The example DEXSeqDataSet code gives and error in version 1.17.4 and 1.16.3.

Thanks

Alex

  countData <- matrix( rpois(10000, 100), nrow=1000 )
  sampleData <- data.frame(
      condition=rep( c("untreated", "treated"), each=5 ) )
  design <- formula( ~ sample + exon + condition:exon )
  groupID <- rep(
      paste0("gene", 1:10),
      each= 100 )
  featureID <- rep(
      paste0("exon", 1:10),
      times= 100 )
  DEXSeqDataSet( countData, sampleData, design,
              featureID, groupID )

Error:

converting counts to integer mode
Error in `$<-.data.frame`(`*tmp*`, "dispersion", value = NA) :
  replacement has 1 row, data has 0

I was also confronted this problem in version 1.16.3 and the "DEXSeqDataSetFromHTSeq"  also returns the error as well.

R version 3.2.2 (2015-08-14)

Platform: x86_64-redhat-linux-gnu (64-bit)

Running under: CentOS release 6.5 (Final)

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=en_US.UTF-8       LC_NAME=C                 

 [9] LC_ADDRESS=C               LC_TELEPHONE=C            

[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:

[1] stats4    parallel  stats     graphics  grDevices utils     datasets 

[8] methods   base     

other attached packages:

 [1] DEXSeq_1.16.3              DESeq2_1.10.0             

 [3] RcppArmadillo_0.6.200.2.0  Rcpp_0.12.2               

 [5] SummarizedExperiment_1.0.1 GenomicRanges_1.22.1      

 [7] GenomeInfoDb_1.6.1         IRanges_2.4.4             

 [9] S4Vectors_0.8.3            Biobase_2.30.0            

[11] BiocGenerics_0.16.1        BiocParallel_1.4.0        

loaded via a namespace (and not attached):

 [1] genefilter_1.52.0    statmod_1.4.22       locfit_1.5-9.1      

 [4] reshape2_1.4.1       splines_3.2.2        lattice_0.20-33     

 [7] colorspace_1.2-6     survival_2.38-3      XML_3.98-1.3        

[10] foreign_0.8-66       DBI_0.3.1            RColorBrewer_1.1-2  

[13] lambda.r_1.1.7       plyr_1.8.3           stringr_1.0.0       

[16] zlibbioc_1.16.0      Biostrings_2.38.2    munsell_0.4.2       

[19] gtable_0.1.2         futile.logger_1.4.1  hwriter_1.3.2       

[22] latticeExtra_0.6-26  geneplotter_1.48.0   biomaRt_2.26.1      

[25] AnnotationDbi_1.32.0 proto_0.3-10         acepack_1.3-3.3     

[28] xtable_1.8-0         scales_0.3.0         Hmisc_3.17-0        

[31] annotate_1.48.0      XVector_0.10.0       Rsamtools_1.22.0    

[34] gridExtra_2.0.0      ggplot2_1.0.1        digest_0.6.8        

[37] stringi_1.0-1        grid_3.2.2           tools_3.2.2         

[40] bitops_1.0-6         magrittr_1.5         RCurl_1.95-4.7      

[43] RSQLite_1.0.0        Formula_1.2-1        cluster_2.0.3       

[46] futile.options_1.0.0 MASS_7.3-45          rpart_4.1-10        

[49] nnet_7.3-11         

Therefore it should be  bugs in function "DEXSeqDataSet". I noticed the last few line of the function

modelFrame <- makeBigModelFrame(dds)

is differ from the old version and the code in git.

@pensona: Thanks a lot for reporting this. I have submitted a fix both in the devel and the release branch, and bumped the version numbers for both cases. The problem was that the makeBigModelFrame function was invoking rowRanges(object)$groupID, however, these resulted to be NULL when the genomic annotation of the exons was not present. I changed it to mcols(object)$groupID and it works fine now. 

@zb.picb: Thanks for pointing to the right direction! Were you providing the file path to the annotation gtf when you were creating your object? I could not reproduce the error with the DEXSeqDataSetFromHTSeq function when I provided the annotation file. But the latest versions in the svn should work now with and without the annotation file. 

Let me know!

Alejandro

@Alejandro, Thanks very much for replying so quick. I didn't provide the annotation file when run DEXSeqDataSetFromHTSeq since I had used it to calculate PAS usage instead of exon and it's absent of annotation.  As your mentioned the fixed version above, is it available in Bioconductor ? I downloaded  and installed the develop version DEXSeq_1.17.4.tar.gz , the same error occurred.  

"other attached packages:

 [1] DEXSeq_1.17.4              DESeq2_1.10.0             

 [3] RcppArmadillo_0.6.200.2.0  Rcpp_0.12.2               

 [5] SummarizedExperiment_1.0.1 GenomicRanges_1.22.1      

 [7] GenomeInfoDb_1.6.1         IRanges_2.4.4             

 [9] S4Vectors_0.8.3            Biobase_2.30.0            

[11] BiocGenerics_0.16.1        BiocParallel_1.4.0"

>  DEXSeqDataSet( countData, sampleData, design,

+               featureID, groupID )

converting counts to integer mode

Error in `$<-.data.frame`(`*tmp*`, "dispersion", value = NA) : 

  replacement has 1 row, data has 0

In addition: Warning message:

In DESeqDataSet(rse, design, ignoreRank = TRUE) :

  900 duplicate rownames were renamed by adding numbers

Do you know if I install the correct version and  what's the reason it still not worked ?  Shall I delete the old version and try it again after installing new one.  Look forward to you help. 

@Alejandro, I replaced the code in DEXSeq_1.17.4.tar.gz with the source from Git and problem solved.  Thanks  a lot.