HyPhy message board - Print Page

HyPhy message board
http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl HYPHY Package >> HyPhy feedback >> Modeltest step by step? http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl?num=1162008862 Message started by qiang on Oct 27^th, 2006 at 9:14pm

Title: Modeltest step by step?
Post by qiang on Oct 27^th, 2006 at 9:14pm

I am a new user.I want implenment Modeltest in PY-PHY from multiple nucleotide sequence alignments.It maybe need a tree file,but I do not kown how to create it.
:-[It's so stupied,but I hope you give me some help.

Title: Re: Modeltest step by step?
Post by Sergei on Oct 27^th, 2006 at 10:03pm

Dear qiang,

HyPhy already has an implementation of ModelTest. You need to go to Standard Analyses (which is found under the 'Analysis' menu in GUI versions), select ModelComparison, and then choose ModelTest.bf

Cheers,
Sergei

Title: Re: Modeltest step by step?
Post by qiang on Oct 28^th, 2006 at 9:37pm

hi,seiger
thank you,i have got it.
but using the selected models , i want to implemente likelihood rate test for a significant deparcrture from the hypothesis of a global moleculer clock.How do I do?

thank you for your attention!

Title: Re: Modeltest step by step?
Post by Sergei on Oct 30^th, 2006 at 11:36am

Dear qiang,

StandardAnalyses>MolecularClock>MolecularClock.bf will do this.

Cheers,
Sergei

Title: Re: Modeltest step by step?
Post by qiang on Oct 31^st, 2006 at 6:03am

"-2(Ln Likelihood Ratio)=416.386
Constrained parameters:31
P-Value:0"

I am not good at statistics.How about this result? Is it reject moleculer clock assumption?

Thank you very much!!

Title: Re: Modeltest step by step?
Post by Sergei on Oct 31^st, 2006 at 6:09am

Dear qiang,

Yes. Low p-value = high significance of your result (i.e. you reject the null hypothesis, in this case molecular clock).

Cheers,
Sergei

Title: Re: Modeltest step by step?
Post by qiang on Nov 4^th, 2006 at 3:49am

seiger,thank you very much.I had worked out!