<div dir="ltr">Hi Matthieu,<div><br></div><div>I am going to use the human self-alignment to detect paralogous genomic regions (particularly non coding regions). But I can not find examples of API for this purpose. Could you pass me some scripts or examples where I can start? Say I have a human genomic coordinate, and want to find its paralogous regions and alignments.</div><div>Many thanks.</div><div>Jinrui</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Mar 13, 2019 at 8:37 AM Matthieu Muffato <<a href="mailto:muffato@ebi.ac.uk">muffato@ebi.ac.uk</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div bgcolor="#FFFFFF">
<p>Hi Jinrui</p>
<p>In all our pairwise alignments, we refine the LastZ alignment
blocks with two steps called "chaining" and "netting" (see
<a class="gmail-m_-1501946497467315318moz-txt-link-freetext" href="http://europepmc.org/articles/PMC4852398" target="_blank">http://europepmc.org/articles/PMC4852398</a> and
<a class="gmail-m_-1501946497467315318moz-txt-link-freetext" href="http://genomewiki.ucsc.edu/index.php/Chains_Nets" target="_blank">http://genomewiki.ucsc.edu/index.php/Chains_Nets</a> for more
information). What you get in our database is the product of these
two steps.<br>
The netting phase is done on the reference species only, we don't
do bidirectional netting. This means that there is very little
overlap / nesting on the reference species (human in the case of
the human vs * alignments). Overlap / nesting is allowed on the
non-reference species, though. For instance, in the human-mouse
alignments, there are 20,000 pairs of blocks that overlap on
human, and 1,900,000 pairs of blocks that overlap on mouse.<br>
</p>
<p>So in this case, yes you can identify human paralogous regions 1)
through the self-alignment and 2) through the human-mouse
alignment (or any pairwise alignment that involves human) by
finding human regions that align to the same region in the other
species</p>
<p>Hope this helps,</p>
<p>Matthieu<br>
</p>
<div class="gmail-m_-1501946497467315318moz-cite-prefix">On 11/03/2019 19:45, Jin-Rui Xu wrote:<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div>Hi Matthieu,</div>
<div><br>
</div>
<div>Thank you very much for your email. </div>
<div><br>
</div>
<div>I am wondering in the human self alignment, one genomic
region may be mapped to multiple other regions. These multiple
hits also exist in e.g. human vs mouse genome alignment.</div>
<div>Does ensembl provide all these multiple regions or just the
best one? Can these multiple hits achieved by compara perl
API?</div>
<div><br>
</div>
<div>Thanks!</div>
<div>Jinrui</div>
<div><br>
</div>
<div><br>
</div>
<div> </div>
<div><br>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Mon, Mar 11, 2019 at 3:05
PM Matthieu Muffato <<a href="mailto:muffato@ebi.ac.uk" target="_blank">muffato@ebi.ac.uk</a>> wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Dear Jinrui,<br>
<br>
We have a human self-alignment, that has been computed with
LastZ and <br>
identifies paralogous regions within the genome. You can
find the whole <br>
alignment on the FTP <br>
<a href="ftp://ftp.ensembl.org/pub/current_maf/ensembl-compara/pairwise_alignments/" rel="noreferrer" target="_blank">ftp://ftp.ensembl.org/pub/current_maf/ensembl-compara/pairwise_alignments/</a>
<br>
but also query specific regions: <br>
<a href="http://rest.ensembl.org/alignment/region/homo_sapiens/17:63997797-64000390:1?species_set=homo_sapiens;content-type=application/json;method=LASTZ_NET" rel="noreferrer" target="_blank">http://rest.ensembl.org/alignment/region/homo_sapiens/17:63997797-64000390:1?species_set=homo_sapiens;content-type=application/json;method=LASTZ_NET</a><br>
<br>
Human is the only species for which we have a
self-alignment.<br>
<br>
Kind regards,<br>
Matthieu<br>
<br>
On 09/03/2019 03:10, Jin-Rui Xu wrote:<br>
> Hello,<br>
><br>
> I just started learning the compara API. However, I am
still not sure <br>
> whether it can address my questions. I am wondering if
someone could <br>
> give me some guidance and example scripts. Here is my
question: (1) I <br>
> want to identify all paralogous DNA fragments (not
neccessarily genes) <br>
> in a genome. One genomic regions may have more than one
duplicate. (2) <br>
> Then, I want to find in which of the other species, the
two paralogous <br>
> DNAs have a common ancestor.<br>
> Alternatively, I can focus on two genomic regions in a
genome to test <br>
> if they are paralogous, and then which species has
their common <br>
> ancestral DNA<br>
> How could I get this done using compara API (version
95)?<br>
><br>
> Many thanks!<br>
><br>
> Jinrui<br>
<br>
-- <br>
Matthieu Muffato, Ph.D.<br>
Ensembl Compara and TreeFam Project Leader<br>
European Bioinformatics Institute (EMBL-EBI)<br>
European Molecular Biology Laboratory<br>
Wellcome Trust Genome Campus, Hinxton<br>
Cambridge, CB10 1SD, United Kingdom<br>
Room A3-145<br>
Phone + 44 (0) 1223 49 4631<br>
Fax + 44 (0) 1223 49 4468<br>
<br>
</blockquote>
</div>
</div>
</blockquote>
<pre class="gmail-m_-1501946497467315318moz-signature" cols="72">--
Matthieu Muffato, Ph.D.
Ensembl Compara and TreeFam Project Leader
European Bioinformatics Institute (EMBL-EBI)
European Molecular Biology Laboratory
Wellcome Trust Genome Campus, Hinxton
Cambridge, CB10 1SD, United Kingdom
Room A3-145
Phone + 44 (0) 1223 49 4631
Fax + 44 (0) 1223 49 4468</pre>
</div>
</blockquote></div>