethod("initialize","eSet") and getMethod("sampleNames<-","AssayData") from the Biobase package are particularly using functions like "checkClass", "assayDataEnvLock", "assayDataStorageMode". >>> >>> In my implementation I would like to use the same functions but I receive errors like: >>> Error in .local(.Object, ...) : could not find function "checkClass" >>> >>> Why? >>> How can I enable my code to use this functions? Because the eSet initialize code seems to be able to use this functions, whit the very same packages loaded. >>> I have the same problem on linux and windows machines. See sessionInfo below. >> These functions are defined in Biobase but not exported. They are >> accessible with >> >> Biobase:::checkClass >> >> and the like, but probably you do not want to do this. Perhaps you want >> to use callNextMethod() or similar, or the functions that are actually >> exported? What specifically do you want to accomplish? Maybe some hints >> in, e.g., >> >> library(oligoClasses) >> selectMethod(initialize, "SnpCallSetPlus") >> >> If I'd written a class to do something special with 'sampleNames', say >> force them to be upper case, then I might do something like >> >> setClass("A", contains="eSet") >> setReplaceMethod("sampleNames", c("A", "character"), >> function(object, value) >> { >> callNextMethod(object=object, value=toupper(value)) >> }) >> >> and then >> >>> a = new("A", exprs=matrix(rnorm(200), 20)) >>> sampleNames(a) >> [1] "1" "2" "3" "4" "5" "6" "7" "8" "9" "10" >>> sampleNames(a) <- letters[1:10] >>> sampleNames(a) >> [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J" >> >> >> Martin >> >>> Grateful for help, Elmar >>> >>> >>> >>>> sessionInfo() >>> R version 2.10.0 (2009-10-26) >>> x86_64-pc-linux-gnu >>> >>> locale: >>> [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C >>> [3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8 >>> [5] LC_MONETARY=C LC_MESSAGES=en_US.UTF-8 >>> [7] LC_PAPER=en_US.UTF-8 LC_NAME=C >>> [9] LC_ADDRESS=C LC_TELEPHONE=C >>> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C >>> >>> attached base packages: >>> [1] stats graphics grDevices utils datasets methods base >>> >>> other attached packages: >>> [1] RBioinf_1.6.0 graph_1.20.0 Biobase_2.6.0 >>> >>> loaded via a namespace (and not attached): >>> [1] cluster_1.12.1 tools_2.10.0 >>> >>>> sessionInfo() >>> R version 2.10.0 (2009-10-26) >>> i386-pc-mingw32 >>> >>> locale: >>> [1] LC_COLLATE=English_United Kingdom.1252 >>> [2] LC_CTYPE=English_United Kingdom.1252 >>> [3] LC_MONETARY=English_United Kingdom.1252 >>> [4] LC_NUMERIC=C >>> [5] LC_TIME=English_United Kingdom.1252 >>> >>> attached base packages: >>> [1] stats graphics grDevices utils datasets methods base >>> >>> other attached packages: >>> [1] RBioinf_1.6.0 graph_1.24.1 Biobase_2.6.0 >>> >>> loaded via a namespace (and not attached): >>> [1] tools_2.10.0 >>> >>> _______________________________________________ >>> Bioconductor mailing list >>> Bioconductor at stat.math.ethz.ch >>> https://stat.ethz.ch/mailman/listinfo/bioconductor >>> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >> >> -- >> Martin Morgan >> 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 >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at stat.math.ethz.ch >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor > -- Martin Morgan 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