-loglik 0.0

[AT]

Hi Claudia and Cécile,

I used SNaQ in my analusis. The RAxML gene tree based on the 1011 genes was used as an input and was summarized by a list of quartet concordance factors. An ASTRAL tree based on 1011 genes was implemented as a starting topology for the optimization (hmax = 0). I used 10 independent runs and hmax values ranging from 0 to 5. 

For hmax = 4 and 5, I get -loglik 0.0. (I check the out files). But I haven't gotten any errors.

Do you think I do something wrong? Could you please give me any advice?

Thank you in advance, 

Alicia 

 

[Cécile Ané]

A score of 0.0 means that your estimated network fit your data perfectly, because the score saved as -loglik is normalized by the best score a network could ever have. This best scenario is when the quartet concordance factors expected from the network match perfectly those observed in the data. A score of 0.0 is something we obtained on quartet CF data that were manually constructed from a network (as in a simulation, but as if with an infinite number of genes), to check that SNaQ was able to recover that network. But with a finite number of genes, a perfect score of 0.0, without any rounding, is indeed suspicious. I encourage you to look further into how the concordance factors match between the data and what is expected from your estimated network.

Cécile.

[Alicia Talavera]

Hi Cécile, 

Thank you so much for your advice. I will check it carefully. 

Best wishes, 

Alicia 

--
You received this message because you are subscribed to a topic in the Google Groups "PhyloNetworks users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/phylonetworks-users/AdAEtm-ofgs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to phylonetworks-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/phylonetworks-users/52aaa89d-2f9d-49db-9fac-205b036c9fdbn%40googlegroups.com.

--

Alicia Talavera Júdez, Margarita Salas Postdoctoral Fellow

Smithsonian National Museum of Natural History

University of Malaga 

Tala...@si.edu