Sonika and Alok Just to confirm: zeros will not cause the the problem that you have reported (tested on dozens of datasets with zero counts). Like Paul, I suspect that you have some NAs in your count matrix. This is unusual. I haven't seen RNA-Seq results with NAs before. I suggest you follow Paul's suggestion. If you find NAs then you can make a decision about removing the tag or setting NAs to zero. If you don't find NAs then we can dig deeper. As an aside I also note that you are using an older version of R and edgeR. I strongly recommend updating to R 2.13 and the corresponding version of edgeR using biocLite(), which will give you edgeR 2.2.5. We have done a lot of development and improvement of the package in the last year. Best wishes Davis > To:?"'bioconductor at r-project.org'" <bioconductor at="" r-project.org=""> > Date:?Wed, 31 Aug 2011 10:02:26 +1000 > Subject:?[BioC] edgeR: handling missing values with Quantile normalisation > Hi there, > > I am analysing RNAseq counts using edgeR package. But I am running into problems because of 'zero' counts for certain tags in my data. > > The code syntax I am using is here: > >> targets <- read.delim(file = "Targets.txt", stringsAsFactors = FALSE) >> targets > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?files ? group description > 1 ?Sample_xx_count.txt.raw control ? something > 2 ?Sample_xx_count.txt.raw control ? something > 3 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 4 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 5 ?Sample_xx_count.txt.raw control ? something > 6 ?Sample_xx_count.txt.raw control ? something > 7 ? ................ > > d <- readDGE(targets, skip = 0, comment.char = "#") > d > > An object of class "DGEList" > $samples > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? files ? group description ?lib.size norm.factors > 1 Sample_xx_count.txt.raw control ? something 498180513 ? ? ? ? ? ?1 > 2 Sample_xx_count.txt.raw control ? something 483775405 ? ? ? ? ? ?1 > 3 Sample_xx_count.txt.raw ?Hi_Pos ? something 368609647 ? ? ? ? ? ?1 > 4 Sample_xx_count.txt.raw ?Hi_Pos ? something 617334315 ? ? ? ? ? ?1 > 5 Sample_xx_count.txt.raw control ? something 678060765 ? ? ? ? ? ?1 > 13 more rows ... > > $counts > ? ? ? ? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? ?7 ? ? 8 ? ? 9 ? ?10 ? ? 11 ? ?12 ? ?13 ? ?14 ? ? 15 16 ? ?17 ? ?18 > Tag1 ? 15923 20323 14867 23098 32484 17223 ?51579 29578 17408 24097 ?34470 31964 17583 17583 ?39460 ?0 30359 25416 > Tag2 ? ? ? ?700 ? 600 ? 200 ? 695 ? 500 ?1300 ? 1425 ?1775 ? 700 ?1974 ? 1300 ?2371 ? 900 ? 900 ? 1689 ?0 ? 898 ?1690 > Tag3 ? ? ?0 ? ? 0 ? 100 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ?100 ?0 ? 100 ? ? 0 > Tag4 ? ? 74008 58753 51648 65233 93828 71047 117340 90551 55000 70124 121393 86106 46197 46197 127290 ?0 98369 79673 > Tag5 ? ? 19868 19385 25500 31215 56684 24096 ?51265 37492 27420 24496 ?32729 24722 24913 24913 ?50448 ?0 39755 55829 > 21887 more rows ... > > > ?d <- calcNormFactors(d) > Error in quantile.default(x, p = q) : > ?missing values and NaN's not allowed if 'na.rm' is FALSE > > Could someone please suggest how to handle the missing values with edgeR normalisation methods ? > > Thank you > Sonika > ------------------- > >> sessionInfo() > R version 2.12.2 (2011-02-25) > Platform: i386-pc-mingw32/i386 (32-bit) > > locale: > [1] LC_COLLATE=English_Australia.1252 ?LC_CTYPE=English_Australia.1252 ? ?LC_MONETARY=English_Australia.1252 > [4] LC_NUMERIC=C ? ? ? ? ? ? ? ? ? ? ? LC_TIME=English_Australia.1252 > > attached base packages: > [1] stats ? ? graphics ?grDevices utils ? ? datasets ?methods ? base > > other attached packages: > [1] edgeR_2.0.5 ?svIDE_0.9-50 > > loaded via a namespace (and not attached): > [1] limma_3.6.9 ? svMisc_0.9-61 tcltk_2.12.2 ?tools_2.12.2 ?XML_3.2-0.2 > > ? ? ? ?[[alternative HTML version deleted]] > > > > > ---------- Forwarded message ---------- > From:?Paul Leo <p.leo at="" uq.edu.au=""> > To:?Sonika Tyagi <sonika.tyagi at="" agrf.org.au=""> > Date:?Wed, 31 Aug 2011 11:07:47 +1000 > Subject:?Re: [BioC] edgeR: handling missing values with Quantile normalisation > > HI Sonika > It is probably not zero's that are causing the problem but NAs, > > Check through the counts array > to see if it contains ?NA's ... someting like.. > > apply(d$counts,2,function(x) sumis.na(x))) > > should get back all zeros.... > > probably setting them to 0 is appropriate. > > > Cheers > Paul > > > > -----Original Message----- > From: Sonika Tyagi <sonika.tyagi at="" agrf.org.au=""> > To: 'bioconductor at r-project.org' <bioconductor at="" r-project.org=""> > Subject: [BioC] edgeR: handling missing values with Quantile > normalisation > Date: Wed, 31 Aug 2011 10:02:26 +1000 > > Hi there, > > I am analysing RNAseq counts using edgeR package. But I am running into problems because of 'zero' counts for certain tags in my data. > > The code syntax I am using is here: > >> targets <- read.delim(file = "Targets.txt", stringsAsFactors = FALSE) >> targets > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?files ? group description > 1 ?Sample_xx_count.txt.raw control ? something > 2 ?Sample_xx_count.txt.raw control ? something > 3 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 4 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 5 ?Sample_xx_count.txt.raw control ? something > 6 ?Sample_xx_count.txt.raw control ? something > 7 ? ................ > > d <- readDGE(targets, skip = 0, comment.char = "#") > d > > An object of class "DGEList" > $samples > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? files ? group description ?lib.size norm.factors > 1 Sample_xx_count.txt.raw control ? something 498180513 ? ? ? ? ? ?1 > 2 Sample_xx_count.txt.raw control ? something 483775405 ? ? ? ? ? ?1 > 3 Sample_xx_count.txt.raw ?Hi_Pos ? something 368609647 ? ? ? ? ? ?1 > 4 Sample_xx_count.txt.raw ?Hi_Pos ? something 617334315 ? ? ? ? ? ?1 > 5 Sample_xx_count.txt.raw control ? something 678060765 ? ? ? ? ? ?1 > 13 more rows ... > > $counts > ? ? ? ? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? ?7 ? ? 8 ? ? 9 ? ?10 ? ? 11 ? ?12 ? ?13 ? ?14 ? ? 15 16 ? ?17 ? ?18 > Tag1 ? 15923 20323 14867 23098 32484 17223 ?51579 29578 17408 24097 ?34470 31964 17583 17583 ?39460 ?0 30359 25416 > Tag2 ? ? ? ?700 ? 600 ? 200 ? 695 ? 500 ?1300 ? 1425 ?1775 ? 700 ?1974 ? 1300 ?2371 ? 900 ? 900 ? 1689 ?0 ? 898 ?1690 > Tag3 ? ? ?0 ? ? 0 ? 100 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ?100 ?0 ? 100 ? ? 0 > Tag4 ? ? 74008 58753 51648 65233 93828 71047 117340 90551 55000 70124 121393 86106 46197 46197 127290 ?0 98369 79673 > Tag5 ? ? 19868 19385 25500 31215 56684 24096 ?51265 37492 27420 24496 ?32729 24722 24913 24913 ?50448 ?0 39755 55829 > 21887 more rows ... > > > ?d <- calcNormFactors(d) > Error in quantile.default(x, p = q) : > ?missing values and NaN's not allowed if 'na.rm' is FALSE > > Could someone please suggest how to handle the missing values with edgeR normalisation methods ? > > Thank you > Sonika > ------------------- > >> sessionInfo() > R version 2.12.2 (2011-02-25) > Platform: i386-pc-mingw32/i386 (32-bit) > > locale: > [1] LC_COLLATE=English_Australia.1252 ?LC_CTYPE=English_Australia.1252 ? ?LC_MONETARY=English_Australia.1252 > [4] LC_NUMERIC=C ? ? ? ? ? ? ? ? ? ? ? LC_TIME=English_Australia.1252 > > attached base packages: > [1] stats ? ? graphics ?grDevices utils ? ? datasets ?methods ? base > > other attached packages: > [1] edgeR_2.0.5 ?svIDE_0.9-50 > > loaded via a namespace (and not attached): > [1] limma_3.6.9 ? svMisc_0.9-61 tcltk_2.12.2 ?tools_2.12.2 ?XML_3.2-0.2 > > ? ? ? ?[[alternative HTML version deleted]] > > _______________________________________________ > 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 > > > > > ---------- Forwarded message ---------- > From:?ALok <foralok at="" gmail.com=""> > To:?Paul Leo <p.leo at="" uq.edu.au=""> > Date:?Wed, 31 Aug 2011 10:47:22 +0530 > Subject:?Re: [BioC] edgeR: handling missing values with Quantile normalisation > Hi Sonika > > You can calculate quantile.default independently with the argument > quantile((x, p = q, ?na.rm = TRUE) > and pass this value to the main function > this will automatically take care of zeros. > > or alternatively you can try other methods ("TMM", "RLE", "quantile") for > calcNormFactors, if that fits in your requirements. > > cheers > Alok > > > On Wed, Aug 31, 2011 at 6:37 AM, Paul Leo <p.leo at="" uq.edu.au=""> wrote: > >> >> HI Sonika >> It is probably not zero's that are causing the problem but NAs, >> >> Check through the counts array >> to see if it contains ?NA's ... someting like.. >> >> apply(d$counts,2,function(x) sumis.na(x))) >> >> should get back all zeros.... >> >> probably setting them to 0 is appropriate. >> >> >> Cheers >> Paul >> >> >> >> -----Original Message----- >> From: Sonika Tyagi <sonika.tyagi at="" agrf.org.au=""> >> To: 'bioconductor at r-project.org' <bioconductor at="" r-project.org=""> >> Subject: [BioC] edgeR: handling missing values with Quantile >> normalisation >> Date: Wed, 31 Aug 2011 10:02:26 +1000 >> >> Hi there, >> >> I am analysing RNAseq counts using edgeR package. But I am running into >> problems because of 'zero' counts for certain tags in my data. >> >> The code syntax I am using is here: >> >> > targets <- read.delim(file = "Targets.txt", stringsAsFactors = FALSE) >> > targets >> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?files ? group description >> 1 ?Sample_xx_count.txt.raw control ? something >> 2 ?Sample_xx_count.txt.raw control ? something >> 3 ?Sample_xx_count.txt.raw ?Hi_Pos ? something >> 4 ?Sample_xx_count.txt.raw ?Hi_Pos ? something >> 5 ?Sample_xx_count.txt.raw control ? something >> 6 ?Sample_xx_count.txt.raw control ? something >> 7 ? ................ >> >> d <- readDGE(targets, skip = 0, comment.char = "#") >> d >> >> An object of class "DGEList" >> $samples >> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? files ? group description ?lib.size >> norm.factors >> 1 Sample_xx_count.txt.raw control ? something 498180513 ? ? ? ? ? ?1 >> 2 Sample_xx_count.txt.raw control ? something 483775405 ? ? ? ? ? ?1 >> 3 Sample_xx_count.txt.raw ?Hi_Pos ? something 368609647 ? ? ? ? ? ?1 >> 4 Sample_xx_count.txt.raw ?Hi_Pos ? something 617334315 ? ? ? ? ? ?1 >> 5 Sample_xx_count.txt.raw control ? something 678060765 ? ? ? ? ? ?1 >> 13 more rows ... >> >> $counts >> ? ? ? ? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? ?7 ? ? 8 ? ? 9 >> ?10 ? ? 11 ? ?12 ? ?13 ? ?14 ? ? 15 16 ? ?17 ? ?18 >> Tag1 ? 15923 20323 14867 23098 32484 17223 ?51579 29578 17408 24097 ?34470 >> 31964 17583 17583 ?39460 ?0 30359 25416 >> Tag2 ? ? ? ?700 ? 600 ? 200 ? 695 ? 500 ?1300 ? 1425 ?1775 ? 700 ?1974 >> 1300 ?2371 ? 900 ? 900 ? 1689 ?0 ? 898 ?1690 >> Tag3 ? ? ?0 ? ? 0 ? 100 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 >> ? 0 ? ? 0 ? ? 0 ? ?100 ?0 ? 100 ? ? 0 >> Tag4 ? ? 74008 58753 51648 65233 93828 71047 117340 90551 55000 70124 >> 121393 86106 46197 46197 127290 ?0 98369 79673 >> Tag5 ? ? 19868 19385 25500 31215 56684 24096 ?51265 37492 27420 24496 >> ?32729 24722 24913 24913 ?50448 ?0 39755 55829 >> 21887 more rows ... >> >> >> ?d <- calcNormFactors(d) >> Error in quantile.default(x, p = q) : >> ?missing values and NaN's not allowed if 'na.rm' is FALSE >> >> Could someone please suggest how to handle the missing values with edgeR >> normalisation methods ? >> >> Thank you >> Sonika >> ------------------- >> >> > sessionInfo() >> R version 2.12.2 (2011-02-25) >> Platform: i386-pc-mingw32/i386 (32-bit) >> >> locale: >> [1] LC_COLLATE=English_Australia.1252 ?LC_CTYPE=English_Australia.1252 >> ?LC_MONETARY=English_Australia.1252 >> [4] LC_NUMERIC=C ? ? ? ? ? ? ? ? ? ? ? LC_TIME=English_Australia.1252 >> >> attached base packages: >> [1] stats ? ? graphics ?grDevices utils ? ? datasets ?methods ? base >> >> other attached packages: >> [1] edgeR_2.0.5 ?svIDE_0.9-50 >> >> loaded via a namespace (and not attached): >> [1] limma_3.6.9 ? svMisc_0.9-61 tcltk_2.12.2 ?tools_2.12.2 ?XML_3.2-0.2 >> >> ? ? ? ?[[alternative HTML version deleted]] >> >> _______________________________________________ >> 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 >> >> _______________________________________________ >> 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 >> > > > > -- > ************************************************************ > Alok Kumar Srivastava > Ph.D scholar > Centre of Computational Biology and Bioinformatics > School of Computational and Integrative Sciences > JNU, New Delhi > ************************************************************ > > ? ? ? ?[[alternative HTML version deleted]] > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > > -- ----------------------------------------------------------- Davis J McCarthy DPhil Candidate University of Oxford E: davis.mccarthy at balliol.ox.ac.uk W: sites.google.com/site/davismcc

Thanks Paul, Alok and Davis for your response! Yes, I had a look at the data again and I didn't find NAs in there. I am upgrading the R and edgeR versions and shall run the code again and see how it goes. Kind regards Sonika -----Original Message----- From: davismcc@googlemail.com [mailto:davismcc@googlemail.com] On Behalf Of Davis McCarthy Sent: Thursday, 1 September 2011 9:11 AM To: bioconductor at r-project.org; Sonika Tyagi; foralok at gmail.com Subject: Re: Bioconductor Digest, Vol 102, Issue 29 Sonika and Alok Just to confirm: zeros will not cause the the problem that you have reported (tested on dozens of datasets with zero counts). Like Paul, I suspect that you have some NAs in your count matrix. This is unusual. I haven't seen RNA-Seq results with NAs before. I suggest you follow Paul's suggestion. If you find NAs then you can make a decision about removing the tag or setting NAs to zero. If you don't find NAs then we can dig deeper. As an aside I also note that you are using an older version of R and edgeR. I strongly recommend updating to R 2.13 and the corresponding version of edgeR using biocLite(), which will give you edgeR 2.2.5. We have done a lot of development and improvement of the package in the last year. Best wishes Davis > To:?"'bioconductor at r-project.org'" <bioconductor at="" r-project.org=""> > Date:?Wed, 31 Aug 2011 10:02:26 +1000 > Subject:?[BioC] edgeR: handling missing values with Quantile normalisation > Hi there, > > I am analysing RNAseq counts using edgeR package. But I am running into problems because of 'zero' counts for certain tags in my data. > > The code syntax I am using is here: > >> targets <- read.delim(file = "Targets.txt", stringsAsFactors = FALSE) >> targets > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?files ? group description > 1 ?Sample_xx_count.txt.raw control ? something > 2 ?Sample_xx_count.txt.raw control ? something > 3 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 4 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 5 ?Sample_xx_count.txt.raw control ? something > 6 ?Sample_xx_count.txt.raw control ? something > 7 ? ................ > > d <- readDGE(targets, skip = 0, comment.char = "#") > d > > An object of class "DGEList" > $samples > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? files ? group description ?lib.size norm.factors > 1 Sample_xx_count.txt.raw control ? something 498180513 ? ? ? ? ? ?1 > 2 Sample_xx_count.txt.raw control ? something 483775405 ? ? ? ? ? ?1 > 3 Sample_xx_count.txt.raw ?Hi_Pos ? something 368609647 ? ? ? ? ? ?1 > 4 Sample_xx_count.txt.raw ?Hi_Pos ? something 617334315 ? ? ? ? ? ?1 > 5 Sample_xx_count.txt.raw control ? something 678060765 ? ? ? ? ? ?1 > 13 more rows ... > > $counts > ? ? ? ? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? ?7 ? ? 8 ? ? 9 ? ?10 ? ? 11 ? ?12 ? ?13 ? ?14 ? ? 15 16 ? ?17 ? ?18 > Tag1 ? 15923 20323 14867 23098 32484 17223 ?51579 29578 17408 24097 ?34470 31964 17583 17583 ?39460 ?0 30359 25416 > Tag2 ? ? ? ?700 ? 600 ? 200 ? 695 ? 500 ?1300 ? 1425 ?1775 ? 700 ?1974 ? 1300 ?2371 ? 900 ? 900 ? 1689 ?0 ? 898 ?1690 > Tag3 ? ? ?0 ? ? 0 ? 100 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ?100 ?0 ? 100 ? ? 0 > Tag4 ? ? 74008 58753 51648 65233 93828 71047 117340 90551 55000 70124 121393 86106 46197 46197 127290 ?0 98369 79673 > Tag5 ? ? 19868 19385 25500 31215 56684 24096 ?51265 37492 27420 24496 ?32729 24722 24913 24913 ?50448 ?0 39755 55829 > 21887 more rows ... > > > ?d <- calcNormFactors(d) > Error in quantile.default(x, p = q) : > ?missing values and NaN's not allowed if 'na.rm' is FALSE > > Could someone please suggest how to handle the missing values with edgeR normalisation methods ? > > Thank you > Sonika > ------------------- > >> sessionInfo() > R version 2.12.2 (2011-02-25) > Platform: i386-pc-mingw32/i386 (32-bit) > > locale: > [1] LC_COLLATE=English_Australia.1252 ?LC_CTYPE=English_Australia.1252 ? ?LC_MONETARY=English_Australia.1252 > [4] LC_NUMERIC=C ? ? ? ? ? ? ? ? ? ? ? LC_TIME=English_Australia.1252 > > attached base packages: > [1] stats ? ? graphics ?grDevices utils ? ? datasets ?methods ? base > > other attached packages: > [1] edgeR_2.0.5 ?svIDE_0.9-50 > > loaded via a namespace (and not attached): > [1] limma_3.6.9 ? svMisc_0.9-61 tcltk_2.12.2 ?tools_2.12.2 ?XML_3.2-0.2 > > ? ? ? ?[[alternative HTML version deleted]] > > > > > ---------- Forwarded message ---------- > From:?Paul Leo <p.leo at="" uq.edu.au=""> > To:?Sonika Tyagi <sonika.tyagi at="" agrf.org.au=""> > Date:?Wed, 31 Aug 2011 11:07:47 +1000 > Subject:?Re: [BioC] edgeR: handling missing values with Quantile normalisation > > HI Sonika > It is probably not zero's that are causing the problem but NAs, > > Check through the counts array > to see if it contains ?NA's ... someting like.. > > apply(d$counts,2,function(x) sumis.na(x))) > > should get back all zeros.... > > probably setting them to 0 is appropriate. > > > Cheers > Paul > > > > -----Original Message----- > From: Sonika Tyagi <sonika.tyagi at="" agrf.org.au=""> > To: 'bioconductor at r-project.org' <bioconductor at="" r-project.org=""> > Subject: [BioC] edgeR: handling missing values with Quantile > normalisation > Date: Wed, 31 Aug 2011 10:02:26 +1000 > > Hi there, > > I am analysing RNAseq counts using edgeR package. But I am running into problems because of 'zero' counts for certain tags in my data. > > The code syntax I am using is here: > >> targets <- read.delim(file = "Targets.txt", stringsAsFactors = FALSE) >> targets > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?files ? group description > 1 ?Sample_xx_count.txt.raw control ? something > 2 ?Sample_xx_count.txt.raw control ? something > 3 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 4 ?Sample_xx_count.txt.raw ?Hi_Pos ? something > 5 ?Sample_xx_count.txt.raw control ? something > 6 ?Sample_xx_count.txt.raw control ? something > 7 ? ................ > > d <- readDGE(targets, skip = 0, comment.char = "#") > d > > An object of class "DGEList" > $samples > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? files ? group description ?lib.size norm.factors > 1 Sample_xx_count.txt.raw control ? something 498180513 ? ? ? ? ? ?1 > 2 Sample_xx_count.txt.raw control ? something 483775405 ? ? ? ? ? ?1 > 3 Sample_xx_count.txt.raw ?Hi_Pos ? something 368609647 ? ? ? ? ? ?1 > 4 Sample_xx_count.txt.raw ?Hi_Pos ? something 617334315 ? ? ? ? ? ?1 > 5 Sample_xx_count.txt.raw control ? something 678060765 ? ? ? ? ? ?1 > 13 more rows ... > > $counts > ? ? ? ? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? ?7 ? ? 8 ? ? 9 ? ?10 ? ? 11 ? ?12 ? ?13 ? ?14 ? ? 15 16 ? ?17 ? ?18 > Tag1 ? 15923 20323 14867 23098 32484 17223 ?51579 29578 17408 24097 ?34470 31964 17583 17583 ?39460 ?0 30359 25416 > Tag2 ? ? ? ?700 ? 600 ? 200 ? 695 ? 500 ?1300 ? 1425 ?1775 ? 700 ?1974 ? 1300 ?2371 ? 900 ? 900 ? 1689 ?0 ? 898 ?1690 > Tag3 ? ? ?0 ? ? 0 ? 100 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ?100 ?0 ? 100 ? ? 0 > Tag4 ? ? 74008 58753 51648 65233 93828 71047 117340 90551 55000 70124 121393 86106 46197 46197 127290 ?0 98369 79673 > Tag5 ? ? 19868 19385 25500 31215 56684 24096 ?51265 37492 27420 24496 ?32729 24722 24913 24913 ?50448 ?0 39755 55829 > 21887 more rows ... > > > ?d <- calcNormFactors(d) > Error in quantile.default(x, p = q) : > ?missing values and NaN's not allowed if 'na.rm' is FALSE > > Could someone please suggest how to handle the missing values with edgeR normalisation methods ? > > Thank you > Sonika > ------------------- > >> sessionInfo() > R version 2.12.2 (2011-02-25) > Platform: i386-pc-mingw32/i386 (32-bit) > > locale: > [1] LC_COLLATE=English_Australia.1252 ?LC_CTYPE=English_Australia.1252 ? ?LC_MONETARY=English_Australia.1252 > [4] LC_NUMERIC=C ? ? ? ? ? ? ? ? ? ? ? LC_TIME=English_Australia.1252 > > attached base packages: > [1] stats ? ? graphics ?grDevices utils ? ? datasets ?methods ? base > > other attached packages: > [1] edgeR_2.0.5 ?svIDE_0.9-50 > > loaded via a namespace (and not attached): > [1] limma_3.6.9 ? svMisc_0.9-61 tcltk_2.12.2 ?tools_2.12.2 ?XML_3.2-0.2 > > ? ? ? ?[[alternative HTML version deleted]] > > _______________________________________________ > 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 > > > > > ---------- Forwarded message ---------- > From:?ALok <foralok at="" gmail.com=""> > To:?Paul Leo <p.leo at="" uq.edu.au=""> > Date:?Wed, 31 Aug 2011 10:47:22 +0530 > Subject:?Re: [BioC] edgeR: handling missing values with Quantile normalisation > Hi Sonika > > You can calculate quantile.default independently with the argument > quantile((x, p = q, ?na.rm = TRUE) > and pass this value to the main function > this will automatically take care of zeros. > > or alternatively you can try other methods ("TMM", "RLE", "quantile") for > calcNormFactors, if that fits in your requirements. > > cheers > Alok > > > On Wed, Aug 31, 2011 at 6:37 AM, Paul Leo <p.leo at="" uq.edu.au=""> wrote: > >> >> HI Sonika >> It is probably not zero's that are causing the problem but NAs, >> >> Check through the counts array >> to see if it contains ?NA's ... someting like.. >> >> apply(d$counts,2,function(x) sumis.na(x))) >> >> should get back all zeros.... >> >> probably setting them to 0 is appropriate. >> >> >> Cheers >> Paul >> >> >> >> -----Original Message----- >> From: Sonika Tyagi <sonika.tyagi at="" agrf.org.au=""> >> To: 'bioconductor at r-project.org' <bioconductor at="" r-project.org=""> >> Subject: [BioC] edgeR: handling missing values with Quantile >> normalisation >> Date: Wed, 31 Aug 2011 10:02:26 +1000 >> >> Hi there, >> >> I am analysing RNAseq counts using edgeR package. But I am running into >> problems because of 'zero' counts for certain tags in my data. >> >> The code syntax I am using is here: >> >> > targets <- read.delim(file = "Targets.txt", stringsAsFactors = FALSE) >> > targets >> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?files ? group description >> 1 ?Sample_xx_count.txt.raw control ? something >> 2 ?Sample_xx_count.txt.raw control ? something >> 3 ?Sample_xx_count.txt.raw ?Hi_Pos ? something >> 4 ?Sample_xx_count.txt.raw ?Hi_Pos ? something >> 5 ?Sample_xx_count.txt.raw control ? something >> 6 ?Sample_xx_count.txt.raw control ? something >> 7 ? ................ >> >> d <- readDGE(targets, skip = 0, comment.char = "#") >> d >> >> An object of class "DGEList" >> $samples >> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? files ? group description ?lib.size >> norm.factors >> 1 Sample_xx_count.txt.raw control ? something 498180513 ? ? ? ? ? ?1 >> 2 Sample_xx_count.txt.raw control ? something 483775405 ? ? ? ? ? ?1 >> 3 Sample_xx_count.txt.raw ?Hi_Pos ? something 368609647 ? ? ? ? ? ?1 >> 4 Sample_xx_count.txt.raw ?Hi_Pos ? something 617334315 ? ? ? ? ? ?1 >> 5 Sample_xx_count.txt.raw control ? something 678060765 ? ? ? ? ? ?1 >> 13 more rows ... >> >> $counts >> ? ? ? ? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? ?7 ? ? 8 ? ? 9 >> ?10 ? ? 11 ? ?12 ? ?13 ? ?14 ? ? 15 16 ? ?17 ? ?18 >> Tag1 ? 15923 20323 14867 23098 32484 17223 ?51579 29578 17408 24097 ?34470 >> 31964 17583 17583 ?39460 ?0 30359 25416 >> Tag2 ? ? ? ?700 ? 600 ? 200 ? 695 ? 500 ?1300 ? 1425 ?1775 ? 700 ?1974 >> 1300 ?2371 ? 900 ? 900 ? 1689 ?0 ? 898 ?1690 >> Tag3 ? ? ?0 ? ? 0 ? 100 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 ? ? 0 ? ? 0 ? ? 0 ? ? ?0 >> ? 0 ? ? 0 ? ? 0 ? ?100 ?0 ? 100 ? ? 0 >> Tag4 ? ? 74008 58753 51648 65233 93828 71047 117340 90551 55000 70124 >> 121393 86106 46197 46197 127290 ?0 98369 79673 >> Tag5 ? ? 19868 19385 25500 31215 56684 24096 ?51265 37492 27420 24496 >> ?32729 24722 24913 24913 ?50448 ?0 39755 55829 >> 21887 more rows ... >> >> >> ?d <- calcNormFactors(d) >> Error in quantile.default(x, p = q) : >> ?missing values and NaN's not allowed if 'na.rm' is FALSE >> >> Could someone please suggest how to handle the missing values with edgeR >> normalisation methods ? >> >> Thank you >> Sonika >> ------------------- >> >> > sessionInfo() >> R version 2.12.2 (2011-02-25) >> Platform: i386-pc-mingw32/i386 (32-bit) >> >> locale: >> [1] LC_COLLATE=English_Australia.1252 ?LC_CTYPE=English_Australia.1252 >> ?LC_MONETARY=English_Australia.1252 >> [4] LC_NUMERIC=C ? ? ? ? ? ? ? ? ? ? ? LC_TIME=English_Australia.1252 >> >> attached base packages: >> [1] stats ? ? graphics ?grDevices utils ? ? datasets ?methods ? base >> >> other attached packages: >> [1] edgeR_2.0.5 ?svIDE_0.9-50 >> >> loaded via a namespace (and not attached): >> [1] limma_3.6.9 ? svMisc_0.9-61 tcltk_2.12.2 ?tools_2.12.2 ?XML_3.2-0.2 >> >> ? ? ? ?[[alternative HTML version deleted]] >> >> _______________________________________________ >> 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 >> >> _______________________________________________ >> 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 >> > > > > -- > ************************************************************ > Alok Kumar Srivastava > Ph.D scholar > Centre of Computational Biology and Bioinformatics > School of Computational and Integrative Sciences > JNU, New Delhi > ************************************************************ > > ? ? ? ?[[alternative HTML version deleted]] > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > > -- ----------------------------------------------------------- Davis J McCarthy DPhil Candidate University of Oxford E: davis.mccarthy at balliol.ox.ac.uk W: sites.google.com/site/davismcc