Comments on: Gene Ontology analysis with Python and Bioconductor

By: Assorted Links of Bioinformatics « Abner’s Postgraduate Days

Assorted Links of Bioinformatics « Abner’s Postgraduate Days — Sat, 24 Mar 2012 16:11:58 +0000

[...] Gene Ontology analysis with Python and Bioconductor « Blue Collar Bioinformatics [...]

By: Zhidong

Zhidong — Wed, 04 Aug 2010 14:29:38 +0000

Thank you, Brad.

By: Brad Chapman

Brad Chapman — Wed, 04 Aug 2010 11:03:09 +0000

Zhidong;
The purpose of the p-value is to select the enriched genes from the background. If you have some other method of identifying the enriched genes, assign them proxy p-values that differentiate them from background. The topDiffGenes function should then be modified to select out the targets you are interested in, and the rest of the analysis can proceed as normal. Hope this helps.

By: Zhidong

Zhidong — Wed, 04 Aug 2010 00:55:17 +0000

Dear Brad,
If I have the predicted target genes of miRNAs with no P values of difference expression, how could I do the functional enrichment analysis of these target genes?

By: Robert

Robert — Thu, 29 Jul 2010 18:01:44 +0000

Hi Brad,

Nice example of GO analysis. However, you can save yourself a bit of work at the end. In each organism specific database in Bioconductor, there is a wrapper to return the mapping of all genes (by Entrez ID) to that particular GO term, thereby saving you the work of traversing the tree.

For example, with the “org.Rn.eg.db” database, you can use: org.Rn.egGO2ALLEGS to return all genes annotated to your terms of interest.

By: Zhidong

Zhidong — Fri, 16 Jul 2010 14:34:45 +0000

Hi Brad,

That is a naive problem for I know little about python and now I have run it successfully following your reply. Thank you very much.

By: Brad Chapman

Brad Chapman — Thu, 15 Jul 2010 15:20:27 +0000

Zhidong;
You’re getting this message because none of the terms in your GO mapping file (ATH_GO_GOSLIM.txt) match the example file (diffexp_example-diffs.csv). You’ll need to supply a real list of terms for the first argument and be sure the names match what is in your GO file. You’ll also need to modify the parse_go_map_file function so you can pull out the Arabidopsis gene names (1st column) and GO terms (6th column).

I updated the code so the error message in this case is more useful and to make it work on the latest rpy2 version. Hope this helps get things working for you.

By: Zhidong

Zhidong — Thu, 15 Jul 2010 09:20:30 +0000

Dear Brad,

Thank you for your great post.

I have try your script and examples but failed(Please see it as following). Could you tell me how to fix it? Thank you in advance.

$python diffexp_go_analysis.py diffexp_example-diffs.csv ATH_GO_GOSLIM.txt

Loading required package: graph
Loading required package: Biobase

Welcome to Bioconductor

Vignettes contain introductory material. To view, type
‘openVignette()’. To cite Bioconductor, see
‘citation(“Biobase”)’ and for packages ‘citation(pkgname)’.

Loading required package: GO.db
Loading required package: AnnotationDbi
Loading required package: DBI
Loading required package: SparseM
Package SparseM (0.85) loaded.
To cite, see citation(“SparseM”)

Attaching package: ‘SparseM’

The following object(s) are masked from ‘package:base’:

backsolve

groupGOTerms: GOBPTerm, GOMFTerm, GOCCTerm environments built.

Building most specific GOs …..Error in split.default(geneID[goodGO], allGO[goodGO]) :
first argument must be a vector
In addition: Warning message:
In is.na(gene2GO) : is.na() applied to non-(list or vector) of type ‘NULL’
Traceback (most recent call last):
File “diffexp_go_analysis.py”, line 130, in
main(sys.argv[1], sys.argv[2])
File “diffexp_go_analysis.py”, line 25, in main
gene_pval, go_pval, topgo_method)
File “diffexp_go_analysis.py”, line 81, in run_topGO
go_data = robjects.r.new(“topGOdata”, **params)
File “/usr/lib64/python2.6/site-packages/rpy2/robjects/__init__.py”, line 423, in __call__
res = super(RFunction, self).__call__(*new_args, **new_kwargs)
rinterface.RRuntimeError: Error in split.default(geneID[goodGO], allGO[goodGO]) :
first argument must be a vector

By: Frederico Arnoldi

Frederico Arnoldi — Tue, 11 May 2010 20:27:40 +0000

Dear Brad Chapman,
Python and R is fantastic combination: the pleasantness of Python together with the amazing power of R and it’s packages. I’ve been using this combination for bioinformatics too, but very probably with fewer familiarity than you. Sometimes I ask myself how robust would be implementing some task with Python/Rpy instead of directly in R. I am sure your posts make Python’s reputation stronger. Thanks for sharing your codes.