[CANAPE] About neo and paleo endemism centres

[Léo-Paul Dagallier]

Dear Biodiverse users,

I have interrogations concerning the CANAPE analysis.

As explained in the CANAPE papers, a centre of neo-endemism indicates an over-representation of range-restricted short branches, and a centre of paleo-endemism indicates an over-representation of range-restricted long branches.

If I understood correctly the underlying procedure, a "short" branche is lower than the theoretical tree branches, and a "long" branche is greater than the theoretical tree banches. Thus the categorisation between neo and paleo relies on the branches length of this theoretical tree (that are all equal). The branch length on the theoretical tree appears to be the mean of the branch length of the original tree. In my case the original tree is dated and the mean branch length is 30 My.

So, my interrogations are:

- Can we extrapolate the CANAPE results as follow: "centres of neo-/paleo-endemism are over-representations of range-restricted taxa that are younger/older than 30 My" ?

- If so, would there be a way to constrain the branch length of the theoretical tree ? for example to test for over-representations of range-restricted taxa that pre- or post-date any geological event (e.g.: mountain uplift, sea closure, etc.).

Thanks a lot !

Léo-Paul Dagallier

[Shawn Laffan]

Hello Léo-Paul,

The comparison tree does use the average branch length, although the implementation in Biodiverse keeps zero length branches as zero, so it is actually the average of the non-zero length branches.  In most cases a tree will not have such branches, though.

Your interpretation is correct for RPD (relative phylogenetic diversity), which is the ratio of PD for the observed tree to PD for the comparison tree.  It perhaps needs more thought, but it should be possible to apply the threshold logic from CANAPE to the PD_obs, PD_alt and RPD scores to identify regions with branches that are longer than the expected 30 My. 

The reason it will not apply to CANAPE is that for phylogenetic endemism (PE) the branch lengths are downweighted by their geographic ranges.  For example, a relatively long branch length (e.g. 60 My) might be widespread.  If it is found in 10 cells then its range weighted length will be 60/10=6.  The same branch on the alternate tree will be given a weighted length of 30/10=3.  However, if another branch of length 120 is found in 20 cells then its weighted length will also be 6, although its weighted length on the alternate tree will be 1.5.

One approach that could be used to assess the branch ages without correcting for ranges is to use the CANAPE regions to define the analysis windows, and then explore the branch lengths of the observed tree more directly using some of the PD branch and clade summary indices. 
https://github.com/shawnlaffan/biodiverse/wiki/Indices#phylogenetic-diversity-node-list
https://github.com/shawnlaffan/biodiverse/wiki/Indices#pd-clade-contributions

There are two ways to do this.

The first is to use a GIS to generate polygons of the CANAPE regions, and then use these polygons to define the neighbourhoods for a spatial analysis. 
https://github.com/shawnlaffan/biodiverse/wiki/SpatialConditions#sp_points_in_same_poly_shape  (this documentation needs a lot of work)

The second approach is to cluster the CANAPE regions, and set the system to calculate the indices I listed above for each node.  This would involve finding the relevant branch on the cluster dendrogram that relates to a CANAPE region, which might be difficult if your study region is large and complex. 
More details of the cluster analysis are at http://biodiverse-analysis-software.blogspot.com/2016/04/more-canape-how-to-restrict-your.html

That blog post describes a means to use analysis results to define neighbourhoods, but it does not provide a means to define non-contiguous regions.  In other words, it will treat all cells that pass the threshold as a single set, regardless of whether they are distinct regions or not. 


The above approaches might help identify regions with branches that pre or post date some event, but I am not certain about that.  If your question does not need the RPD/RPE scores, then you could also generate your own tree and use that in a post-CANAPE spatial or cluster analysis.  i.e., run CANAPE using your original tree, and then calculate the PD scores for the CANAPE regions using some other tree.  Biodiverse will use whichever tree is selected at the top of the window when it runs its analyses, so there is a high degree of flexibility there.

You can also visualise several of the clade indices on the tree, so if you know which branches are interesting then you can see what their contributions are to the analysis for some neighbourhood.  http://biodiverse-analysis-software.blogspot.com/2017/09/visualise-spatial-analysis-results-on.html


Hopefully that all helps.  I am sure there are points that I have missed and/or that need further clarification, so please do ask any follow-up questions. 


Regards,
Shawn.

--
You received this message because you are subscribed to the Google Groups "Biodiverse Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biodiverse-use...@googlegroups.com.
To post to this group, send email to biodiver...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/biodiverse-users/c64d5e92-1da1-48b5-8325-a49a9999bb3c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

-- 
Assoc Prof Shawn Laffan
  School of Biological, Earth and Environmental Sciences
  UNSW, Sydney 2052, Australia
  Tel +61 2 9385 8093
  http://www.bees.unsw.edu.au/staff/shawn-laffan
  https://shawnlaffan.github.io/biodiverse (free diversity analysis software)
  
  Asia-Pacific Editor
  International Journal of Geographical Information Science
  http://www.tandf.co.uk/journals/ijgis

  UNSW CRICOS Provider Code 00098G

[Léo-Paul Dagallier]

Hello Shawn,

Thanks a lot for these clarifications and suggestions ! That indeed helps a lot !
I will explore what you propose and I'll let you know if any question or suggestion comes to my mind.

Thanks a lot again,
Regards,
Léo-Paul

[Shawn Laffan]

Hello Paul,

You are welcome, and I am glad it has been of help.

Regards,
Shawn.

To view this discussion on the web visit https://groups.google.com/d/msgid/biodiverse-users/a8f552ad-6b23-40e0-817d-f958248e19fc%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.