Re: mapnh-counts/rates

[Julien Yann Dutheil]

Dear Paulina,

Sorry for the late answer. The normalization of counts is optional, and only makes sense in the case where the parameters of interest are dependent on the counts AND the acestral frequencies (eg, theta parameter in a T92 model, but not the kappa = transition / transversion parameter). This normalization is used when performing a clustering of branches, for instance for using with partnh. In all cases, only the counts are output, not the normalized ones (but Laurent correct me if I'm wrong).

cheers,

Julien.

On Thu, Jun 5, 2014 at 3:23 PM, Pau Bolivar <paubo...@gmail.com> wrote:

Dear Julien

I want to implement the L95 (Lobry 1995) non-reversible model to calculate different kind of substitution rates (strong-to weak, weak-to-strong, strong-to-strong and weak-to-weak) per branch. I understand that mapNH can count different kinds of substitutions with map.type=All or GC.

In the manual it says that for GC in case of a non-stationary models the counts are normalized by the frequencies of the ancestral node. Does this mean that the counts of A or T to G or C and viceversa are directly comparable and already corrected by the ancestral sequence GC content?.  If this is correct, Is this the case for map.type all as well? (I am also interested in looking at rates of strong-to-strong and weak-to-weak).  

I want to calculate the different classes of substitution rates per site for each branch  (not just total counts per branch) so I was thinking of getting these rates using the number of substitution counts and number of possible AT to GC (and viceversa) substitutions. However if this counts are already corrected by that then the rates per sites can be calculated with the total number of sites? Does this make sense or am I completely lost here...

Thanks for your time and help! 

Best regards,

Paulina

--

---
You received this message because you are subscribed to the Google Groups "testnh-help-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to testnh-help-fo...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
Julien Y. Dutheil, Ph-D
0 (+49) 6421 178 986

§ Max Planck Institute for Terrestrial Microbiology
Department of Organismic Interactions
Marburg -- GERMANY

§ Intitute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE

[Laurent Guéguen]

Dear Paulina & Julien,

yes indeed, the usual output of mapnh does not normalize the counts, so the output is exactly the expected number of subsitutions on the branches.

I have implemented an option to normalize the counts as I think you intend to do, but it is not yet approved by peers. Moreover, it is not yet implemented for non-reversible models,
but it would not be difficult for me to do it (just wait July, I hope).

CHeers,
Laurent