[BiO BB] promoter sequence analysis for TFBS .. using phylogeny

Samantha Fox bioinfosm at gmail.com
Fri Oct 28 17:08:10 EDT 2005


Hi Thomas,
Sorry for the delay in responding, but thanks much for the reply.

Well I have 2000 sets of orthologous sequences for these 8 species. And
their phylogenetic distance is quite well known.
So actually it would be more useful to use the phylogenetic information, and
then perform alignments to pick out Transcription factor binding sites, or
be more confident of finding TFBS, etc.

I hope to have explained it better .. would you have any suggestions in this
regard !

Cheers,
Samantha


On 10/22/05, Thomas Keane <thomas.m.keane at nuim.ie> wrote:
>
> I'd advise downloading your own copy of phyml and running it on your
> computer - there is a link to download a copy somewhere on the website.
> You should note that phyml only recognises phylip format files. Its
> basically software for building phylogenetic trees from a group of
> alignments.
>
> I havent used MEME - what we do in our lab is run Clustalw to align the
> sequences then GBlocks to remove any badly aligned areas
> (http://molevol.ibmb.csic.es/Gblocks/Gblocks.html) then modelgenerator
> to get the ML model then phyml to get our bootstrapped trees :-)
>
> Thomas
>
> Samantha Fox wrote:
>
> > Dear Thomas,
> > Thanks for the response. I could use the ModelGenerator, but PHYML
> > online execution gave some error. I will also have to read their
> > paper, to see what exactly the software does .... and what I should
> > expect as output.
> >
> > One simple thing I did was run MEME on each ortholog set, and the
> > blocks come out nicely conserved ... sequentially as well. However I
> > was not sure of what next, and didnt go further.
> > Another task was to run clustalw on each ortholog set separately and
> > see what sequences come out conserved in majority sets, etc.. Most
> > results again were repitition of known .. so not really interesting ...
> >
> > Any comments on this ....... or some other ideas ???
> >
> > Cheers ...
> > Samantha
> >
> > On 10/21/05, *Thomas Keane* <thomas.m.keane at nuim.ie
> > <mailto:thomas.m.keane at nuim.ie>> wrote:
> >
> > If you want to use maximum likelihood then I would suggest that
> > you use
> > Phyml (http://atgc.lirmm.fr/phyml/) - you can download your own copy.
> > You will also need to find the optimal ML model first - you could use
> > Modelgenerator (http://bioinf.nuim.ie/software/modelgenerator) to do
> > this as it creates scripts to start Phyml with the optimal model.
> >
> > Thomas
> >
> > Samantha Fox wrote:
> >
> > > Hii...
> > > I am looking for various options to analyze a set of few hundered bp
> > > long orthologous sequences from a group of phylogenetically related
> > > species. Its like 5-8 homologous promoter dna sequences per gene
> > for
> > > around thousand genes. The motivation is to get the conserved motifs
> > > which have remained constant under selection.
> > >
> > > What can be the best ways to do this analysis, I looked at some
> > > phylogenetic tools, but most have limitations like web-based,
> > just 2
> > > sequences, etc.
> > >
> > > All your suggestions... and anyone experienced with similar
> > stuff. ...
> > > I welcome it all...
> > >
> > > Thanks.
> > > Samantha
> > >
> >
> >------------------------------------------------------------------------
> >
> > >
> > >_______________________________________________
> > >Bioinformatics.Org general
> > forum - BiO_Bulletin_Board at bioinformatics.org
> > <mailto:BiO_Bulletin_Board at bioinformatics.org>
> > > https://bioinformatics.org/mailman/listinfo/bio_bulletin_board
> > >
> > >
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.bioinformatics.org/pipermail/bbb/attachments/20051028/94a1cbda/attachment.html>


More information about the BBB mailing list