[BioC] get al microRNA target sites for a given gene

[James F. Reid]

Dear Javier,

On 16/10/12 15:03, James W. MacDonald wrote:
> Hi Javier,
>
> On 10/16/2012 3:06 AM, Javier Pérez Florido wrote:
>> Hi James,
>> Thanks for your reply.
>> What I want is something like this:
>> http://docs.bioinfo.cipf.es/projects/cellbase/wiki/Feature_rest_ws_api
>>
>> Cellbase retrieves all microRNA targets sites for a given gene, for
>> example, for BRCA2 gene:
>> *
>> http://ws.bioinfo.cipf.es/cellbase/rest/latest/hsa/feature/gene/BRCA2/mirna_target
>>
>>
>>
>>
>> *I can use this webservice through R, but would like to know if some
>> Bioconductor package can handle this....
>
> There is the mirbase.db package, but you would need to poke around to
> see what it can do - I have never found it particularly helpful for my
> use cases.
>
> Best,
>
> Jim
>
>
>> Thanks again
>> Javier
>>
>>
>> On 15/10/12 17:53, James W. MacDonald wrote:
>>> Hi Javier,
>>>
>>> On 10/15/2012 11:42 AM, Javier Pérez Florido wrote:
>>>> Dear list,
>>>> How can I obtain all microRNA target sites for a given gene?
>>>
>>> Do you want miRNA target *sites* or target genes? Also, what do you
>>> mean by 'all'? To my knowledge, miRNA->mRNA mapping is based on
>>> sequence similarity, which implies a cutoff that can be varied (so
>>> 'all' is a relative term).
>>>
>>> Anyway, you could use the microCosm files that you can download here:
>>>
>>> http://www.ebi.ac.uk/enright-srv/microcosm/cgi-bin/targets/v5/download.pl
>>>
>>>
>>> The files contain the Ensembl transcript ID, as well as the chr,
>>> start and end. So a simple tapply() (or by()) will give you the
>>> by-transcript miRNAs, and you can go from there.
>>>
>>> Best,
>>>
>>> Jim
>>>
>>>
>>>> I've checked other packages such as miRBase, but it provides
>>>> information in the other way....(input is a microRNA).
>>>>
>>>> Thanks,
>>>> Javier
>>>>
>>>> _______________________________________________
>>>> Bioconductor mailing list
>>>> Bioconductor at r-project.org
>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>>> Search the archives:
>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>
>>
>

I don't think that a package which aggregates microRNA target site 
prediction from different resources, such as the one you mentioned 
(cellbase), is currently available on Bioconductor.
The only one that I know of that is a _single_ source of predicted 
target sites is targetscan which is on Bioconductor for both human and 
mouse which is updated regularly.
So for example from your example with BRCA2:

 > library("targetscan.Hs.eg.db")
 > library("org.Hs.eg.db")
 > get(get("BRCA2", revmap(org.Hs.egSYMBOL)), targetscan.Hs.egTARGETS)
Error in .checkKeys(value, Lkeys(x), x at ifnotfound) :
   value for "675" not found
this is because no conserved sites are predicted in the 3' UTR of BRCA2 
by targetscan, on the other hand BRCA1 does:
 > get(get("BRCA1", revmap(org.Hs.egSYMBOL)), targetscan.Hs.egTARGETS)
  [1] "miR-132/212/212-3p" "miR-132/212/212-3p" "miR-132/212/212-3p"
  [4] "miR-132/212/212-3p" "miR-132/212/212-3p" "miR-7/7ab"
  [7] "miR-7/7ab"          "miR-7/7ab"          "miR-7/7ab"
[10] "miR-7/7ab"          "miR-205/205ab"      "miR-205/205ab"
[13] "miR-205/205ab"      "miR-205/205ab"      "miR-205/205ab"
[16] "miR-218/218a"       "miR-218/218a"       "miR-218/218a"
[19] "miR-218/218a"       "miR-218/218a"       "miR-125a-3p/1554"
[22] "miR-125a-3p/1554"   "miR-125a-3p/1554"   "miR-125a-3p/1554"
[25] "miR-125a-3p/1554"
notice the repetitions indicating multiple target sites within the UTR.

HTH,
J.