[Mesquitelist] Testing phylogeographical hypotheses using coalescent simulations

[Wayne Maddison]

Could you send me the file saved so that it is doing the analysis?  I think I'll need to see it to understand what is going wrong.  
Wayne

On 14 Jun 2011, at 4:49 AM, Mark Ravinet wrote:

> Dear Mesquite users,
> 
> I am somewhat new to Mesquite and I would very much appreciate your 
> input regarding my issues with attempting to follow the hypothesis 
> testing methodology of Knowles (2001, Mol Ecol) and Knowles & Maddison 
> (2002, Mol Ecol). I understand the theory behind the approach but 
> unfortunately, no one else in my department has used Mesquite for this 
> purpose and I am unsure of whether I am conducting the analysis correctly.
> 
> Essentially I am attempting to infer the number of refugia involved in 
> colonization of Northern Europe by three-spine stickleback using my 
> mitochondrial dataset for the British Isles and additionally, previously 
> published sequences for the rest of Europe.
> 
> For a simple case, I have three hypotheses - i.e. one refugia 
> (fragmentation, the null hypothesis), two refugia and three refugia. 
> After Knowles, I have attempted to model 1000 genealogies constrained 
> within the null hypothesis and then I have attempted to constrain this 
> to one of  the alternative hypothesis models to generate an expected 
> distribution of discordance.
> 
> To do this, I have one file with a taxa block for each hypothesis and 
> one main taxa block for the mtDNA haplotypes. Then I have set up 
> associations between each hypothesis and the haplotypes before using 
> coalescent simulations to make a new tree block of 1000 genealogies 
> constrained within the displayed hypothesis tree.
> 
> The issue comes when I try to use Mesquite to calculate the discordance 
> between the simulated gene trees from the null hypothesis and the 
> alternative hypothesis. To do this, I have been using the Analysis > 
> charts > trees commands, selecting the haplotypes taxa block for the 
> tree values and then choosing stored trees > null hypothesis to generate 
> a distribution of Slatkin's S. In order to generate the distribution, I 
> use the association for the displayed hypothesis tree.
> 
> My problem is that whenever I do this, I seem to get exactly the same 
> distribution of discordance for when the genealogies are constrained 
> under the null tree - i.e. there is no difference between genealogical 
> discordance within the null tree and between it and the alternative 
> hypothesis tree. At first I thought this might be a problem with my 
> data, but it also occurs when I use a 'model' dataset I concocted to 
> test the problem.
> 
> I must apologise for the long email, but this is becoming a really very 
> frustrating situation. If anyone could shed any light on what I might be 
> doing wrong, or even if anyone has a simple protocol for running these 
> type of analyses I might be able to follow I would very much appreciate it.
> 
> Many thanks in advance
> 
> Mark Ravinet
> 
> -- 
> ------------------------------
> Mark Ravinet
> PhD Student
> School of Biological Sciences
> Queen's University Belfast
> Medical Biology Centre
> Lisburn Road
> Belfast
> BT9 7BL
> Tel:02890972103
> Mob: 07841675863
> Email: mravinet01 at qub.ac.uk
> ------------------------------
> 
> _______________________________________________
> Mesquitelist mailing list
> Mesquitelist at mesquiteproject.org
> http://mesquiteproject.org/mailman/listinfo/mesquitelist

--------------
Wayne Maddison
    Professor and Canada Research Chair, Depts. of Zoology and Botany
    & Co-Director, Beaty Biodiversity Museum
6270 University Boulevard
University of British Columbia
Vancouver, BC  V6T 1Z4   Canada

Home page: http://salticidae.org/wpm/home.html
Mesquite: http://mesquiteproject.org; MacClade: http://macclade.org
Beaty Biodiversity Museum: http://beatymuseum.ubc.ca