Question

Problem reproducing PCA plot in DeSeq2 totuorial

0

Entering edit mode

sarahmacdonald86 • 0

@sarahmacdonald86-12162

Last seen 7.7 years ago

Hello,

I am following the Differential analysis of count data – the DESeq2 package vignette which is great, however when I come to and can produce heatmaps, however when i try a PCA plot i get the following error.

plotPCA(rld, intgroup=c("LesionScore", "Treatment")

Error in (function (classes, fdef, mtable) :
unable to find an inherited method for function ‘exprs’ for signature ‘"DESeqTransform"’

My data has been input as a matrix following the phyloseq DeSeq2 tutorial

I have found this link where someone has the same problem, however I tried the suggested fix and this didnt work for me

http://seqanswers.com/forums/archive/index.php/t-28350.html

I have been using a mixture of install.packages() and biocLite(), do I need to go through and install all packages using bioLite()?

Cheers,

Sarah

DeSeq2 PCA plot • 4.3k views

ADD COMMENT • link 7.9 years ago sarahmacdonald86 • 0

0

Entering edit mode

Hi James,

I tried this and simply get this:

[1] TRUE

Help!

cheers,

Sarah

ADD REPLY • link 7.9 years ago sarahmacdonald86 • 0

0

Entering edit mode

To add comments, please click the ADD COMMENT button and use the box that pops up, rather than the 'Add your answer' box below.

What is the output from sessionInfo()?

ADD REPLY • link 7.9 years ago James W. MacDonald 67k

0

Entering edit mode

Sorry yes, better to add comment.

> sessionInfo()

R version 3.3.2 (2016-10-31)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: macOS Sierra 10.12

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

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

other packages:
[1] vsn_3.42.3 BiocInstaller_1.24.0
[3] RColorBrewer_1.1-2 pheatmap_1.0.8
[5] plyr_1.8.4 DESeq_1.26.0
[7] lattice_0.20-34 locfit_1.5-9.1
[9] DESeq2_1.14.1 SummarizedExperiment_1.4.0
[11] Biobase_2.34.0 GenomicRanges_1.26.2
[13] GenomeInfoDb_1.10.2 IRanges_2.8.1
[15] S4Vectors_0.12.1 BiocGenerics_0.20.0
[17] devtools_1.12.0 ggplot2_2.2.1
[19] phyloseq_1.19.1

loaded via a namespace (and not attached):
[1] jsonlite_1.2 splines_3.3.2 foreach_1.4.3
[4] Formula_1.2-1 assertthat_0.1 affy_1.52.0
[7] latticeExtra_0.6-28 RSQLite_1.1-2 backports_1.0.4
[10] limma_3.30.7 digest_0.6.11 XVector_0.14.0
[13] checkmate_1.8.2 colorspace_1.3-2 preprocessCore_1.36.0
[16] htmltools_0.3.5 Matrix_1.2-7.1 XML_3.98-1.5
[19] genefilter_1.56.0 zlibbioc_1.20.0 xtable_1.8-2
[22] scales_0.4.1 affyio_1.44.0 BiocParallel_1.8.1
[25] htmlTable_1.8 tibble_1.2 annotate_1.52.1
[28] mgcv_1.8-16 withr_1.0.2 nnet_7.3-12
[31] lazyeval_0.2.0 survival_2.40-1 magrittr_1.5
[34] memoise_1.0.0 nlme_3.1-128 MASS_7.3-45
[37] foreign_0.8-67 vegan_2.4-1 tools_3.3.2
[40] data.table_1.10.0 stringr_1.1.0 munsell_0.4.3
[43] cluster_2.0.5 AnnotationDbi_1.36.0 Biostrings_2.42.1
[46] ade4_1.7-5 rhdf5_2.18.0 grid_3.3.2
[49] RCurl_1.95-4.8 iterators_1.0.8 biomformat_1.2.0
[52] igraph_1.0.1 labeling_0.3 bitops_1.0-6
[55] base64enc_0.1-3 gtable_0.2.0 codetools_0.2-15
[58] multtest_2.30.0 DBI_0.5-1 reshape2_1.4.2
[61] gridExtra_2.2.1 knitr_1.15.1 Hmisc_4.0-2
[64] permute_0.9-4 ape_4.0 stringi_1.1.2
[67] Rcpp_0.12.8 geneplotter_1.52.0 rpart_4.1-10
[70] acepack_1.4.1

ADD REPLY • link 7.9 years ago sarahmacdonald86 • 0

score 0 · Answer 1 · 2017-01-12

0

Entering edit mode

James W. MacDonald 67k

@james-w-macdonald-5106

Last seen 8 hours ago

United States

You can check your installation by doing

library(BiocInstaller)

biocValid()

And following the instructions that pop up.

ADD COMMENT • link 7.9 years ago James W. MacDonald 67k

1

Entering edit mode

I'd guess that you have another package loaded which is masking plotPCA, by not using the generic in BiocGenerics. There are a set of packages that "play by the rules" allowing for proper dispatch, but others may mask it. The user can get around this in the end by using DESeq2::plotPCA.

ADD REPLY • link 7.9 years ago Michael Love 42k

0

Entering edit mode

Hi Michael,

Yes this seemed to be the problem, loaded a new session with only the packages I needed and this worked.

Thanks everyone.

ADD REPLY • link 7.9 years ago sarahmacdonald86 • 0

0

Entering edit mode

Ha! it seems Jim is the offender!

I wouldn't have written the comment above in such a pedantic way if I knew it was you :)

https://github.com/Bioconductor-mirror/affycoretools/blob/dcefa745c46b88a8f8d8ce6435687514b60389c3/R/affystart.R#L377

ADD REPLY • link 7.9 years ago Michael Love 42k

0

Entering edit mode

Urgh. Is this the part where I wave my cane and tell you damn kids to get off my lawn?

I'll change my version of the function to S4 to lower my OQ (offender quotient) to a lower but probably still offensive level. ;-D

ADD REPLY • link 7.9 years ago James W. MacDonald 67k

1

Entering edit mode

So one of the reasons I never converted plotPCA in affycoretools to S4 was because I felt I would be hijacking the method for ExpressionSets. I'm not sure what happens if two packages get loaded with the same methods dispatching on the same signature, but I've made the changes so DESeq2 and affycoretools no longer clash.

> library(affycoretools)
> library(DESeq2)
> plotPCA
nonstandardGenericFunction for "plotPCA" defined from package "BiocGenerics"

function (object, ...)
{
    standardGeneric("plotPCA")
}
<environment: 0x1c57e70>
Methods may be defined for arguments: object
Use  showMethods("plotPCA")  for currently available ones.
> showMethods(plotPCA)
Function: plotPCA (package BiocGenerics)
object="DESeqTransform"
object="ExpressionSet"