HyPhy message board
http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl
HYPHY Package >> HyPhy feedback >> Defining a model
http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl?num=1095771502

Message started by Cor on Sep 21st, 2004 at 5:58am

Title: Defining a model
Post by Cor on Sep 21st, 2004 at 5:58am
Hi HyPhy Makers!
 I'm a new HyPhy user and find it interesting. I subjected my data to Modeltest analysis and now at a lost how to proceed with the analysis. The last part of the output says,  
To define this model for analyses, use 'custom' model with the following options:
Model string:012345
Model options: Rate variation, then choose gamma+Inv distribution
Equilibrium frequencies option: Observed.
The GUI example seem not applicable and I don't know  ??? how to arrive at the 'custom' model dialog box from the main menu to select the instructed options.
Wish the HyPhy manual would be further expanded to be user-friendly to those who have the least computer programming background like me.
More power,
Cor :)

Title: Re: Defining a model
Post by Sergei on Sep 21st, 2004 at 7:39am
In order to fit a custom nucleotide model to a nucleotide alignment, the following steps should be taken:

1). Invoke Standard Analyses (menu option under 'Analysis' for Mac/Windows or the list of analyses when HyPhy is launched from a command line).

2). Choose Basic Analyses->AnalyzeNucProtData.bf

3). Locate the file with your nucleotide alignment

4). Choose CUSTOM in the list of models that is presented next.

5). Fill in the options when prompted

Hope this helps,
Sergei

Title: Re: Defining a model
Post by Cor on Sep 24th, 2004 at 7:30am
Thanks for the very prompt reply, Sir. :)
 I defined the model alright but the ensuing output was not as I expected. It was worth trying the analysis though since it made me understand a little better how to use HyPhy.  I thought of using the recommended model (model string 012345 by Modeltest) with HyPhy's analyses tools via the console window. By invoking the Model window of the object inpector, my objective was almost realized. The on-line GUI example in defining a model indicates "t" ( branch length), and "c" (variable category for gamma). In this regard, I would like to know if "R" could represent transition/transversion ratio (instead of non-synonymous/ synonymous ratio). Is there a way to represent the invariable site in the model? Is it also labeled as variable category?
 All the best,
 Cor ;)

Title: Re: Defining a model
Post by Sergei on Sep 25th, 2004 at 9:52pm
While it is possible to define Gamma+Inv through the graphical model widget, a simple nucleotide fit is easier to perform via the Standard Analyses route. Did that not work for you?

Also, in order to use models defined in the graphical widget, one needs to perform the entire analysis in the graphical shell: look at the first few pages of Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login Login for a walk-through.


HyPhy message board » Powered by YaBB 2.5.2!
YaBB Forum Software © 2000-2024. All Rights Reserved.