On 12/02/2014 11:34, "Gordon Brown" <gordon.brown at="" cruk.cam.ac.uk=""> wrote: >Hi, Sharona, > >Part of the problem is that some of your peaks have width zero (or 1, >depending on the exact definition of "narrow" format). For example, line >12826 of synch_PolII_rep1_narrowPeakn.bed is > >--------------- >chr1 84765176 84765176 . 9.263 . 9.263 1.709270e+00 2.493753e-02 0 >--------------- > >When DiffBind tries to normalise the score based on the width of the peak, >it comes out as N/0, which R shows as infinity. Things go downhill from >there. This is a bug in DiffBind (if the width is zero, we should just >make it 1, for example). But in the meantime, to work around it, just >remove those peaks (or add 1 to the right coordinate). (In bed format, >intervals are represented as half-open, i.e. the right end is one past the >end of the interval; I'll have to check the exact definition of "narrow" >format to see if we should be interpreting the coordinates differently.) > >Regarding memory usage, the latest version of DiffBind is *much* more >memory efficient than previous ones, so upgrading should take care of >that. > >Cheers, > > - Gord > > >On 2014-02-11 18:10, "Sharona Elgavish" <sharonae at="" ekmd.huji.ac.il=""> wrote: > >>Dear all, >> >>I'm trying to upload my Chip-Seq data to DiffBind and receive the >>following error: >> >>> pol = dba(sampleSheet="pol_narrow_bam_allpeaks.csv", >>>peakFormat="narrow") >> >>polsyn1 hella pol synchronized 1 narrow >>polsyn2 hella pol synchronized 2 narrow >>polunsyn1 hella pol unsynchronized 1 narrow >>polunsyn2 hella pol unsynchronized 2 narrow >>Error in if (res >= minval) { : missing value where TRUE/FALSE needed >> >>> sessionInfo() >>R version 3.0.1 (2013-05-16) >>Platform: x86_64-pc-linux-gnu (64-bit) >> >>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=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 >> [7] LC_PAPER=C LC_NAME=C >> [9] LC_ADDRESS=C LC_TELEPHONE=C >>[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C >> >>attached base packages: >>[1] parallel stats graphics grDevices utils datasets methods >>[8] base >> >>other attached packages: >>[1] DiffBind_1.6.2 Biobase_2.20.1 GenomicRanges_1.12.5 >>[4] IRanges_1.18.4 BiocGenerics_0.6.0 BiocInstaller_1.10.4 >> >>loaded via a namespace (and not attached): >> [1] amap_0.8-7 bitops_1.0-6 caTools_1.16 edgeR_3.2.4 >> [5] gdata_2.13.2 gplots_2.12.1 gtools_3.2.1 >>KernSmooth_2.23-10 >> [9] limma_3.16.8 RColorBrewer_1.0-5 stats4_3.0.1 tools_3.0.1 >> >>[13] zlibbioc_1.6.0 >> >> >> >>I get that error message whether I install DiffBind through the >>Bioconductor that is, biocLite("DiffBind"), or by installing >>DiffBind_1.8.3 by >>install.packages(DiffBind_1.8.3.tar.gz") >> >>When I took only the first 4000 peaks in each of the peak files (I have >>4) I easily manged to upload the dba. However, running dba.count was >>successful only upon using the bUseSummarizeOverlaps=TRUE, >>bParallel=FALSE options >> >>The peak files are in narrowPeaks format and were generated using the >>program PeakSeq. >> >>The peak files contain 10,000-32,000 peaks >> >>Am I out of memory? In case I am, how can I increase the memory during >>the upload (and not at dba.count)? >> >>I attached the sampleSheet and the 4 peaks files. >> >> >>I would much appreciate any help >>Thanks! >> >>Sharona >> >