HyPhy message board | |
http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl
HYPHY Package >> HyPhy feedback >> Spidermonkey questions http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl?num=1345845681 Message started by Alex_Wong on Aug 24th, 2012 at 3:01pm |
Title: Spidermonkey questions Post by Alex_Wong on Aug 24th, 2012 at 3:01pm
Hello,
I am potentially interested in using Spidermonkey to look for co-evolution between proteins. Before I dive in too deeply, I have a few questions - (1) Are there any in-principle concerns with using Spidermonkey on multiple proteins instead of just one? Presumably I could input a concatenated alignment to do so? (2) Is there a minimum number of taxa that are recommended to see a signal if there is one? (3) Again on the topic of power - given the choice between closely related taxa or more distantly related taxa, is there a preference? Any guidance would be much appreciated! Thanks, Alex Wong |
Title: Re: Spidermonkey questions Post by Sergei on Aug 28th, 2012 at 9:48am
Hi Alex,
In response to your queries: 1). No, so long as the proteins have the same gene trees. Spidermonkey assumes a single phylogeny for all sites. 2). Spidermonkey derives power from co-occuring substitutions, which are a function of the number of taxa and sequence divergence. I would say that you need at least 5 substitutions at a site to gain good power. 3). Basically you want to maximize the number of substitutions without losing homology. I would include a mix of distantly and closely related taxa -- the former give you power, the latter help map substitutions more accurately. Sergei |
HyPhy message board » Powered by YaBB 2.5.2! YaBB Forum Software © 2000-2024. All Rights Reserved. |