<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Dear Ensembl Development community,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
I previously asked the Ensembl Helpdesk the following question</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<blockquote style="margin-top: 0px; margin-bottom: 0px;">
<div>
<div>
<div>
<div dir="ltr">
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
Hello Ensembl help desk -</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
<br>
</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
Thank you for your essential resource.</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
<br>
</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
I would like to study vertebrate paralogues arising from the original two whole genome duplications in the ancestral vertebrate ancestor, as well as related research questions.</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
However, while Ensembl has supertrees where each paralogue is found in a separate genetree, it is impossible to know how the genetrees are related - </div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
the supertree is not a real tree. </div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
<br>
</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
What is the Ensembl recommendation for turning supertrees into real trees? Has Ensembl already done this, or know of and recommend an outside effort/s?</div>
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
What do you think about only taking genes from the ten species with best quality genomes in the supertree, and then running the protein tree algorithm on these ten percent of the total sequences - maybe this way the number will not be above 400 sequences, and
the computation will be practical to accomplish; Or, maybe I could make a consensus sequence for each genetree, and then run the protein tree algorithm on the consensus sequences?</div>
<br>
</div>
</div>
</div>
</div>
</blockquote>
<div>
<div>
<div>
<div dir="ltr">
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
and got the following answer</div>
<blockquote style="margin-top: 0px; margin-bottom: 0px;">
<div style="color:black; font-size:12pt; font-family:Calibri,Arial,Helvetica,sans-serif">
<font size="2"><span style="font-size:11pt">I have discussed this with our developers and they advise the following.<br>
<br>
Some of our "supertrees" are indeed flat.<br>
<br>
There are many ways of building supertrees. Both ways that you suggest are<br>
possible, but we cannot advise on what would be better. There are other<br>
methods based on sequence concatenation, matrices, etc. We cannot give a<br>
recommendation on the best approach.<br>
<br>
Please note that our tree-size limit is 1,500 genes, not 400.<br>
<br>
Finally, the most important limitation for us is how much computation we can<br>
afford during our production time in the Ensembl release cycle.<br>
<br>
We use TreeBest, which is not multi-threaded, but there are alternatives such<br>
as RAxML, ExAML that can run on dozens / hundreds of CPUs. Mafft has options<br>
to deal with large multiple alignments, and there are other tools such as<br>
PASTA. You may be able to build the tree and the alignment without using<br>
supertree methods.</span></font><br>
</div>
</blockquote>
</div>
</div>
</div>
</div>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
What is good about the response, and what would you further suggest?</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Sincerely,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Joseph (Yossi) Steinberger</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
</body>
</html>