Hi all, I am trying to apply frma with an alternative Brainarray CDF to a microarray dataset. I cannot get this to work and found a relevant post on the bioC mailing list: http://article.gmane.org/gmane.science. biology.informatics.conductor/47887/match=frma+alternative+cdf I get the same error as mentioned on the post: > test <- ReadAffy() > test at cdfName <- "hgu133plus2hsentrezgcdf" > library(frma) > test2 <- frma(test) Error in rcModelWPLM(y = x1, w = w.tmp, row.effects = pe.tmp, input.scale = x4) : row.effects should sum to zero > test2 <- frma(test, summarize = "random_effect") Error in if (convi < 1e-04) break : missing value where TRUE/FALSE needed I get the same error with the most recent version of the brain array CDF (18.0.0). The suggested solution on the page mentioned above (brain array version 15 and hgu133afrmavecs version 1.2.0) dord not solve the problem for me. Does anyone know which versions of the brainarray CDFs, and hgu133afrmavecs go together? thanks a lot, Tine -- output of sessionInfo(): > sessionInfo() R version 3.0.3 (2014-03-06) Platform: x86_64-apple-darwin10.8.0 (64-bit) locale: [1] C/UTF-8/C/C/C/C attached base packages: [1] parallel stats graphics grDevices utils datasets methods [8] base other attached packages: [1] hgu133plus2frmavecs_1.2.0 frma_1.12.0 [3] hgu133plus2hsentrezgcdf_15.0.0 AnnotationDbi_1.22.6 [5] affy_1.38.1 Biobase_2.20.1 [7] BiocGenerics_0.6.0 loaded via a namespace (and not attached): [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 [7] RSQLite_0.11.4 affxparser_1.32.3 affyio_1.28.0 [10] bit_1.1-12 codetools_0.2-8 ff_2.2-13 [13] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 [16] oligoClasses_1.22.0 preprocessCore_1.22.0 splines_3.0.3 [19] stats4_3.0.3 tools_3.0.3 zlibbioc_1.6.0 -- Sent via the guest posting facility at bioconductor.org.

Tine, I'm a bit behind on updating the frmavecs alternative CDFs. Try using: hgu133plus2frmavecs_1.3.0 with hgu133plus2hsentrezgcdf_16.0.0 A couple other things to try: -- update to the latest version of preprocessCore (that's where rcModelWPLM comes from). -- not sure if this matters, but I always pass the cdfname to ReadAffy() instead of changing it afterwards Best, Matt On Sun, Jul 27, 2014 at 3:37 PM, Tine Casneuf [guest] < guest@bioconductor.org> wrote: > Hi all, > > I am trying to apply frma with an alternative Brainarray CDF to a > microarray dataset. I cannot get this to work and found a relevant post on > the bioC mailing list: > http://article.gmane.org/gmane.science.biology.informatics.conductor /47887/match=frma+alternative+cdf > > I get the same error as mentioned on the post: > > > test <- ReadAffy() > > test@cdfName <- "hgu133plus2hsentrezgcdf" > > library(frma) > > test2 <- frma(test) > Error in rcModelWPLM(y = x1, w = w.tmp, row.effects = pe.tmp, input.scale > = x4) : > row.effects should sum to zero > > test2 <- frma(test, summarize = "random_effect") > Error in if (convi < 1e-04) break : missing value where TRUE/FALSE needed > > > I get the same error with the most recent version of the brain array CDF > (18.0.0). The suggested solution on the page mentioned above (brain array > version 15 and hgu133afrmavecs version 1.2.0) dord not solve the problem > for me. > > Does anyone know which versions of the brainarray CDFs, and > hgu133afrmavecs go together? > > thanks a lot, > > Tine > > -- output of sessionInfo(): > > > sessionInfo() > R version 3.0.3 (2014-03-06) > Platform: x86_64-apple-darwin10.8.0 (64-bit) > > locale: > [1] C/UTF-8/C/C/C/C > > attached base packages: > [1] parallel stats graphics grDevices utils datasets methods > [8] base > > other attached packages: > [1] hgu133plus2frmavecs_1.2.0 frma_1.12.0 > [3] hgu133plus2hsentrezgcdf_15.0.0 AnnotationDbi_1.22.6 > [5] affy_1.38.1 Biobase_2.20.1 > [7] BiocGenerics_0.6.0 > > loaded via a namespace (and not attached): > [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 > [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 > [7] RSQLite_0.11.4 affxparser_1.32.3 affyio_1.28.0 > [10] bit_1.1-12 codetools_0.2-8 ff_2.2-13 > [13] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 > [16] oligoClasses_1.22.0 preprocessCore_1.22.0 splines_3.0.3 > [19] stats4_3.0.3 tools_3.0.3 zlibbioc_1.6.0 > > > -- > Sent via the guest posting facility at bioconductor.org. > -- Matthew N McCall, PhD 112 Arvine Heights Rochester, NY 14611 Cell: 202-222-5880 [[alternative HTML version deleted]]

Yes, the intention is to be one version behind the latest Brainarray version. So hgu133plus2frmavecs_1.3.0 should really go with hgu133plus2hsentrezgcdf_17.0.0, but this year has been very hectic, so I missed an update. Best, Matt On Mon, Jul 28, 2014 at 3:55 AM, Tine Casneuf <tineke.casneuf@gmail.com> wrote: > Hi Matt, > > thanks a lot, this works indeed (hgu133plus2frmavecs_1.3.0 with > hgu133plus2hsentrezgcdf_16.0.0). Is your intention to keep updating these > to be able to work with the more recent versions of Brainarray? > > thanks again, > > tine > > > > On Sun, Jul 27, 2014 at 11:26 PM, Matthew McCall <mccallm@gmail.com> > wrote: > >> Tine, >> >> I'm a bit behind on updating the frmavecs alternative CDFs. Try using: >> hgu133plus2frmavecs_1.3.0 with hgu133plus2hsentrezgcdf_16.0.0 >> >> A couple other things to try: >> -- update to the latest version of preprocessCore (that's where >> rcModelWPLM comes from). >> -- not sure if this matters, but I always pass the cdfname to ReadAffy() >> instead of changing it afterwards >> >> Best, >> Matt >> >> >> >> >> On Sun, Jul 27, 2014 at 3:37 PM, Tine Casneuf [guest] < >> guest@bioconductor.org> wrote: >> >>> Hi all, >>> >>> I am trying to apply frma with an alternative Brainarray CDF to a >>> microarray dataset. I cannot get this to work and found a relevant post on >>> the bioC mailing list: >>> http://article.gmane.org/gmane.science.biology.informatics.conduct or/47887/match=frma+alternative+cdf >>> >>> I get the same error as mentioned on the post: >>> >>> > test <- ReadAffy() >>> > test@cdfName <- "hgu133plus2hsentrezgcdf" >>> > library(frma) >>> > test2 <- frma(test) >>> Error in rcModelWPLM(y = x1, w = w.tmp, row.effects = pe.tmp, >>> input.scale = x4) : >>> row.effects should sum to zero >>> > test2 <- frma(test, summarize = "random_effect") >>> Error in if (convi < 1e-04) break : missing value where TRUE/FALSE needed >>> >>> >>> I get the same error with the most recent version of the brain array CDF >>> (18.0.0). The suggested solution on the page mentioned above (brain array >>> version 15 and hgu133afrmavecs version 1.2.0) dord not solve the problem >>> for me. >>> >>> Does anyone know which versions of the brainarray CDFs, and >>> hgu133afrmavecs go together? >>> >>> thanks a lot, >>> >>> Tine >>> >>> -- output of sessionInfo(): >>> >>> > sessionInfo() >>> R version 3.0.3 (2014-03-06) >>> Platform: x86_64-apple-darwin10.8.0 (64-bit) >>> >>> locale: >>> [1] C/UTF-8/C/C/C/C >>> >>> attached base packages: >>> [1] parallel stats graphics grDevices utils datasets methods >>> [8] base >>> >>> other attached packages: >>> [1] hgu133plus2frmavecs_1.2.0 frma_1.12.0 >>> [3] hgu133plus2hsentrezgcdf_15.0.0 AnnotationDbi_1.22.6 >>> [5] affy_1.38.1 Biobase_2.20.1 >>> [7] BiocGenerics_0.6.0 >>> >>> loaded via a namespace (and not attached): >>> [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 >>> [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 >>> [7] RSQLite_0.11.4 affxparser_1.32.3 affyio_1.28.0 >>> [10] bit_1.1-12 codetools_0.2-8 ff_2.2-13 >>> [13] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 >>> [16] oligoClasses_1.22.0 preprocessCore_1.22.0 splines_3.0.3 >>> [19] stats4_3.0.3 tools_3.0.3 zlibbioc_1.6.0 >>> >>> >>> -- >>> Sent via the guest posting facility at bioconductor.org. >>> >> >> >> >> -- >> Matthew N McCall, PhD >> 112 Arvine Heights >> Rochester, NY 14611 >> Cell: 202-222-5880 >> >> > -- Matthew N McCall, PhD 112 Arvine Heights Rochester, NY 14611 Cell: 202-222-5880 [[alternative HTML version deleted]]

Happy to help. The Gene Expression Barcode (barcode function in the frma package) is only implemented for the Affy CDF. It should produce an error when you try to barcode the output from frma using an alternative CDF, but I believe you overrode the error by supplying the argument -- platform = "GPL570". Best, Matt On Mon, Jul 28, 2014 at 9:58 AM, Tine Casneuf <tineke.casneuf@gmail.com> wrote: > Hi Matt, > > thanks in advance. Not a problem, it works with the previous version. > > Can i bother you with another question? With an example dataset, I tried > to estimate which genes are expressed, but found an substantial difference > between using the Affy CDF versus the alternative one: > > library(frma) > > library(affy) > > library(GEOquery) > > getGEOSuppFiles("GSE20986") > > dir("GSE20986/") > > untar("GSE20986/GSE20986_RAW.tar", exdir="data") > > cels <- list.files("data/", pattern = "[gz]") > > sapply(paste("data", cels, sep="/"), gunzip) > > myCELs <- paste("data/", dir("data/", pattern = ".CEL$"), sep = "") > > > ## affy CDF > > batch <- ReadAffy(filenames = myCELs) > > eset <- frma(batch, summarize = "random_effect") > > apc <- barcode(eset, cutoff = (-log10(0.05)), output = "binary") > > selPbs <- sum(rowSums(apc) > 1) > > selPbs > > > ## alternative CDF > > batchBA <- ReadAffy(filenames = myCELs, cdfname = "hgu133plus2hsentrezg") > > esetBA <- frma(batchBA, summarize = "random_effect") > > apcBA <- barcode(esetBA, cutoff = (-log10(0.05)), output = "binary", > platform = "GPL570") > > selPbsBA <- sum(rowSums(apcBA) > 1) > > selPbsBA > > > selPbs ## the Affy CDF > > [1] 30117 > > selPbsBA ## the brain array CDF > > [1] 117 > > Would you happen to know why with the Affy CDF the number of expressed > genes is ~30.000 and with the alternative only ~100 genes, 300 times lower? > > Thanks a lot, > Best, > > Tine > > > As before, > sessionInfo() > > R version 3.0.3 (2014-03-06) > > Platform: x86_64-apple-darwin10.8.0 (64-bit) > > > locale: > > [1] C > > > attached base packages: > > [1] parallel stats graphics grDevices utils datasets methods > > [8] base > > > other attached packages: > > [1] hgu133plus2hsentrezgcdf_16.0.0 hgu133plus2frmavecs_1.3.0 > > [3] hgu133plus2cdf_2.12.0 AnnotationDbi_1.22.6 > > [5] frma_1.12.0 affy_1.38.1 > > [7] GEOquery_2.26.2 Biobase_2.20.1 > > [9] BiocGenerics_0.6.0 plyr_1.8.1 > > [11] rj_1.1.3-1 > > > loaded via a namespace (and not attached): > > [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 > > [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 > > [7] RCurl_1.95-4.1 RSQLite_0.11.4 Rcpp_0.11.2 > > [10] XML_3.95-0.2 affxparser_1.32.3 affyio_1.28.0 > > [13] bit_1.1-12 codetools_0.2-8 ff_2.2-13 > > [16] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 > > [19] oligoClasses_1.22.0 preprocessCore_1.22.0 rj.gd_1.1.3-1 > > [22] splines_3.0.3 stats4_3.0.3 tools_3.0.3 > > [25] zlibbioc_1.6.0 > > > > On Mon, Jul 28, 2014 at 1:44 PM, Matthew McCall <mccallm@gmail.com> wrote: > >> Yes, the intention is to be one version behind the latest Brainarray >> version. So hgu133plus2frmavecs_1.3.0 should really go with >> hgu133plus2hsentrezgcdf_17.0.0, but this year has been very hectic, so I >> missed an update. >> >> Best, >> Matt >> >> >> On Mon, Jul 28, 2014 at 3:55 AM, Tine Casneuf <tineke.casneuf@gmail.com> >> wrote: >> >>> Hi Matt, >>> >>> thanks a lot, this works indeed (hgu133plus2frmavecs_1.3.0 with >>> hgu133plus2hsentrezgcdf_16.0.0). Is your intention to keep updating >>> these to be able to work with the more recent versions of Brainarray? >>> >>> thanks again, >>> >>> tine >>> >>> >>> >>> On Sun, Jul 27, 2014 at 11:26 PM, Matthew McCall <mccallm@gmail.com> >>> wrote: >>> >>>> Tine, >>>> >>>> I'm a bit behind on updating the frmavecs alternative CDFs. Try using: >>>> hgu133plus2frmavecs_1.3.0 with hgu133plus2hsentrezgcdf_16.0.0 >>>> >>>> A couple other things to try: >>>> -- update to the latest version of preprocessCore (that's where >>>> rcModelWPLM comes from). >>>> -- not sure if this matters, but I always pass the cdfname to >>>> ReadAffy() instead of changing it afterwards >>>> >>>> Best, >>>> Matt >>>> >>>> >>>> >>>> >>>> On Sun, Jul 27, 2014 at 3:37 PM, Tine Casneuf [guest] < >>>> guest@bioconductor.org> wrote: >>>> >>>>> Hi all, >>>>> >>>>> I am trying to apply frma with an alternative Brainarray CDF to a >>>>> microarray dataset. I cannot get this to work and found a relevant post on >>>>> the bioC mailing list: >>>>> http://article.gmane.org/gmane.science.biology.informatics.condu ctor/47887/match=frma+alternative+cdf >>>>> >>>>> I get the same error as mentioned on the post: >>>>> >>>>> > test <- ReadAffy() >>>>> > test@cdfName <- "hgu133plus2hsentrezgcdf" >>>>> > library(frma) >>>>> > test2 <- frma(test) >>>>> Error in rcModelWPLM(y = x1, w = w.tmp, row.effects = pe.tmp, >>>>> input.scale = x4) : >>>>> row.effects should sum to zero >>>>> > test2 <- frma(test, summarize = "random_effect") >>>>> Error in if (convi < 1e-04) break : missing value where TRUE/FALSE >>>>> needed >>>>> >>>>> >>>>> I get the same error with the most recent version of the brain array >>>>> CDF (18.0.0). The suggested solution on the page mentioned above (brain >>>>> array version 15 and hgu133afrmavecs version 1.2.0) dord not solve the >>>>> problem for me. >>>>> >>>>> Does anyone know which versions of the brainarray CDFs, and >>>>> hgu133afrmavecs go together? >>>>> >>>>> thanks a lot, >>>>> >>>>> Tine >>>>> >>>>> -- output of sessionInfo(): >>>>> >>>>> > sessionInfo() >>>>> R version 3.0.3 (2014-03-06) >>>>> Platform: x86_64-apple-darwin10.8.0 (64-bit) >>>>> >>>>> locale: >>>>> [1] C/UTF-8/C/C/C/C >>>>> >>>>> attached base packages: >>>>> [1] parallel stats graphics grDevices utils datasets methods >>>>> [8] base >>>>> >>>>> other attached packages: >>>>> [1] hgu133plus2frmavecs_1.2.0 frma_1.12.0 >>>>> [3] hgu133plus2hsentrezgcdf_15.0.0 AnnotationDbi_1.22.6 >>>>> [5] affy_1.38.1 Biobase_2.20.1 >>>>> [7] BiocGenerics_0.6.0 >>>>> >>>>> loaded via a namespace (and not attached): >>>>> [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 >>>>> [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 >>>>> [7] RSQLite_0.11.4 affxparser_1.32.3 affyio_1.28.0 >>>>> [10] bit_1.1-12 codetools_0.2-8 ff_2.2-13 >>>>> [13] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 >>>>> [16] oligoClasses_1.22.0 preprocessCore_1.22.0 splines_3.0.3 >>>>> [19] stats4_3.0.3 tools_3.0.3 zlibbioc_1.6.0 >>>>> >>>>> >>>>> -- >>>>> Sent via the guest posting facility at bioconductor.org. >>>>> >>>> >>>> >>>> >>>> -- >>>> Matthew N McCall, PhD >>>> 112 Arvine Heights >>>> Rochester, NY 14611 >>>> Cell: 202-222-5880 >>>> >>>> >>> >> >> >> -- >> Matthew N McCall, PhD >> 112 Arvine Heights >> Rochester, NY 14611 >> Cell: 202-222-5880 >> >> > -- Matthew N McCall, PhD 112 Arvine Heights Rochester, NY 14611 Cell: 202-222-5880 [[alternative HTML version deleted]]

The barcode function uses precomputed distributions that are based on data preprocessed with the Affy CDF. Comparing values from data preprocessed with an alternative CDF to these distributions doesn't make any sense -- the probesets are fundamentally different. Matt On Mon, Jul 28, 2014 at 10:14 AM, Tine Casneuf <tineke.casneuf@gmail.com> wrote: > Hi Matt, > > that is correct, it indeed only works with the alternative CDF eset object > if , platform = "GPL570" is added. > > Do you know why the number of genes expressed differs to such a great > extent: > > 30.000 genes with Affy CDF versus 110 with the alternative (after adding , > platform = "GPL570") ? > > thanks, > > tine > > > > > On Mon, Jul 28, 2014 at 4:09 PM, Matthew McCall <mccallm@gmail.com> wrote: > >> Happy to help. The Gene Expression Barcode (barcode function in the frma >> package) is only implemented for the Affy CDF. It should produce an error >> when you try to barcode the output from frma using an alternative CDF, but >> I believe you overrode the error by supplying the argument -- platform = >> "GPL570". >> >> Best, >> Matt >> >> >> >> On Mon, Jul 28, 2014 at 9:58 AM, Tine Casneuf <tineke.casneuf@gmail.com> >> wrote: >> >>> Hi Matt, >>> >>> thanks in advance. Not a problem, it works with the previous version. >>> >>> Can i bother you with another question? With an example dataset, I tried >>> to estimate which genes are expressed, but found an substantial difference >>> between using the Affy CDF versus the alternative one: >>> >>> library(frma) >>> >>> library(affy) >>> >>> library(GEOquery) >>> >>> getGEOSuppFiles("GSE20986") >>> >>> dir("GSE20986/") >>> >>> untar("GSE20986/GSE20986_RAW.tar", exdir="data") >>> >>> cels <- list.files("data/", pattern = "[gz]") >>> >>> sapply(paste("data", cels, sep="/"), gunzip) >>> >>> myCELs <- paste("data/", dir("data/", pattern = ".CEL$"), sep = "") >>> >>> >>> ## affy CDF >>> >>> batch <- ReadAffy(filenames = myCELs) >>> >>> eset <- frma(batch, summarize = "random_effect") >>> >>> apc <- barcode(eset, cutoff = (-log10(0.05)), output = "binary") >>> >>> selPbs <- sum(rowSums(apc) > 1) >>> >>> selPbs >>> >>> >>> ## alternative CDF >>> >>> batchBA <- ReadAffy(filenames = myCELs, cdfname = "hgu133plus2hsentrezg" >>> ) >>> >>> esetBA <- frma(batchBA, summarize = "random_effect") >>> >>> apcBA <- barcode(esetBA, cutoff = (-log10(0.05)), output = "binary", >>> platform = "GPL570") >>> >>> selPbsBA <- sum(rowSums(apcBA) > 1) >>> >>> selPbsBA >>> >>> > selPbs ## the Affy CDF >>> >>> [1] 30117 >>> > selPbsBA ## the brain array CDF >>> >>> [1] 117 >>> >>> Would you happen to know why with the Affy CDF the number of expressed >>> genes is ~30.000 and with the alternative only ~100 genes, 300 times lower? >>> >>> Thanks a lot, >>> Best, >>> >>> Tine >>> >>> >>> As before, > sessionInfo() >>> >>> R version 3.0.3 (2014-03-06) >>> >>> Platform: x86_64-apple-darwin10.8.0 (64-bit) >>> >>> >>> locale: >>> >>> [1] C >>> >>> >>> attached base packages: >>> >>> [1] parallel stats graphics grDevices utils datasets methods >>> >>> [8] base >>> >>> >>> other attached packages: >>> >>> [1] hgu133plus2hsentrezgcdf_16.0.0 hgu133plus2frmavecs_1.3.0 >>> >>> [3] hgu133plus2cdf_2.12.0 AnnotationDbi_1.22.6 >>> >>> [5] frma_1.12.0 affy_1.38.1 >>> >>> [7] GEOquery_2.26.2 Biobase_2.20.1 >>> >>> [9] BiocGenerics_0.6.0 plyr_1.8.1 >>> >>> [11] rj_1.1.3-1 >>> >>> >>> loaded via a namespace (and not attached): >>> >>> [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 >>> >>> [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 >>> >>> [7] RCurl_1.95-4.1 RSQLite_0.11.4 Rcpp_0.11.2 >>> >>> [10] XML_3.95-0.2 affxparser_1.32.3 affyio_1.28.0 >>> >>> [13] bit_1.1-12 codetools_0.2-8 ff_2.2-13 >>> >>> [16] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 >>> >>> [19] oligoClasses_1.22.0 preprocessCore_1.22.0 rj.gd_1.1.3-1 >>> >>> [22] splines_3.0.3 stats4_3.0.3 tools_3.0.3 >>> >>> [25] zlibbioc_1.6.0 >>> >>> >>> >>> On Mon, Jul 28, 2014 at 1:44 PM, Matthew McCall <mccallm@gmail.com> >>> wrote: >>> >>>> Yes, the intention is to be one version behind the latest Brainarray >>>> version. So hgu133plus2frmavecs_1.3.0 should really go with >>>> hgu133plus2hsentrezgcdf_17.0.0, but this year has been very hectic, so I >>>> missed an update. >>>> >>>> Best, >>>> Matt >>>> >>>> >>>> On Mon, Jul 28, 2014 at 3:55 AM, Tine Casneuf <tineke.casneuf@gmail.com>>>> > wrote: >>>> >>>>> Hi Matt, >>>>> >>>>> thanks a lot, this works indeed (hgu133plus2frmavecs_1.3.0 with >>>>> hgu133plus2hsentrezgcdf_16.0.0). Is your intention to keep updating >>>>> these to be able to work with the more recent versions of Brainarray? >>>>> >>>>> thanks again, >>>>> >>>>> tine >>>>> >>>>> >>>>> >>>>> On Sun, Jul 27, 2014 at 11:26 PM, Matthew McCall <mccallm@gmail.com> >>>>> wrote: >>>>> >>>>>> Tine, >>>>>> >>>>>> I'm a bit behind on updating the frmavecs alternative CDFs. Try using: >>>>>> hgu133plus2frmavecs_1.3.0 with hgu133plus2hsentrezgcdf_16.0.0 >>>>>> >>>>>> A couple other things to try: >>>>>> -- update to the latest version of preprocessCore (that's where >>>>>> rcModelWPLM comes from). >>>>>> -- not sure if this matters, but I always pass the cdfname to >>>>>> ReadAffy() instead of changing it afterwards >>>>>> >>>>>> Best, >>>>>> Matt >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On Sun, Jul 27, 2014 at 3:37 PM, Tine Casneuf [guest] < >>>>>> guest@bioconductor.org> wrote: >>>>>> >>>>>>> Hi all, >>>>>>> >>>>>>> I am trying to apply frma with an alternative Brainarray CDF to a >>>>>>> microarray dataset. I cannot get this to work and found a relevant post on >>>>>>> the bioC mailing list: >>>>>>> http://article.gmane.org/gmane.science.biology.informatics.con ductor/47887/match=frma+alternative+cdf >>>>>>> >>>>>>> I get the same error as mentioned on the post: >>>>>>> >>>>>>> > test <- ReadAffy() >>>>>>> > test@cdfName <- "hgu133plus2hsentrezgcdf" >>>>>>> > library(frma) >>>>>>> > test2 <- frma(test) >>>>>>> Error in rcModelWPLM(y = x1, w = w.tmp, row.effects = pe.tmp, >>>>>>> input.scale = x4) : >>>>>>> row.effects should sum to zero >>>>>>> > test2 <- frma(test, summarize = "random_effect") >>>>>>> Error in if (convi < 1e-04) break : missing value where TRUE/FALSE >>>>>>> needed >>>>>>> >>>>>>> >>>>>>> I get the same error with the most recent version of the brain array >>>>>>> CDF (18.0.0). The suggested solution on the page mentioned above (brain >>>>>>> array version 15 and hgu133afrmavecs version 1.2.0) dord not solve the >>>>>>> problem for me. >>>>>>> >>>>>>> Does anyone know which versions of the brainarray CDFs, and >>>>>>> hgu133afrmavecs go together? >>>>>>> >>>>>>> thanks a lot, >>>>>>> >>>>>>> Tine >>>>>>> >>>>>>> -- output of sessionInfo(): >>>>>>> >>>>>>> > sessionInfo() >>>>>>> R version 3.0.3 (2014-03-06) >>>>>>> Platform: x86_64-apple-darwin10.8.0 (64-bit) >>>>>>> >>>>>>> locale: >>>>>>> [1] C/UTF-8/C/C/C/C >>>>>>> >>>>>>> attached base packages: >>>>>>> [1] parallel stats graphics grDevices utils datasets >>>>>>> methods >>>>>>> [8] base >>>>>>> >>>>>>> other attached packages: >>>>>>> [1] hgu133plus2frmavecs_1.2.0 frma_1.12.0 >>>>>>> [3] hgu133plus2hsentrezgcdf_15.0.0 AnnotationDbi_1.22.6 >>>>>>> [5] affy_1.38.1 Biobase_2.20.1 >>>>>>> [7] BiocGenerics_0.6.0 >>>>>>> >>>>>>> loaded via a namespace (and not attached): >>>>>>> [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 >>>>>>> [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 >>>>>>> [7] RSQLite_0.11.4 affxparser_1.32.3 affyio_1.28.0 >>>>>>> [10] bit_1.1-12 codetools_0.2-8 ff_2.2-13 >>>>>>> [13] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 >>>>>>> [16] oligoClasses_1.22.0 preprocessCore_1.22.0 splines_3.0.3 >>>>>>> [19] stats4_3.0.3 tools_3.0.3 zlibbioc_1.6.0 >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> Sent via the guest posting facility at bioconductor.org. >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Matthew N McCall, PhD >>>>>> 112 Arvine Heights >>>>>> Rochester, NY 14611 >>>>>> Cell: 202-222-5880 >>>>>> >>>>>> >>>>> >>>> >>>> >>>> -- >>>> Matthew N McCall, PhD >>>> 112 Arvine Heights >>>> Rochester, NY 14611 >>>> Cell: 202-222-5880 >>>> >>>> >>> >> >> >> -- >> Matthew N McCall, PhD >> 112 Arvine Heights >> Rochester, NY 14611 >> Cell: 202-222-5880 >> >> > -- Matthew N McCall, PhD 112 Arvine Heights Rochester, NY 14611 Cell: 202-222-5880 [[alternative HTML version deleted]]

Tine, No plans for a BrainArray preprocessed barcode at the moment. If it is something you're very interested in, I'd be happy to chat about options. Best, Matt On Mon, Jul 28, 2014 at 10:30 AM, Tine Casneuf <tineke.casneuf@gmail.com> wrote: > Hi Matt, > > As I had expected :) > > Assuming this is not something I as a user can do, do you have plans to > develop the barcode for the brainarray CDF parsed data? -I assume this > would be quite an effort as it would require rerunning for every other > Brainarray version? - > > thank you! > tine > > > > On Mon, Jul 28, 2014 at 4:20 PM, Matthew McCall <mccallm@gmail.com> wrote: > >> The barcode function uses precomputed distributions that are based on >> data preprocessed with the Affy CDF. Comparing values from data >> preprocessed with an alternative CDF to these distributions doesn't make >> any sense -- the probesets are fundamentally different. >> >> Matt >> >> >> On Mon, Jul 28, 2014 at 10:14 AM, Tine Casneuf <tineke.casneuf@gmail.com> >> wrote: >> >>> Hi Matt, >>> >>> that is correct, it indeed only works with the alternative CDF eset >>> object if , platform = "GPL570" is added. >>> >>> Do you know why the number of genes expressed differs to such a great >>> extent: >>> >>> 30.000 genes with Affy CDF versus 110 with the alternative (after >>> adding , platform = "GPL570") ? >>> >>> thanks, >>> >>> tine >>> >>> >>> >>> >>> On Mon, Jul 28, 2014 at 4:09 PM, Matthew McCall <mccallm@gmail.com> >>> wrote: >>> >>>> Happy to help. The Gene Expression Barcode (barcode function in the >>>> frma package) is only implemented for the Affy CDF. It should produce an >>>> error when you try to barcode the output from frma using an alternative >>>> CDF, but I believe you overrode the error by supplying the argument -- >>>> platform = "GPL570". >>>> >>>> Best, >>>> Matt >>>> >>>> >>>> >>>> On Mon, Jul 28, 2014 at 9:58 AM, Tine Casneuf <tineke.casneuf@gmail.com>>>> > wrote: >>>> >>>>> Hi Matt, >>>>> >>>>> thanks in advance. Not a problem, it works with the previous version. >>>>> >>>>> Can i bother you with another question? With an example dataset, I >>>>> tried to estimate which genes are expressed, but found an substantial >>>>> difference between using the Affy CDF versus the alternative one: >>>>> >>>>> library(frma) >>>>> >>>>> library(affy) >>>>> >>>>> library(GEOquery) >>>>> >>>>> getGEOSuppFiles("GSE20986") >>>>> >>>>> dir("GSE20986/") >>>>> >>>>> untar("GSE20986/GSE20986_RAW.tar", exdir="data") >>>>> >>>>> cels <- list.files("data/", pattern = "[gz]") >>>>> >>>>> sapply(paste("data", cels, sep="/"), gunzip) >>>>> >>>>> myCELs <- paste("data/", dir("data/", pattern = ".CEL$"), sep = "") >>>>> >>>>> >>>>> ## affy CDF >>>>> >>>>> batch <- ReadAffy(filenames = myCELs) >>>>> >>>>> eset <- frma(batch, summarize = "random_effect") >>>>> >>>>> apc <- barcode(eset, cutoff = (-log10(0.05)), output = "binary") >>>>> >>>>> selPbs <- sum(rowSums(apc) > 1) >>>>> >>>>> selPbs >>>>> >>>>> >>>>> ## alternative CDF >>>>> >>>>> batchBA <- ReadAffy(filenames = myCELs, cdfname = >>>>> "hgu133plus2hsentrezg") >>>>> >>>>> esetBA <- frma(batchBA, summarize = "random_effect") >>>>> >>>>> apcBA <- barcode(esetBA, cutoff = (-log10(0.05)), output = "binary", >>>>> platform = "GPL570") >>>>> >>>>> selPbsBA <- sum(rowSums(apcBA) > 1) >>>>> >>>>> selPbsBA >>>>> >>>>> > selPbs ## the Affy CDF >>>>> >>>>> [1] 30117 >>>>> > selPbsBA ## the brain array CDF >>>>> >>>>> [1] 117 >>>>> >>>>> Would you happen to know why with the Affy CDF the number of expressed >>>>> genes is ~30.000 and with the alternative only ~100 genes, 300 times lower? >>>>> >>>>> Thanks a lot, >>>>> Best, >>>>> >>>>> Tine >>>>> >>>>> >>>>> As before, > sessionInfo() >>>>> >>>>> R version 3.0.3 (2014-03-06) >>>>> >>>>> Platform: x86_64-apple-darwin10.8.0 (64-bit) >>>>> >>>>> >>>>> locale: >>>>> >>>>> [1] C >>>>> >>>>> >>>>> attached base packages: >>>>> >>>>> [1] parallel stats graphics grDevices utils datasets >>>>> methods >>>>> >>>>> [8] base >>>>> >>>>> >>>>> other attached packages: >>>>> >>>>> [1] hgu133plus2hsentrezgcdf_16.0.0 hgu133plus2frmavecs_1.3.0 >>>>> >>>>> [3] hgu133plus2cdf_2.12.0 AnnotationDbi_1.22.6 >>>>> >>>>> [5] frma_1.12.0 affy_1.38.1 >>>>> >>>>> [7] GEOquery_2.26.2 Biobase_2.20.1 >>>>> >>>>> [9] BiocGenerics_0.6.0 plyr_1.8.1 >>>>> >>>>> [11] rj_1.1.3-1 >>>>> >>>>> >>>>> loaded via a namespace (and not attached): >>>>> >>>>> [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 >>>>> >>>>> [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 >>>>> >>>>> [7] RCurl_1.95-4.1 RSQLite_0.11.4 Rcpp_0.11.2 >>>>> >>>>> [10] XML_3.95-0.2 affxparser_1.32.3 affyio_1.28.0 >>>>> >>>>> [13] bit_1.1-12 codetools_0.2-8 ff_2.2-13 >>>>> >>>>> [16] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 >>>>> >>>>> [19] oligoClasses_1.22.0 preprocessCore_1.22.0 rj.gd_1.1.3-1 >>>>> >>>>> [22] splines_3.0.3 stats4_3.0.3 tools_3.0.3 >>>>> >>>>> [25] zlibbioc_1.6.0 >>>>> >>>>> >>>>> >>>>> On Mon, Jul 28, 2014 at 1:44 PM, Matthew McCall <mccallm@gmail.com> >>>>> wrote: >>>>> >>>>>> Yes, the intention is to be one version behind the latest Brainarray >>>>>> version. So hgu133plus2frmavecs_1.3.0 should really go with >>>>>> hgu133plus2hsentrezgcdf_17.0.0, but this year has been very hectic, so I >>>>>> missed an update. >>>>>> >>>>>> Best, >>>>>> Matt >>>>>> >>>>>> >>>>>> On Mon, Jul 28, 2014 at 3:55 AM, Tine Casneuf < >>>>>> tineke.casneuf@gmail.com> wrote: >>>>>> >>>>>>> Hi Matt, >>>>>>> >>>>>>> thanks a lot, this works indeed (hgu133plus2frmavecs_1.3.0 with >>>>>>> hgu133plus2hsentrezgcdf_16.0.0). Is your intention to keep updating >>>>>>> these to be able to work with the more recent versions of Brainarray? >>>>>>> >>>>>>> thanks again, >>>>>>> >>>>>>> tine >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Sun, Jul 27, 2014 at 11:26 PM, Matthew McCall <mccallm@gmail.com> >>>>>>> wrote: >>>>>>> >>>>>>>> Tine, >>>>>>>> >>>>>>>> I'm a bit behind on updating the frmavecs alternative CDFs. Try >>>>>>>> using: >>>>>>>> hgu133plus2frmavecs_1.3.0 with hgu133plus2hsentrezgcdf_16.0.0 >>>>>>>> >>>>>>>> A couple other things to try: >>>>>>>> -- update to the latest version of preprocessCore (that's where >>>>>>>> rcModelWPLM comes from). >>>>>>>> -- not sure if this matters, but I always pass the cdfname to >>>>>>>> ReadAffy() instead of changing it afterwards >>>>>>>> >>>>>>>> Best, >>>>>>>> Matt >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Sun, Jul 27, 2014 at 3:37 PM, Tine Casneuf [guest] < >>>>>>>> guest@bioconductor.org> wrote: >>>>>>>> >>>>>>>>> Hi all, >>>>>>>>> >>>>>>>>> I am trying to apply frma with an alternative Brainarray CDF to a >>>>>>>>> microarray dataset. I cannot get this to work and found a relevant post on >>>>>>>>> the bioC mailing list: >>>>>>>>> http://article.gmane.org/gmane.science.biology.informatics.c onductor/47887/match=frma+alternative+cdf >>>>>>>>> >>>>>>>>> I get the same error as mentioned on the post: >>>>>>>>> >>>>>>>>> > test <- ReadAffy() >>>>>>>>> > test@cdfName <- "hgu133plus2hsentrezgcdf" >>>>>>>>> > library(frma) >>>>>>>>> > test2 <- frma(test) >>>>>>>>> Error in rcModelWPLM(y = x1, w = w.tmp, row.effects = pe.tmp, >>>>>>>>> input.scale = x4) : >>>>>>>>> row.effects should sum to zero >>>>>>>>> > test2 <- frma(test, summarize = "random_effect") >>>>>>>>> Error in if (convi < 1e-04) break : missing value where TRUE/FALSE >>>>>>>>> needed >>>>>>>>> >>>>>>>>> >>>>>>>>> I get the same error with the most recent version of the brain >>>>>>>>> array CDF (18.0.0). The suggested solution on the page mentioned above >>>>>>>>> (brain array version 15 and hgu133afrmavecs version 1.2.0) dord not solve >>>>>>>>> the problem for me. >>>>>>>>> >>>>>>>>> Does anyone know which versions of the brainarray CDFs, and >>>>>>>>> hgu133afrmavecs go together? >>>>>>>>> >>>>>>>>> thanks a lot, >>>>>>>>> >>>>>>>>> Tine >>>>>>>>> >>>>>>>>> -- output of sessionInfo(): >>>>>>>>> >>>>>>>>> > sessionInfo() >>>>>>>>> R version 3.0.3 (2014-03-06) >>>>>>>>> Platform: x86_64-apple-darwin10.8.0 (64-bit) >>>>>>>>> >>>>>>>>> locale: >>>>>>>>> [1] C/UTF-8/C/C/C/C >>>>>>>>> >>>>>>>>> attached base packages: >>>>>>>>> [1] parallel stats graphics grDevices utils datasets >>>>>>>>> methods >>>>>>>>> [8] base >>>>>>>>> >>>>>>>>> other attached packages: >>>>>>>>> [1] hgu133plus2frmavecs_1.2.0 frma_1.12.0 >>>>>>>>> [3] hgu133plus2hsentrezgcdf_15.0.0 AnnotationDbi_1.22.6 >>>>>>>>> [5] affy_1.38.1 Biobase_2.20.1 >>>>>>>>> [7] BiocGenerics_0.6.0 >>>>>>>>> >>>>>>>>> loaded via a namespace (and not attached): >>>>>>>>> [1] BiocInstaller_1.12.1 Biostrings_2.28.0 DBI_0.2-7 >>>>>>>>> [4] GenomicRanges_1.12.5 IRanges_1.18.4 MASS_7.3-33 >>>>>>>>> [7] RSQLite_0.11.4 affxparser_1.32.3 affyio_1.28.0 >>>>>>>>> [10] bit_1.1-12 codetools_0.2-8 ff_2.2-13 >>>>>>>>> [13] foreach_1.4.2 iterators_1.0.7 oligo_1.24.2 >>>>>>>>> [16] oligoClasses_1.22.0 preprocessCore_1.22.0 splines_3.0.3 >>>>>>>>> [19] stats4_3.0.3 tools_3.0.3 zlibbioc_1.6.0 >>>>>>>>> >>>>>>>>> >>>>>>>>> -- >>>>>>>>> Sent via the guest posting facility at bioconductor.org. >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> Matthew N McCall, PhD >>>>>>>> 112 Arvine Heights >>>>>>>> Rochester, NY 14611 >>>>>>>> Cell: 202-222-5880 >>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Matthew N McCall, PhD >>>>>> 112 Arvine Heights >>>>>> Rochester, NY 14611 >>>>>> Cell: 202-222-5880 >>>>>> >>>>>> >>>>> >>>> >>>> >>>> -- >>>> Matthew N McCall, PhD >>>> 112 Arvine Heights >>>> Rochester, NY 14611 >>>> Cell: 202-222-5880 >>>> >>>> >>> >> >> >> -- >> Matthew N McCall, PhD >> 112 Arvine Heights >> Rochester, NY 14611 >> Cell: 202-222-5880 >> >> > -- Matthew N McCall, PhD 112 Arvine Heights Rochester, NY 14611 Cell: 202-222-5880 [[alternative HTML version deleted]]