rate variation in ML models

newsmgr at merrimack.edu newsmgr at merrimack.edu
Tue Oct 14 08:17:35 EST 1997


Relay-Version: ANU News - V6.2.0 06/23/97 OpenVMS AXP V6.2; site chasm
Path: chasm!cam-news-feed2.bbnplanet.com!cam-news-hub1.bbnplanet.com!news.bbnplanet.com!news-feed1.tiac.net!newsfeed.internetmci.com!194.162.162.196!newsfeed.nacamar.de!univ-lyon1.fr!news
Newsgroups: bionet.molbio.evolution
Subject: Re: rate variation in ML models
Message-ID: <61v0aj$aci at tempo.univ-lyon1.fr>
From: Nicolas_Galtier
Date: 14 Oct 1997 05:35:15 GMT
Reply-To: galtier at acnuc.univ-lyon1.fr
References: <61tkeo$bst at net.bio.net>
Distribution: world
Organization: Universite Claude Bernard - Lyon 1
Lines: 35
NNTP-Posting-Host: acnuc.univ-lyon1.fr

Nicolas Galtier wrote

>> I had reached somewhat different conclusions about rate variation among sites
>> in usual implementations of the ML method.

Joe Felsenstein replied:

> If you look at Yang's paper in Genetics in 1995:
>
> Yang-Z.
>    A space-time process model for the evolution of DNA sequences.
>       Genetics.  1995 Feb.  139(2).  P 993-1005.
>  
> you will see him using an autocorrelated model which is somewhat different
> from Gary's and my 1996 Hidden Markov Model approach (though it too is in
> effect a Hidden Markov Model method).  It does not assume i.i.d.


Now I see: I was refering to Yang's 1993 (MBE 10:1396) and 1994 (JME 39:306)
papers, not to the paper you cite above. A brief look at the PAML doc shows
that both algorithms are available. 

Therefore accounting for unequal rates among sites was performed either 
relaxing the independance assumption (Felsenstein Churchill 96, Yang 95) 
or not (C option in former DNAML, Yang 93-94). I wonder wether these
approaches have significantly different results, i.e. wether the
improvement of the tree-making methods mainly comes from assuming
unequal rates or from assuming non-independant sites (for common data sets, 
say rRNA).

Nicolas.







More information about the Mol-evol mailing list