clarification on *-clade_results.txt

[Carmen Gibbs]

Hi!

I am seeking a bit of clarification on the *-clade_results.txt output file.  I assume that this is a tally of all of the families that have an increase or decrease for each node, regardless of the magnitude or significance of that change.  Am I correct?  (I am wondering if I can say that these are significant changes or not).

"If you open file results/base_clade_results.txt, you will see, for each branch, how many families underwent expansions and contractions."

Thanks!

Carmen

[Hahn, Matthew]

Hi Carmen,

Yes, that is all families that have changed, regardless of the size of the change. There is no such thing as a “significant” change in a family in CAFE, though there can be rapidly evolving families that have significantly different rates on a branch.

Hope that helps,

Matt

--
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/1857f6a4-fc15-483c-a4fa-cc49db7cc24cn%40googlegroups.com.

[Carmen Gibbs]

Thanks for the response Matt!

I'm trying to properly represent the CAFE5 analysis that I recently performed on my data and this helps.

I think where I also have my brain in a knot is with the CafePlotter result file, which I recognize is a third party.  The result_summary.tsv file produced lists the significant families and gives a count, change, and p-value for each family and node.  Perhaps those p-values indicate the significantly different rates on a branch.

Carmen

[Carmen Gibbs]

Hi Matt,

Sorry, a follow-up clarification question regarding model_branch_probabilities.txt

So the probabilities in model_branch_probabilities.txt cannot be used to describe significant contractions or expansions then (as I have seen others do) but instead must be described as branches with significantly different evolutionary rates from the overall evolutionary rate for that family.

Am I right?

Thanks,

Carmen

"model_branch_probabilities.txt

Contains a tab-separated list of the probabilities calculated for each clade and significant family. Probabilities are displayed as "N/A" if the parent and child have the same value. In the case of the Gamma model, only contains significant families that are calculated to be rapidly changing."

[Hahn, Matthew]

Yes, exactly. If a family has a significantly different rate across the whole tree, this output tries to identify which branch or branches are most responsible. Nothing to do with significant contractions or expansions!

cheers,

Matt

To view this discussion on the web visit https://groups.google.com/d/msgid/hahnlabcafe/6b9746b9-e5f2-4324-a8f6-669a77dbb90bn%40googlegroups.com.

[Carmen Gibbs]

Ok thanks!