miRNA target predictions

[german nudelman]

Dear MirMaid developers and its lucky users!

MirMaid RESTful services seem to very handy and in obtaining info
about miRNAs, but it seems to me that being able to retrieve also
predicted targets for miRNA via a web service would be an enormously
useful feature.

I wonder whether such a feature is planed.

Could anybody suggest what would be best way to go around being able
to retrieve predicted targets for miRNA via a web service/API as of
today.

best,
German Nudelman
Sr. Bioinformatics developer,
Mount Sinai Medical Center, NYC

[Raoul B.]

Actually it depends on which prediction algorithm do you use.
I wrote a small rails application for targetscan, but never integrated
into mirmaid. Probbably the effort to integrate it into mirmaid is
trivial.

Do you guys have tested mirmaid with ruby 1.9.2 and rails 2.3.8 ?

[Morten Lindow]

On Mon, Sep 27, 2010 at 7:40 PM, german nudelman <german....@gmail.com> wrote:

predicted targets for miRNA via a web service would be an enormously
useful feature.

I agree and it has been on the wish list since the start. The reason it has not been done yet is that it is not straightforward how to beautifully model target-mRNA: e.g. should we use refseq-sequences or ensembl-transcripts? which variants?, how should conservation and orthology be modeled? Can we make a general target prediction modelling that allows for different algorithms that use different target sequences? 

That being said, I know Anders is working on something.  Have you found a pragmatic solution Anders?
 

Could anybody suggest what would be best way to go around being able
to retrieve predicted targets for miRNA via a web service/API as of
today.

Ensembl had something at one point, but I haven't checked if they still have in the more recent assemblies.

--
Morten Lindow, Ph.d., Group Leader
Bioinformatics, Santaris Pharma A/S
office: +45 4517 9867; mobile: +45 2849 9682
m...@santaris.com | morten...@gmail.com
microRNA information for computers: http://www.mirmaid.org

[Anders Jacobsen]

Hi German,

There is no target prediction plugin, yet. It is probably the plugin
that is highest on our wish-list, but as Morten pointed out, there are
some important design issues in order to make the plugin useful and
scalable. Ideally we would like to support multiple organisms,
prediction methods and a unified sequence set.

Raoul: We have not yet tested with Ruby >=1.9 or Rails 2.3.8. I think
we will make an effort in a few months to bring miRMaid in-sync with
Ruby 1.9.2 and Rails 3.

-Anders

> --
> You received this message because you are subscribed to the Google Groups
> "MirMaid" group.
> To post to this group, send email to mir...@googlegroups.com.
> To unsubscribe from this group, send email to
> mirmaid+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/mirmaid?hl=en.
>

[Raoul B.]

Anders,

On 29 Set, 20:36, Anders Jacobsen <anders...@gmail.com> wrote:
>
> Raoul: We have not yet tested with Ruby >=1.9 or Rails 2.3.8. I think
> we will make an effort in a few months to bring miRMaid in-sync with
> Ruby 1.9.2 and Rails 3.

I have a lot of interest on this topic.

about the prediction plugin, it seems to me that you are trying to
generalize as much as possible, unfortunately this process requires
spending a lot of time in design.
What about provide a quick plugin for targetscan for instance?


>
> -AndersOn Tue, Sep 28, 2010 at 2:53 AM, Morten Lindow <morten.lin...@gmail.com> wrote:
>
> > On Mon, Sep 27, 2010 at 7:40 PM, german nudelman <german.nudel...@gmail.com>

> > wrote:
>
> >> predicted targets for miRNA via a web service would be an enormously
> >> useful feature.
>
> > I agree and it has been on the wish list since the start. The reason it has
> > not been done yet is that it is not straightforward how to beautifully model
> > target-mRNA: e.g. should we use refseq-sequences or ensembl-transcripts?
> > which variants?, how should conservation and orthology be modeled? Can we
> > make a general target prediction modelling that allows for different
> > algorithms that use different target sequences?
>
> > That being said, I know Anders is working on something.  Have you found a
> > pragmatic solution Anders?
>
> >> Could anybody suggest what would be best way to go around being able
> >> to retrieve predicted targets for miRNA via a web service/API as of
> >> today.
>
> > Ensembl had something at one point, but I haven't checked if they still have
> > in the more recent assemblies.
>
> > --
> > Morten Lindow, Ph.d., Group Leader
> > Bioinformatics, Santaris Pharma A/S
> > office: +45 4517 9867; mobile: +45 2849 9682

> > m...@santaris.com | morten.lin...@gmail.com

[Anders Jacobsen]

Hi Raoul,

Your point is well taken. Maybe we could talk about the work you
already did on TargetScan and get something simple working quickly.

-Anders