Is it a limitation of TestNH?

[Giap Nguyen]

Dear all,

 

From recent posts, I've learnt that Mapnh has an option that allows very short branchs cluster with their parents. As I understand, that option equals to Dn & Ds could not be counted, for example. Consequently, it is impossible to estimate omega values in some nodes of tree? A small piece is given:

 

A3937_Italy_2003:0.1212121212):Inf):0.1):0):0.05555555556,((((((A3990_Italy_2004:0.1111111111,A9962_Italy_2004:0.05555555556):0,(A9991_Italy_2003:0,A9989_Italy_2004:0):0):NaN

 

Would you please tell give me advice on my situation above? And what should I do in case I still want to estimate omega values for all nodes?

 

Thank you very much

Best regards

Giap Nguyen

 

[Julien Yann Dutheil]

Hi Nguyen,

The "auto" clustering is for when Dn and Ds are two small. It is not
that they cannot be counted: they are counted as zero or a small
number, which means that the corresponding omega value has a large
uncertainty, or is undefined (division by zero). Therefore Partnh will
not try to estimate an omega value for these branch only, but rather
will estimate a joint value with the parent node.

For you situation, please describe what is the tree you provide.

Best,

Julien.

> --
>
>
>



--
Julien Y. Dutheil, Ph-D
0 (+49) 6421 178 530

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

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