Entering edit mode
Dear Michael,
I am using DESeq2 for a factorial design. I have Celltype and
condition as colData. I have used the formula ~ Celltype +
Cellypte:condition as suggested by Steve Lianoglou elsewhere and then
I look at the single comparisons in only one Celltype
Everything works fine and I have good adj p values, genes make sene...
The problem arises when I want to make an MA plot of the 2 conditions
for one cell type. I call the function as DESeq2:: since I known there
are some. Here is my code:
> ddsFull
class: DESeqDataSet
dim: 54668 18
exptData(0):
assays(1): counts
rownames(54668): 5S_rRNA 61E3.4 ... snosnR66 yR211F11.2
rowData metadata column names(0):
colnames(18): r3_1394_A r3_2435_A ... r3_1395_A r3_1395_B
colData names(2): condition Celltype
> design(ddsFull)<-formula(~ Celltype + Celltype:condition)
> ddsFull.fact<-DESeq(ddsFull)
estimating size factors
estimating dispersions
gene-wise dispersion estimates
mean-dispersion relationship
final dispersion estimates
fitting model and testing
18 rows did not converge in beta, labelled in mcols(object)$betaConv.
Use larger maxit argument with nbinomWaldTest
> resultsNames(ddsFull.fact)
[1] "Intercept" "Celltype_R6_vs_R3"
"CelltypeR3.conditionKD" "CelltypeR6.conditionKD"
"CelltypeR3.conditionOE" "CelltypeR6.conditionOE"
Now, the problem:
If I use plotMA as for the vignette:
> DESeq2::plotMA(ddsFull.fact)
I obtain a right graph but for the last value in the design formula:
But if I try to plot for my contrasts of interest:
> DESeq2::plotMA(ddsFull.fact, lfcColname="CelltypeR3.conditionKD")
It looks as if the red points are spread over with no correlation to
fold change, even with red points over the 0 fold change.
I did not find the pvalueColname variable anymore on plotMA, I guess
is the same as log2FC.
If I create the results objects by:
> resFull.KD<-results(ddsFull.fact,name="CelltypeR3.conditionKD")
> resFull.OE<-results(ddsFull.fact,name="CelltypeR3.conditionOE")
and try to make my own MA plot by:
> col=ifelse(resFull.KD$padj>=0.1,"black","red3");plot(resFull.KD$base
Mean,resFull.KD$log2FoldChange,log="x",col=col)
It looks pretty convincing with the red points being those external to
the cloud of maximal fold changes an high baseMean as it should
> sessionInfo()
R version 3.0.2 (2013-09-25)
Platform: x86_64-apple-darwin10.8.0 (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 parallel stats graphics grDevices utils
datasets methods base
other attached packages:
[1] parathyroidSE_1.0.2 pasilla_0.2.16 DESeq_1.13.3
lattice_0.20-23 locfit_1.5-9.1 DEXSeq_1.7.22
BiocInstaller_1.12.0
[8] gplots_2.11.3 MASS_7.3-29
KernSmooth_2.23-10 caTools_1.14 gdata_2.13.2
gtools_3.1.0 AnnotationDbi_1.23.27
[15] Biobase_2.21.7 RSQLite_0.11.4 DBI_0.2-7
edgeR_3.3.8 limma_3.17.26 DESeq2_1.2.0
RcppArmadillo_0.3.910.0
[22] Rcpp_0.10.4 GenomicRanges_1.13.51 XVector_0.1.4
IRanges_1.19.38 BiocGenerics_0.7.5
loaded via a namespace (and not attached):
[1] annotate_1.39.0 biomaRt_2.17.3
Biostrings_2.29.19 biovizBase_1.9.4 bitops_1.0-6
BSgenome_1.29.1 cluster_1.14.4
[8] colorspace_1.2-3 dichromat_2.0-0 digest_0.6.3
evaluate_0.5 formatR_0.9 genefilter_1.43.0
geneplotter_1.39.05
[15] GenomicFeatures_1.13.47 ggplot2_0.9.3.1 GO.db_2.10.1
gridSVG_1.3-1 gtable_0.1.2 Gviz_1.5.15
Hmisc_3.12-2
[22] httpuv_1.1.0 hwriter_1.3
interactiveDisplay_1.0.0 knitr_1.5 labeling_0.2
latticeExtra_0.6-26 munsell_0.4.2
[29] plyr_1.8 proto_0.3-10
RColorBrewer_1.0-5 RCurl_1.95-4.1 reshape2_1.2.2
RJSONIO_1.0-3 rpart_4.1-3
[36] Rsamtools_1.13.48 rtracklayer_1.21.12 scales_0.2.3
shiny_0.7.0 splines_3.0.2 statmod_1.4.18
stats4_3.0.2
[43] stringr_0.6.2 survival_2.37-4 tools_3.0.2
XML_3.95-0.2 xtable_1.7-1 zlibbioc_1.7.0
The same code had been used with older versions of DESeq2 (<1.13) with
good results, the ones that did not contain the 'contrast' parameter
in results(). Now with the newest I have this issue.
What can be the trouble?
Thanks
Jose M. Garcia Manteiga PhD
Data analyst in Functional Genomics
Center for Translational Genomics and Bioinformatics
DIBIT2-A3 Room 21
San Raffaele Scientific Institute
Via Olgettina 58
20132 Milano
Italy
Office: +39 02 26439114
[[alternative HTML version deleted]]
