unlink clock models?

[Ruhfel Brad]

Dear all,

I am curious about linking or unlinking the clock models in a partitioned analysis (partitioned by gene in my case, 3 plastid genes and one mitochodrial gene).

Is there some rule of thumb here? Or a statistic of some sort to decide whether or not it is appropriate to link or unlink the clock models?

I have been searching around the net to find something on this but haven't found anything yet.

Also, am i correct that if you unlink the trees that you simply get a topology for each partition (gene trees) and not one single species tree?

Many thanks in advance,
-Brad

[julien]

Unlinking clock models make sense only if you wish to infer a species
tree from independant (unlinked) genes tree... like nuclear VS
mitochondrial, plastid VS mitochondrial and so on...

So in you case, you may want to relax the mitochondrial clock with
respect to the clock used for your three plastid genes together...
Just one thing, this method, described and published as Star Beast (or
*Beast) use coalescent information from gene trees to sample the
posterior distribution of the species tree. Which more or less
directly implies to have at least a couple of individuals for each
taxon... The method also assumes no gene flow and, like classic
phylogenetic methods, no recombination.

See the original paper by Heled and Drummond here
http://mbe.oxfordjournals.org/content/27/3/570.abstract

J.

[Ruhfel Brad]

Hi Julian,

Thanks for the reply.

There is also the option to unlink the clock models and use a tree prior like "Speciation:Yule process". This is not a coalescent approach, and so you wouldn't need multiple individuals per species, correct?

Also it seems pretty clear that different genes in the chloroplast are evolving at different rates, if this is the case, wouldn't different clock models for the chloroplast genes make sense also?

Best regards,
-Brad

> --
> You received this message because you are subscribed to the Google Groups "beast-users" group.
> To post to this group, send email to beast...@googlegroups.com.
> To unsubscribe from this group, send email to beast-users...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/beast-users?hl=en.
>