Question

searching for edgeR version 2.2.6

1

Entering edit mode

elena.capel ▴ 10

@elenacapel-12192

Last seen 7.9 years ago

Hi,

I need to run an old version of edgeR 2.2.6. I found it on Bioconductor:

https://bioc.ism.ac.jp/packages/2.8/bioc/html/edgeR.html

The problem is that when I install it, the package corresponds actually to a newer version 3.8.6:

tar xzvf edgeR_2.2.6.tar
/usr/bin/R --vanilla --args TMM BH tagwise 5

R version 3.1.2 (2014-10-31) -- "Pumpkin Helmet"
Copyright (C) 2014 The R Foundation for Statistical Computing
Platform: x86_64-apple-darwin10.8.0 (64-bit)
> Args <- commandArgs(TRUE)
> options(error=expression(NULL))
> setwd("/Users/Elena/Tnseq_ESSENTIALS_analysis_All_with_BG20h_runB_14jan2017/")
> library(edgeR)
Le chargement a nécessité le package : limma
Message d'avis :
le package ‘limma’ a été compilé avec la version R 3.1.3 

> sessionInfo()
R version 3.1.2 (2014-10-31)
Platform: x86_64-apple-darwin10.8.0 (64-bit)
locale:
[1] fr_FR.UTF-8/fr_FR.UTF-8/fr_FR.UTF-8/C/fr_FR.UTF-8/fr_FR.UTF-8
attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     
other attached packages:
[1] edgeR_3.8.6  limma_3.22.7

Any suggestion please? Thanks a lot in advance!

edgeR • 1.4k views

ADD COMMENT • link updated 7.9 years ago by Martin Morgan 25k • written 7.9 years ago by elena.capel ▴ 10

0

Entering edit mode

Why do you want to install edgeR 2.2.6? That seems a very strange thing indeed to do.

ADD REPLY • link 7.9 years ago Gordon Smyth 51k

0

Entering edit mode

Actually, the reason why I want to use that version is that I already have some command lines working on that version and not on the new ones. In particular, when I enter the following command I get an error:

> de.tagwise <- glmLRT(d, glmfit.tgw, coef=4:9)

Erreur dans glmLRT(d, glmfit.tgw, coef = 4:9) :

glmfit must be an DGEGLM object (usually produced by glmFit).

ADD REPLY • link 7.9 years ago elena.capel ▴ 10

score 2 · Answer 1 · 2017-01-18

2

Entering edit mode

Martin Morgan 25k

@martin-morgan-1513

Last seen 4 months ago

United States

Bioconductor has 'release' versions tied to versions of R. edgeR 2.2.6 was in Bioconductor release 2.8. From this page, Bioconductor 2.8 was based on R 2.13. It was released in April, 2011 -- this is a very very old version of the software, and it is important to ask what the justification is for using this version.

Your strategy will be to install R version 2.13, and then use standard Bioconductor installation commands

source("http://bioconductor.org/biocLite.R")
biocLite("edgeR")

to install the desired version of edgeR. This may require some significant additional effort, especially identifying contemporary versions of CRAN packages.

From what you posted, it seemed that you downloaded but did not install the old version of edgeR. One could install the old version into R from the downloaded tarball (via install.packages("edgeR_2.2.6.tar.gz", repos=NULL, type="source'). (Actually, this version does seem to install, and indeed pass R CMD check, which itself is somewhat surprising). However, it is not sufficient to install an old edgeR into a modern R -- too many changes have occurred for the packages to interoperate reliably.

ADD COMMENT • link 7.9 years ago Martin Morgan 25k

0

Entering edit mode

Thanks for your answer. Actually, the reason why I want to use that version is that I already have some command lines working on that version and not on the new ones. In particular, when I enter the following command I get an error:

> de.tagwise <- glmLRT(d, glmfit.tgw, coef=4:9)

Erreur dans glmLRT(d, glmfit.tgw, coef = 4:9) :

glmfit must be an DGEGLM object (usually produced by glmFit).

ADD REPLY • link 7.9 years ago elena.capel ▴ 10

0

Entering edit mode

I would advise updating your analysis script, rather than depending on an out-of-date version. There have been many methodological improvements that you should make use of. Probably you should do:

de.tagwise <- glmLRT(glmfit.tgw, coef=4:9)

... if glmfit.tgw is the DGEGLM object produced by glmFit. In fact, I don't even remember a time when glmLRT accepted both a DGEList (presumably that's what d is) and a DGEGLM object.

ADD REPLY • link 7.9 years ago Aaron Lun ★ 28k

0

Entering edit mode

I absolutely agree with Aaron. Bioconductor 2.8 was released in April 2011, so by using edgeR v2.2.6 you're missing out on almost 6 years of methodological improvements and bug fixes. I *can* remember a time when glmLRT accepted those arguments, but nevertheless strongly advise you to update your scripts to use the current release version. On 18/01/2017 16:13, Aaron Lun [bioc] wrote: > Activity on a post you are following on support.bioconductor.org > <https: support.bioconductor.org=""> > > User Aaron Lun <https: support.bioconductor.org="" u="" 6732=""/> wrote > Comment: searching for edgeR version 2.2.6 > <https: support.bioconductor.org="" p="" 91347="" #91365="">: > > I would advise updating your analysis script, rather than depending on > an out-of-date version. There have been many methodological > improvements since then. In fact, I don't even remember a time when > |glmLRT| accepted those kind of arguments. > > ------------------------------------------------------------------------ > > Post tags: edgeR > > You may reply via email or visit > C: searching for edgeR version 2.2.6 > -- Davis McCarthy NHMRC Early Career Fellow Stegle Group EMBL-EBI, Cambridge, UK www.ebi.ac.uk

ADD REPLY • link 7.9 years ago davis ▴ 90

0

Entering edit mode

Thanks a lot Aaron.

I totally agree with you but since I am not an "R" professional I was just trying to reproduce some scripts. By the way, I tried changing the script as you suggested and now it worked!

ADD REPLY • link 7.9 years ago elena.capel ▴ 10