Morphological tree search

[Mark]

Hi,

I am attempting to estimate a phylogeny based on morphological data on IQ-tree. The analyses run but when I try independent runs the optimal tree is not converging on the same likelihood (i.e, some have better and some have worse final likelihoods). My input is shown here:

 iqtree -s morpho.phy -m MK+ASC+G -pers 0.1 -nstop 2000 -ntop 1000 -ninit 5000

Looking at the 'log' files it appears the searches find a local optimum at the start of the run and tend to stay there. I have altered the values of -pers from around (0.1 to 0.9) and increased the limit of -nstop and -ninit  for different runs, and then run multiple searches with identical values. 

Would anyone be able to offer any suggestions?

Thanks,

Mark

[Bui Quang Minh]

Dear Mark, 

This sounds like your alignment does not have enough phylogenetic signals … What is the size of your alignment (#sequences, #sites)? It’s difficult to analyze such data and the problem of local optima is quite well known (not specific to IQ-TREE but also other ML software).

In fact, we recommended to repeat several runs and what you did sounds good to me. Then you can take the tree with highest likelihood out of those runs. You can also do the tree topology tests for all resulting trees, and compute the strict consensus tree of trees that passed the test, which can be more stable.

If you have other sequences (e.g. ribosomal RNAs) for the same species set, then you can add them into the analysis, so that you have more phylogenetic information.

Hope that helps a little bit, 

Minh

-- 
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.
For more options, visit https://groups.google.com/d/optout.

--

Bui Quang Minh

Center for Integrative Bioinformatics Vienna (CIBIV)

Campus Vienna Biocenter 5, VBC5, Ebene 1

A-1030 Vienna, Austria

Phone: ++43 1 4277 74326

Email: minh.bui (AT) univie.ac.at