Tool:DEA.R: Command line script for differential expression analysis with R packages (DESeq2/edgeR/limma-voom)

Question

[WouterDeCoster]

DEA.R

I wrote a command line Rscript to perform differential expression analysis:

• Conveniently
• Quick
• Reproducibly

The script uses the R packages DESeq2, edgeR and limma-voom. It's available on github. There is still room for improvement, so if you would be interested in contributing I'd definitely welcome that.

Features

• Performs counting using featureCounts
• Allows specification of covariates
• Rigorous checking of input data
• Creates various plots
• Creates detailed tables and lists of differentially expressed genes

Usage

DEA.R  < sample info file > < annotation.gtf >

with:

• sampleInfoFile: file specifying samples with additional covariates
• annotation.gtf An annotation file in gtf format matching the reference genome used for alignment

More information can be found in the README on github

The script performs a few sanity checks on input data and then executes differential expression analysis with DESeq2, edgeR and limma-voom. The results are written to tab delimited files and the script saves all potential useful plots in image files.

Changelog

2017-05-20: major update in code and usage, added readme
2017-05-27: minor update in code and usage

[GenoMax]

@All: If you have your own implementations that offer similar functionality then include them here. That way we will have a collection in one post.

[GenoMax]

Can you add an example command line (and some dummy input file/comparison conditions) to this post? People may not realize the utility of the script at the link above by just looking at it.

[WouterDeCoster]

Oh, right, edited the original post. Thanks!

[Istvan Albert]

Needs a readme, example usage, example and test datasets and if it all checks out think about writing it up in say:

BMC: sourcecode for biology and medicine

[WouterDeCoster]

Thanks for the advice, and I definitely agree with the first parts and I'll make sure to add those. But since this is just code put together from manuals and forum posts here and stackoverflow, does it make sense in publishing that? I'm not sure that I can take credit.

[Sinji]

Looks like you could publish it as a 'Workflow' manuscript. Though Istvan will be able to better comment on it's feasibility.

[ivivek_ngs]

Way to go. I saw it earlier. Was waiting for you to put it here. Great job done. I am thinking if I should also put the automate script with salmon or not that gives both count and tpm tables. Post that I usually use a procedural script based on my need. But definitely the DEA.r will be worth taking a shot for enchantments if I use it for our lab as well. Let me know what you guys think.

[olechnwin]

Thinking about using this instead of my own script. Are there examples of outputs including tables and figures that are generated?

[WouterDeCoster]

Not right now, but I could generate those. Note that this isn't actively maintained anymore and there are multiple alternatives including GUIs. But you are most welcome to use the tool :)

[olechnwin]

Thanks for letting me know this!

Do you happen to know what are the alternatives? I'm looking more for command line script since it takes a long time to process these rna-seq files.

Answer 1

[Devon Ryan]

Our RNAseq pipeline is here and the DESeq2 Rscript that's part of it here. I really can't claim credit for any of that, it's all written by a post-doc (he's much better at writing understandable code than me, so I'm happy that he maintains these). If anyone needs a ChIPseq pipeline then have a look at his "snakemake_workflows" repository.

Answer 2

[h.mon]

I was searching for DESeq2 citation and stumbled upon these four: hppRNA, SARTools, Glimma, EnrichmentBrowser. I am not involved with any of those, but they may be sources of interesting ideas for those developing ppelines.

In fact, taking the idea from EnrichmentBrowser, you could add functional enrichment analysis to your pipeline.

[WouterDeCoster]

Thanks for the suggestions!