CAFE5 significant increases/decreases

[Theo Llewellyn]

Hi all,

I am using CAFE5 and have generated results for a tree with 37 taxa and ~5000 gene families. In the README it says the _model_\_family\_results.txt file should have the following format:

        #FamilyID   pvalue  *   orang   gibbon  chimp   human     
        0           0.436   n   d       d       c       c
        1           0.209   n   c       i       c       c
        2           0.002   y   c       c       i       i

My file only has the first three columns, is it supposed to have the others? I can filter this for significantly changing families (which gives me 66 families) but I'd like to know whether these are increasing or decreasing and at which nodes so I can plot them on my tree. The Base_clade_results.txt file I guess is showing all increases/decreases but I'm looking for just the significant ones.

Any ideas how I could do this?

Many Thanks,

Theo Llewellyn

PhD researcher, Imperial College London

[Dan Vanderpool]

Hello Theo,

Each family is already plotted on the tree.  That is what the .tre file is.  If you want to look at just the trees with significant changes you can filter this file as shown in the manual here with a couple of grep commands like this.

echo $'#nexus\nbegin trees;'>Significant_trees.tre
grep "*" Base_asr.tre >>Significant_trees.tre
echo "end;">>Significant_trees.tre

If you open this in Dendroscope or fig tree you can see which branches are significant (if you are viewing only one tree at a time you will need to go to menu View->Show Borders to see the family ID)

If you want the significant events by branch you can look at the branch probabilities file.  Similarly the model_change.tab file shows the change reconstructed for each branch giving the direction and magnitude (but all of this information is also on the trees).  We will likely add a summary tree that has significant increases and decreases plotted for each branch, in the future, currently we are trying to avoid having to use additional scripts after the run is complete and have not yet implemented this in C++.

dv   

--
You received this message because you are subscribed to the Google Groups "hahnlab-cafe" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hahnlabcafe...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hahnlabcafe/650e932b-945a-4553-a9eb-3a49c37c04bfn%40googlegroups.com.

__________________________

Dan Vanderpool
Postdoctoral Scholar
Department of Biology
Indiana University
Hahn Lab, Jordan Hall 249B
1001 East Third Street
Bloomington, IN 47405
Email: ddvand...@gmail.com

[Theo Llewellyn]

Hi Dan,

That's really helpful. Thanks for that information, with the combination of those file I can get what I need.

Best,

Theo