Concerning the collection of predicted targets,

I would suggest you to use miRWalk database (http://mirwalk.uni-hd.de). Because, this database offers a comparative platform of possible miRNA-target predictions using 10 different data sets i.e. miRWalk, Dianamt, miRanda, RNA22, miRDB, TargetScan, RNAhybrid, PITA, PICTAR4 and PICTAR5.

You can use following tools...

Miranda, ([email protected])

MiRDB, (mirdb.org)

miRNA – Target Gene Prediction at EMBL

RNAhybrid, : (http://bibiserv.techfak.uni-bielefeld.de/rnahybrid/)

Tarbase(http://diana.cslab.ece.ntua.gr/tarbase/),

there are a lot more...

You can do this analysis by creating your own data sets (KEGG and GO foreground gene sets) vs known genes in R using BuildKEGG, limma.

Data Sets required for this analysis are:

1. You input list (number of targets of a given miRNA).

2. Gene sets of KEGG pathways and GO.

3. Known genes (background set).

4. Fisher's exact test

5. Multiple testing.

Eg. Enrichment of miR-x within Mapk pathway with respect to all known genes

Step1. 2x2 table

Col1\tCol2

125 (predicted targets within Mapk \t 2500 (predicted targets of mir-x in genome

250 (MAPK set) \t 19965 (known genes as background set)

Step2: Fishers' exact test or chi-square

Hope this helps!

Concerning the collection of predicted targets,

I would suggest you to use miRWalk database (http://mirwalk.uni-hd.de). Because, this database offers a comparative platform of possible miRNA-target predictions using 10 different data sets i.e. miRWalk, Dianamt, miRanda, RNA22, miRDB, TargetScan, RNAhybrid, PITA, PICTAR4 and PICTAR5.

Did you already give a look to sylamer?

Http://www.ebi.ac.uk/enright/sylamer/

There are number of bioconductor tools and I suggest that is the best place to start with. Any way I will give you some links here and check them out. It is also better to use BSGenome package where you can write your own codes to analyze miRNA mainly the coordinates and surrounding areas in a quite efficient manner.

Some of these packages are not relevant to you but if you look at the examples some times you might find a way around to use some of these packages to answer you questions or to design ideas for your problem oriented scripts.

http://www.bioconductor.org/packages/2.11/bioc/html/miRNApath.html

http://www.bioconductor.org/packages/2.11/bioc/html/MmPalateMiRNA.html

http://www.bioconductor.org/packages/2.12/bioc/html/RmiR.html

http://www.bioconductor.org/packages/2.12/data/annotation/html/RmiR.Hs.miRNA.html

http://www.bioconductor.org/packages/2.12/BiocViews.html#___AnnotationData,Homo_sapiens,CustomDBSchema,miRNA

http://www.bioconductor.org/packages/2.12/BiocViews.html#___Software,GeneExpression,Microarray,TimeCourse,Visualization

Hope this would at least give you a start.

In addition to my previous answer I also strongly recommend Harsh suggestions.

We wrote CORNA for just this task:

http://bioinformatics.oxfordjournals.org/content/25/6/832.long

As Mick said most of the tools are mainly organism oriented or a particular question specific, but the beauty is that they provide enormous amount of ideas as to how one could tackle problem in hand in a creative manner.

Don't know if these packages could work for you: DESeq or EdgeR.

DESeq is a tool to perform normalization of RNA-Seq libraries and identifying differentially expressed genes and EdgeR does more or less the same, but produce different results some times. It is not that they disagree in producing different sets of DE genes, but differences in number of genes differentially expressed.

So they could be used to analyze differentially regulated miRNA, but it depends on whether your sequencing approach cover total transcriptom, total miRNA or a subset of miRNA I think, since both of these tools try to address reasons behind libraries with different sequencing depths and resizing the libraries to obtain realistic profiling of the expression levels.

I don't know if this package could work for you:

http://www.broadinstitute.org/cancer/software/gsea/wiki/index.php/R-GSEA_Readme

I would recommend you to try my package http://bioconductor.org/packages/release/bioc/html/clusterProfiler.html