Question: NA in padj in the DEXSeq package
gravatar for gil.hornung
2.4 years ago by
gil.hornung0 wrote:


I'm trying to use DEXSeq on RNA-seq data. There is one gene I know is supposed to have differential exon usage, because a few of the exons were knocked-out.

When I look at the data of this gene I see a beautiful plot that suggests differential splicing (as expected, see attached figure), and also very good p-values for differential splicing in several of the exon (see attached figure). However, the padj values are all NA.

I suspect that there is some filtering that is taking place, but I do not know how to control it.


With gratitude,


Gil Hornung

Bioinformatics Analyst


Weizmann Institute of Science


Table of DEXSeq resultsDEXSeq Plot

ADD COMMENTlink modified 2.4 years ago • written 2.4 years ago by gil.hornung0
gravatar for Alejandro Reyes
2.4 years ago by
Alejandro Reyes1.5k
Dana-Farber Cancer Institute, Boston, USA
Alejandro Reyes1.5k wrote:

Dear Gil Hornung,

Thanks for discussing this issue.

One 'pre-eliminary' question, could you include the output of your sessionInfo()?

I guess that the exons that are knocked out are E012, E013, E014 and E015? These exons have indeed a very tiny p-value and the reason they have NA is because they do not pass the independent filtering step. Independent filtering improves power by filtering on low counts, but it is an optional step. The fact that these exons are discarded means is that they are expressed lowly compared to other exons in the genome, but as I mentioned, this step is not 100% necessary. To remove the independent filtering step, you could adjust for multiple testing on the p-values of your DEXSeqDataSet object by using the p.adjust function specifying the parameter method="BH".



ADD COMMENTlink written 2.4 years ago by Alejandro Reyes1.5k
gravatar for gil.hornung
2.4 years ago by
gil.hornung0 wrote:

Dear Alejandro,

Thank you for the fast reply. I suspected that it has something to do with the independent filtering, which I know from DESeq2.

Your reply solved my problem.


Here is the sessionInfo:

R version 3.1.1 (2014-07-10)
Platform: x86_64-unknown-linux-gnu (64-bit)

 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               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    LC_PAPER=en_US.UTF-8       LC_NAME=C                 

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

other attached packages:
 [1] DEXSeq_1.12.2           BiocParallel_1.0.3      Biobase_2.26.0          amap_0.8-14             matrixStats_0.14.2     
 [6] DESeq2_1.6.3            RcppArmadillo_0.5.000.0 Rcpp_0.12.1             GenomicRanges_1.18.4    GenomeInfoDb_1.2.5     
[11] IRanges_2.0.1           S4Vectors_0.4.0         BiocGenerics_0.12.1     ggplot2_1.0.1           gplots_2.17.0          
[16] RColorBrewer_1.1-2     

loaded via a namespace (and not attached):
 [1] acepack_1.3-3.3      annotate_1.44.0      AnnotationDbi_1.28.2 base64enc_0.1-3      BatchJobs_1.6        BBmisc_1.9          
 [7] biomaRt_2.22.0       Biostrings_2.34.1    bitops_1.0-6         brew_1.0-6           caTools_1.17.1       checkmate_1.5.2     
[13] cluster_2.0.3        codetools_0.2-14     colorspace_1.2-6     DBI_0.3.1            digest_0.6.8         fail_1.2            
[19] foreach_1.4.2        foreign_0.8-66       Formula_1.2-1        gdata_2.17.0         genefilter_1.48.1    geneplotter_1.44.0  
[25] grid_3.1.1           gridExtra_2.0.0      gtable_0.1.2         gtools_3.5.0         Hmisc_3.17-0         hwriter_1.3.2       
[31] iterators_1.0.7      KernSmooth_2.23-15   lattice_0.20-31      latticeExtra_0.6-26  locfit_1.5-9.1       magrittr_1.5        
[37] MASS_7.3-40          munsell_0.4.2        nnet_7.3-11          plyr_1.8.2           proto_0.3-10         RCurl_1.95-4.7      
[43] reshape2_1.4.1       rpart_4.1-10         Rsamtools_1.18.3     RSQLite_1.0.0        scales_0.3.0         sendmailR_1.2-1     
[49] splines_3.1.1        statmod_1.4.21       stringi_0.5-5        stringr_1.0.0        survival_2.38-3      tools_3.1.1         
[55] XML_3.98-1.3         xtable_1.7-4         XVector_0.6.0        zlibbioc_1.12.0     
ADD COMMENTlink written 2.4 years ago by gil.hornung0
gravatar for gil.hornung
2.4 years ago by
gil.hornung0 wrote:

A related issue:

I also seem to get NA in the coefficients of some genes, and the log2FoldChange. The counts of such cases are reasonable.

Table of such cases are below, and image of one of these cases is attached

ADD COMMENTlink modified 2.4 years ago • written 2.4 years ago by gil.hornung0

Do you get NAs for all the exons of the gene or only for invididual exons?

ADD REPLYlink written 2.3 years ago by Alejandro Reyes1.5k

Only individual exons

ADD REPLYlink written 2.3 years ago by gil.hornung0
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 136 users visited in the last hour