RNA-Seq Pathway and Gene-set Analysis Workflows in R/Bioconductor with GAGE/Pathview

[bigmw]

The gage package (2.12.0) now includes a new tutorial, "RNA-Seq Data Pathway and Gene-set Analysis Workflows". Note you need to update to current release versions of R(3.0.2)/ Bioconductor(2.13) to use all the features. Please check it out:

http://bioconductor.org/packages/release/bioc/html/gage.html

http://bioconductor.org/packages/release/bioc/vignettes/gage/inst/doc/RNA-seqWorkflow.pdf

We first cover a full workflow from preparation, reads counting, data preprocessing, gene set test, to pathway visualization in about 40 lines of codes. The same workflow can be used for GO analysis or other types of gene set analysis too. We also describe joint workflows, i.e. to do gene-level analysis using one of the major RNA-Seq analysis tools, DEseq/DEseq2, edgeR, limma and Cufflinks, and feed the results into GAGE/Pahview for pathway analysis or visualization. All these workflows are implemented in R/Bioconductor.
Comments and questions are welcome. Thanks!

[bigmw]

GAGE and Pathview can be used independent of each other. GAGE does pathway and Gene-set Analysis, and works on other tyeps of gene sets than pathways, like GO, coexpressed/coregulated gene sets, TF or miRNA target lists etc. Pathview may integrate and visualizeuser data onto pathway graphs independent of pathway analysis procedure.

Pathview package available at:

http://bioconductor.org/packages/release/bioc/html/pathview.html

Here is the info page with example output:

http://pathview.r-forge.r-project.org/

[entrez]

I’ve both gage and pathview installed on my computer. I tried to follow the example in the native workflow. Things work well, except I didn’t get 4 samples show up in the same graph (or nodes with 4 slices) as in Figure 2 of the workflow document, instead I got 4 separate graphs. What I might have done wrong?

[bigmw]

What versions of gage, pathview and Bioconductor you have?

[entrez]

gage 2.10.0, pathview 1.1.4 and Bioconductor 2.12

[bigmw]

pathview 1.1.4 does not show multiple samples/states in the same graph, you need to upgrade to the current release, which is 1.2.0: http://bioconductor.org/packages/rel.../pathview.html.
I would recommend to do an overall upgrade to R 3.0.2/Bioconductor 2.13, which will update your pathview and gage to the latest version too.

[bigmw]

If you don’t know how to upgrade Bioc, please check:

http://www.bioconductor.org/install/#update-bioconductor-packages

Here is some work around if you get problems:

https://stat.ethz.ch/pipermail/bioconductor/2013-October/055642.html

[entrez]

Can I use the workflow (with necessary changes) for microarray data analysis? If so, how?

[bigmw]

GAGE/Pathview workflow can be applied for microarray data analysis. Please check the main tutorials of gage and pathview for details:

http://bioconductor.org/packages/release/bioc/vignettes/gage/inst/doc/gage.pdf

http://bioconductor.org/packages/release/bioc/vignettes/pathview/inst/doc/pathview.pdf

[bigmw]

Pathview is actually applicable to any data mappable to pathways, including gene, protein, metabolite, genetics, literature, and others. The tutorial describes examples on metabolite/compound data too.

[crazyhottommy]

Hi I was playing around with GAGE, one question is that I got the count table by HTSeq, and the ids are gene names for each row, how can I change the gene names to GO term ids?

Thanks

[bigmw]

You don’t have to change gene names/IDs to GO term IDs. GAGE (or other gene set analysis tools) requires two major input data objects: your expression data (vector or matrix-like) and gene set list (list of gene ID vectors). Make sure your gene IDs in expression data and gene set list are the same type, i.e. both are Entrez Gene IDs, or both gene symbols, etc.
You may want to go through the basics and common use of gage described in the main gage vignette:

http://bioconductor.org/packages/release/bioc/vignettes/gage/inst/doc/gage.pdf

if you want a quick start, section 1, 6 and 7 (page 1, 4-8) would be enough. You will see examples for both KEGG and GO analysis.

[bigmw]

If you follow the RNA-seq workflows (links in the first post above), we can actually work on the demo examples from Step 2. In other words, we can start with the pre-mapped raw read counts data (from previous steps), i.e. hnrnp.cnts stored in gageData. I would suggest you to run the demo example and explore gage/pathview functions and input/output data by yourself.

[crazyhottommy]

Thank you!

[wilson90]

In your vignette, we are suppose to provide our annotation file.
I wonder where have you obtained "kegg.gs"?
and I want to use GO annotation. So where can I obtain "GO.gs" in R?
Thank you.

Frustrated user