error in cummeRbund dispersionPlot

[frymor]

Hi,

I am trying to run cummeRbund with the cuffdiff results.
these is my workflow:

library(cummeRbund)

setwd("cummeRbund/")

cuff <- readCufflinks(dir=cuffFile_dir, gtfFile="merged.gtf", genome="dm3" , rebuild=TRUE)

everything seems to run good. all tables are ok, but I am getting these warnings:

Warning messages:

1: attributes are not identical across measure variables; they will be dropped 
...
8: attributes are not identical across measure variables; they will be dropped

I have read here, that these messages can be ignored, so I am not sure they're connected to my problem.

I am than trying to execute the DispersionPlot commend. But this gives me an error:

> disp<-dispersionPlot(genes(cuff))
> disp
[B]Error in seq.default(min, max, by = by) : 
  'from' cannot be NA, NaN or infinite[/B]

Does anyone have an idea why I get this error?
How can I prevent this from happening?

thanks

Assa

#####
> sessionInfo()
R version 3.1.1 (2014-07-10)
Platform: x86_64-apple-darwin13.1.0 (64-bit)

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] grid stats4 parallel stats graphics grDevices utils
[8] datasets methods base

other attached packages:
[1] mgcv_1.8-3 nlme_3.1-118 cummeRbund_2.8.2
[4] Gviz_1.10.2 rtracklayer_1.26.1 GenomicRanges_1.18.1
[7] GenomeInfoDb_1.2.2 IRanges_2.0.0 S4Vectors_0.4.0
[10] fastcluster_1.1.13 reshape2_1.4 ggplot2_1.0.0
[13] RSQLite_1.0.0 DBI_0.3.1 BiocGenerics_0.12.0

loaded via a namespace (and not attached):
[1] acepack_1.3-3.3 AnnotationDbi_1.28.1 base64enc_0.1-2
[4] BatchJobs_1.5 BBmisc_1.8 Biobase_2.26.0
[7] BiocParallel_1.0.0 biomaRt_2.22.0 Biostrings_2.34.0
[10] biovizBase_1.14.0 bitops_1.0-6 brew_1.0-6
[13] BSgenome_1.34.0 checkmate_1.5.0 cluster_1.15.3
[16] codetools_0.2-9 colorspace_1.2-4 dichromat_2.0-0
[19] digest_0.6.4 fail_1.2 foreach_1.4.2
[22] foreign_0.8-61 Formula_1.1-2 GenomicAlignments_1.2.1
[25] GenomicFeatures_1.18.2 gtable_0.1.2 Hmisc_3.14-5
[28] iterators_1.0.7 labeling_0.3 lattice_0.20-29
[31] latticeExtra_0.6-26 MASS_7.3-35 Matrix_1.1-4
[34] matrixStats_0.10.3 munsell_0.4.2 nnet_7.3-8
[37] plyr_1.8.1 proto_0.3-10 R.methodsS3_1.6.1
[40] RColorBrewer_1.0-5 Rcpp_0.11.3 RCurl_1.95-4.3
[43] rpart_4.1-8 Rsamtools_1.18.1 scales_0.2.4
[46] sendmailR_1.2-1 splines_3.1.1 stringr_0.6.2
[49] survival_2.37-7 tools_3.1.1 VariantAnnotation_1.12.3
[52] XML_3.98-1.1 XVector_0.6.0 zlibbioc_1.12.0

[jake13]

I'm having the same issue if anyone has a solution?

[chodar]

Same issue for me. In other forum I read that "#" character in gtf file have to be removed. In my case I dont have # but still have the error.

Thanks.

[Greg Deakin]

I had the same issue as well - it was caused by some extremely low dispersion (and count) values. Some were of the order of 10e-308, which is lower than the minimum floating point value supported by R (on my architecture anyway). I set these to 0 and the plot then worked.

I need to read up a bit more about cuffdiff - obviously it normalises the count values, but I can't see why any should be much less than 1.

Greg

[zsanli]

how to delete most 0 values

I had the same issue as well - it was caused by some extremely low dispersion (and count) values. Some were of the order of 10e-308, which is lower than the minimum floating point value supported by R (on my architecture anyway). I set these to 0 and the plot then worked.

I need to read up a bit more about cuffdiff - obviously it normalises the count values, but I can't see why any should be much less than 1.

Greg

I have same trouble when I draw disperse graph. Do you know how to code to delete most small values and small counts e.g. 2e-168