Hi Christian, Chuck (and lists) It seems that the problem may be the strange behaviour of 'unstack' inside a function.
See this thread in the R mailing list: http://tolstoy.newcastle.edu.au/R/help/04/03/1160.html Anyway, I got round the problem by using 'aggregate' instead of converting to a list and then tapply to sum values of metric. Probably more efficient as well. Thanks for the help offered. My function now looks like this (for the record!) and behaves as it should. makeMetric <- function(deMirPresGenes, deMirs){ #need to match position of each miR in deMirPresGenes with its FC to form a vector of FC in correct order fcVector <- as.numeric(with (deMirs, FC[match(deMirPresGenes[,4], Probe)] ) ) #multiply fc by context score for each interaction metric <- fcVector * as.numeric(deMirPresGenes[,11]) geneMetric <- cbind(deMirPresGenes[,2], metric) colnames(geneMetric) <- c('sym', 'metric') #make cumul by aggregate listMetric <- aggregate(as.numeric(geneMetric[,2]), list(geneMetric[,1]), sum)#returns a dataframe colnames(listMetric) <- c('symbol','cumulMetric') #return whole list return(listMetric)# dataframe } Cheers i --- On Sat, 18/12/10, cstrato <cstr...@aon.at> wrote: > From: cstrato <cstr...@aon.at> > Subject: Re: [BioC] problem with function > To: "Iain Gallagher" <iaingallag...@btopenworld.com> > Cc: "bioconductor" <bioconduc...@stat.math.ethz.ch> > Date: Saturday, 18 December, 2010, 14:40 > You need to do: > > cumulMetric <- function(deMirPresGenes, deMirs){ > fc <- deMirs > fcVector <- as.numeric(with (fc, > FC[match(deMirPresGenes[,4], Probe)] ) ) > > metric <- fcVector * > as.numeric(deMirPresGenes[,11]) > geneMetric <- > as.data.frame(cbind(deMirPresGenes[,2], > as.numeric(metric))) > colnames(geneMetric) <- c('y', 'x') > > listMetric <- unstack(geneMetric, x ~ > y) > listMetric <- > as.data.frame(sapply(listMetric,sum)) #returns a dataframe > colnames(listMetric) <- > c('cumulMetric') > > return(listMetric) > } > > Regards > Christian > > On 12/17/10 11:52 PM, Iain Gallagher wrote: > > ok... done. Not really any further forward here. > > > > print statements after creating fcVector, metric and > geneMetric (see output below). They all look ok in terms of > structure and length. But the error persists and listMetric > is not made?!?! Odd. > > > > I have added some comments to the output below. > > > >> tf2<-cumulMetric(tf1, deMirs$up)#deMirs$up is a > dataframe (see prev posts) > > [1] 2.63 2.63 3.13 2.63 3.13 2.74 # print fcVector - > looks ok > > [1] -0.35505 -0.34979 -1.03290 -1.22558 -0.61348 > -0.86584 # print metric - looks ok > > [1] 1045 # lengthof metric - is correct > > sym > metric # print geneMetric - looks ok > > [1,] "AAK1" "-0.35505" > > [2,] "ABCA1" "-0.34979" > > [3,] "ABCA2" "-1.0329" > > [4,] "ABCB10" "-1.22558" > > [5,] "ABCE1" "-0.61348" > > [6,] "ABCF3" "-0.86584" > > [1] 1045 # nrow of geneMetric - is correct > > Error in eval(expr, envir, enclos) : object > 'geneMetric' not found > >> > > > > cheers > > > > i > > --- On Fri, 17/12/10, cstrato<cstr...@aon.at> > wrote: > > > >> From: cstrato<cstr...@aon.at> > >> Subject: Re: [BioC] problem with function > >> To: "Iain Gallagher"<iaingallag...@btopenworld.com> > >> Cc: "bioconductor"<bioconduc...@stat.math.ethz.ch> > >> Date: Friday, 17 December, 2010, 22:38 > >> At the moment I have no idea, but > >> what I would do in this case is to put > >> print() statements after each line to see where it > fails. > >> > >> Christian > >> > >> On 12/17/10 10:59 PM, Iain Gallagher wrote: > >>> Hi > >>> > >>> FC is the second column of the deMirs > variable. deMirs > >> is a dataframe with 2 columns - Probe (e.g. > hsa-miR-145) and > >> FC (e.g 1.45). Using 'with' allows me to use > deMirs as an > >> 'environment'. I thus don't have to pass FC > explicitly. > >>> > >>> Cheers > >>> > >>> i > >>> > >>> --- On Fri, 17/12/10, cstrato<cstr...@aon.at> > >> wrote: > >>> > >>>> From: cstrato<cstr...@aon.at> > >>>> Subject: Re: [BioC] problem with function > >>>> To: "Iain Gallagher"<iaingallag...@btopenworld.com> > >>>> Cc: "bioconductor"<bioconduc...@stat.math.ethz.ch> > >>>> Date: Friday, 17 December, 2010, 20:39 > >>>> What is FC[]? It is not passed > >>>> to the function. Christan > >>>> > >>>> On 12/17/10 8:11 PM, Iain Gallagher > wrote: > >>>>> Sorry. > >>>>> > >>>>> That was a typo. In my script > >> deMirPresGenes1[,4] is > >>>> deMirPresGenes[,4]. > >>>>> > >>>>> Just to be sure I'm going about this > the right > >> way > >>>> though I should say that at the moment I > assign > >> the output > >>>> of another function to a variable called > 'tf1' - > >> this object > >>>> is the same as the deMirPresGenes is my > previous > >> email. > >>>>> > >>>>> This is then fed to my problem > function using > >>>> positional matching. > >>>>> > >>>>> e.g. tf2<- cumulMetric(tf1, > deMirs) > >>>>> > >>>>> Which leads to: > >>>>> > >>>>> Error in eval(expr, envir, enclos) : > object > >>>> 'geneMetric' not found > >>>>> > >>>>> Hey ho! > >>>>> > >>>>> i > >>>>> > >>>>> --- On Fri, 17/12/10, cstrato<cstr...@aon.at> > >>>> wrote: > >>>>> > >>>>>> From: cstrato<cstr...@aon.at> > >>>>>> Subject: Re: [BioC] problem with > function > >>>>>> To: "Iain Gallagher"<iaingallag...@btopenworld.com> > >>>>>> Cc: "bioconductor"<bioconduc...@stat.math.ethz.ch> > >>>>>> Date: Friday, 17 December, 2010, > 18:40 > >>>>>> I am not sure but I would say > that > >>>>>> deMirPresGenes1 does not exist. > >>>>>> > >>>>>> Regards > >>>>>> Christian > >>>>>> > >>>>>> > >>>>>> On 12/17/10 6:42 PM, Iain > Gallagher > >> wrote: > >>>>>>> Hello List > >>>>>>> > >>>>>>> I wonder if someone would help > me with > >> the > >>>> following > >>>>>> function. > >>>>>>> > >>>>>>> cumulMetric<- > >> function(deMirPresGenes, > >>>> deMirs){ > >>>>>>> > >>>>>>> #need to match position of > each miR > >> in > >>>> deMirPresGenes > >>>>>> with its FC to form a vector of FC > in > >> correct > >>>> order > >>>>>>> > fc<- > >> deMirs > >>>>>>> > >> fcVector<- > >>>> as.numeric(with (fc, > >>>>>> FC[match(deMirPresGenes1[,4], > Probe)] ) ) > >>>>>>> > >>>>>>> > #multiply > >> fc by context > >>>> score for > >>>>>> each interaction > >>>>>>> > metric<- > >> fcVector * > >>>>>> as.numeric(deMirPresGenes[,11]) > >>>>>>> > >> geneMetric<- > >>>>>> cbind(deMirPresGenes[,2], > >> as.numeric(metric)) > >>>>>>> > >>>>>>> > >> #make > >>>> cumul > >>>>>> weighted score > >>>>>>> > >> listMetric<- > >>>> unstack(geneMetric, > >>>>>> > >> as.numeric(geneMetric[,2])~geneMetric[,1]) > >>>>>>> > >> listMetric<- > >>>>>> > as.data.frame(sapply(listMetric,sum)) > >> #returns a > >>>> dataframe > >>>>>>> > >> colnames(listMetric)<- > >>>>>> c('cumulMetric') > >>>>>>> > >>>>>>> > #return > >> whole list > >>>>>>> > >> return(listMetric) > >>>>>>> } > >>>>>>> > >>>>>>> deMirPresGenes looks like > this: > >>>>>>> > >>>>>>> Gene.ID > >>>>>> Gene.Symbol > Species.ID > >>>>>> miRNA Site.type > >>>>>> UTR_start UTR_end > >>>>>> X3pairing_contr > >>>>>> local_AU_contr > >>>>>> position_contr > >>>>>> context_score > >> context_percentile > >>>>>>> 22848 AAK1 > >>>>>> 9606 hsa-miR-183 > >>>>>> 2 1546 > >>>>>> 1552 -0.026 > >>>>>> -0.047 0.099 > >>>>>> -0.135 47 > >>>>>>> 19 ABCA1 > >>>>>> 9606 hsa-miR-183 > >>>>>> 2 1366 > >>>>>> 1372 -0.011 > >>>>>> -0.048 0.087 > >>>>>> -0.133 46 > >>>>>>> 20 ABCA2 > >>>>>> 9606 hsa-miR-495 > >>>>>> 2 666 > >>>>>> 672 -0.042 > >>>>>> -0.092 -0.035 > >>>>>> -0.33 93 > >>>>>>> 23456 ABCB10 > >>>>>> 9606 hsa-miR-183 > >>>>>> 3 1475 > >>>>>> 1481 0.003 > >>>>>> -0.109 -0.05 > >>>>>> -0.466 98 > >>>>>>> 6059 ABCE1 > >>>>>> 9606 hsa-miR-495 > >>>>>> 2 1474 > >>>>>> 1480 0.005 > >>>>>> -0.046 0.006 > >>>>>> -0.196 58 > >>>>>>> 55324 ABCF3 > >>>>>> 9606 hsa-miR-1275 > >>>>>> 3 90 > >>>>>> 96 0.007 > >>>>>> 0.042 -0.055 > >>>>>> -0.316 94 > >>>>>>> > >>>>>>> > >>>>>>> The aim of the function is to > extract > >> a > >>>> dataframe of > >>>>>> gene symbols along with a weighted > score > >> from the > >>>> above > >>>>>> data. The weighted score is the FC > column > >> of > >>>> deMirs * the > >>>>>> context_score column of > deMirPresGenes. > >> This is > >>>> easy peasy! > >>>>>>> > >>>>>>> Where I'm falling down is that > if I > >> run this > >>>> function > >>>>>> it complains that 'geneMetric' > can't be > >> found. Hmm > >>>> - I've > >>>>>> run it all line by line (i.e. not > as a > >> function) > >>>> and it > >>>>>> works but wrapped up like this it > fails! > >>>>>>> > >>>>>>> e.g. > >>>>>>> > >>>>>>>> testF2<- > cumulMetric(testF1, > >>>> deMirs$up) > >>>>>>> Error in eval(expr, envir, > enclos) : > >> object > >>>>>> 'geneMetric' not found > >>>>>>> > >>>>>>> deMirs$up looks like this: > >>>>>>> > >>>>>>> Probe FC > >>>>>>> hsa-miR-183 2.63 > >>>>>>> hsa-miR-1275 > 2.74 > >>>>>>> hsa-miR-495 3.13 > >>>>>>> hsa-miR-886-3p > 3.73 > >>>>>>> hsa-miR-886-5p > 3.97 > >>>>>>> hsa-miR-144* > 6.62 > >>>>>>> hsa-miR-451 7.94 > >>>>>>> > >>>>>>> Could someone possibly point > out where > >> I > >>>> falling > >>>>>> down. > >>>>>>> > >>>>>>> Thanks > >>>>>>> > >>>>>>> i > >>>>>>> > >>>>>>>> sessionInfo() > >>>>>>> R version 2.12.0 (2010-10-15) > >>>>>>> Platform: x86_64-pc-linux-gnu > >> (64-bit) > >>>>>>> > >>>>>>> locale: > >>>>>>> > [1] > >>>> LC_CTYPE=en_GB.utf8 > >>>>>> > >> LC_NUMERIC=C > >>>>>>> > [3] > >>>> LC_TIME=en_GB.utf8 > >>>>>> > >> LC_COLLATE=en_GB.utf8 > >>>>>>> > [5] > >> LC_MONETARY=C > >>>>>> > >>>> > LC_MESSAGES=en_GB.utf8 > >>>>>>> > [7] > >>>> LC_PAPER=en_GB.utf8 > >>>>>> > LC_NAME=C > >>>>>>> > [9] > >> LC_ADDRESS=C > >>>>>> > >> LC_TELEPHONE=C > >>>>>>> [11] > LC_MEASUREMENT=en_GB.utf8 > >>>> LC_IDENTIFICATION=C > >>>>>>> > >>>>>>> attached base packages: > >>>>>>> [1] stats > >> graphics > >>>>>> grDevices utils > >> datasets > >>>>>> methods base > >>>>>>> > >>>>>>> loaded via a namespace (and > not > >> attached): > >>>>>>> [1] tools_2.12.0 > >>>>>>>> > >>>>>>> > >>>>>>> > >>>> > _______________________________________________ > >>>>>>> Bioconductor mailing list > >>>>>>> bioconduc...@r-project.org > >>>>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor > >>>>>>> Search the archives: > >>>>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor > >>>>>>> > >>>>>> > >>>>> > >>>>> > >> _______________________________________________ > >>>>> Bioconductor mailing list > >>>>> bioconduc...@r-project.org > >>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor > >>>>> Search the archives: > >>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor > >>>>> > >>>> > >>> > >> > > > ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
