Groups keyboard shortcuts have been updated
Dismiss
See shortcuts
Question on Turnover analysis
30 views
Skip to first unread message
Gabriela Procópio Camacho
Apr 4, 2023, 5:13:10 PM4/4/23
to biodiver...@googlegroups.com
Dear Shawn,
I am trying to run a factorial exploratory analysis on the PhyloRWT results, to try and determine a good cutoff for the number of clusters. However, I haven't been able to generate an output from the cluster analysis that I could input to the factorial exploratory analysis. For other work, I would use a raster with the turnover values for each cell, but the only file from the cluster analysis is a matrix. Could you point me in the right direction?
Thank you in advance.
Cheers,
Gabi
Gabriela P. Camacho, Ph.D.
Professora e Curadora de Hymenoptera - MZUSP
Pronomes: ela/dela
T +55 11 2065-8100
gabieco.c...@gmail.com
Museu de Zoologia da Universidade de São Paulo
Av. Nazaré, 481 - Ipiranga
04263-000 São Paulo - SP
Brasil
Shawn Laffan
Apr 4, 2023, 6:42:26 PM4/4/23
to biodiver...@googlegroups.com, Gabriela Procópio Camacho
Hello Gabri,
A good question. I have a few comments/responses and requests for clarification.
Could you please send a link to the factorial exploratory analysis method? It would be useful to see how it works.
Regarding your raster with turnover values, is it from a turnover matrix or is it a moving window analysis? The latter can be generated in Biodiverse but the approach is very different to the turnover used in the cluster analyses.
If you need to reduce the matrix then you can use NMDS or similar to identify the main components. That will still give you a multi-layer data set, but you might only need the top three or so layers. An example is in Fig 4 in González-Orozco et al. (2014a, https://onlinelibrary.wiley.com/doi/10.1111/ddi.12129 ).
As for the cutoff, it always depends what one is interested in. Each region will also have subregions. These might be just as interesting and will probably have different cutoff values on the tree. An example is in González-Orozco et al. (2014b, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0092558 ) and also González-Orozco et al. (2014a, Fig 3).
Another possible approach is to optimise the value of an index calculated for each node in the tree. Care needs to be taken when choosing the index, though, as indices like PE and WE are additive so will always increase as one approaches the root of the tree. Possibly one could also look at the rate of change of the indices of each node relative to those of the parent (relative or absolute as appropriate).
Regards,
Shawn.
A good question. I have a few comments/responses and requests for clarification.
Could you please send a link to the factorial exploratory analysis method? It would be useful to see how it works.
Regarding your raster with turnover values, is it from a turnover matrix or is it a moving window analysis? The latter can be generated in Biodiverse but the approach is very different to the turnover used in the cluster analyses.
If you need to reduce the matrix then you can use NMDS or similar to identify the main components. That will still give you a multi-layer data set, but you might only need the top three or so layers. An example is in Fig 4 in González-Orozco et al. (2014a, https://onlinelibrary.wiley.com/doi/10.1111/ddi.12129 ).
As for the cutoff, it always depends what one is interested in. Each region will also have subregions. These might be just as interesting and will probably have different cutoff values on the tree. An example is in González-Orozco et al. (2014b, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0092558 ) and also González-Orozco et al. (2014a, Fig 3).
Another possible approach is to optimise the value of an index calculated for each node in the tree. Care needs to be taken when choosing the index, though, as indices like PE and WE are additive so will always increase as one approaches the root of the tree. Possibly one could also look at the rate of change of the indices of each node relative to those of the parent (relative or absolute as appropriate).
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 view this discussion on the web visit https://groups.google.com/d/msgid/biodiverse-users/CA%2B%3DicGJ7ddYk9i9jLQvOp_d212JGwEykwo%2BqND_U9tZhNfaADA%40mail.gmail.com.
Reply all
Reply to author
Forward
0 new messages