controlling the species displayed on a biplot

An embedded and charset-unspecified text was scrubbed...
Name: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-ecology/attachments/20110521/fb02c67f/attachment.pl>
Hi All,

I have done a IndVal analysis on a dataset which has identified 27 species
out of 57 as being significant arbiters of site differences. What I want to
do is a PCA biplot (data is hellinger transformed prior to this) with only
these 27 species on the ordination, not all 57, How can I control the
species I want displayed?

cheers

Andy

Andrew,

    We need a little more information.  Which IndVal function did you 
use?  It should return a vector of p-values you can use as a mask [p-val 
<= 0.05,] on your taxon matrix.  But are you saying you want to use all 
57 in constructing the PCA and only identify 27, or that you only want 
to use those 27 in the construction of the PCA?

Dave Roberts
An embedded and charset-unspecified text was scrubbed...
Name: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-ecology/attachments/20110522/f71ebab8/attachment.pl>
Hi Dave,

I used the indval function from the package labdsv (your package Dave, I
believe) and confirmed these results with output from the indicspecies
package. I want to use all the species to construct the PCA ordination but
only want to show the species vectors for those 27 species identified as
significant group discriminators.

Andy
If you really want to do this Canoco-esque type of thing then it is
easy, but you need to build the plot up yourself. Here is an exmaple:

## load packages and example data
require(vegan)
require(labdsv)
data(dune)
data(dune.env)

## Indicators of Management type
set.seed(38)
inds <- with(dune.env, indval(dune, Management))

## ordinate
dune.pca <- rda(dune, scale = TRUE)

## exract scores you want to plot presume species and sites
dune.site <- scores(dune.pca, display = "sites", scaling = 3)
dune.spp  <- scores(dune.pca, display = "species", 
                    scaling = 3)[inds$pval <= 0.05, ]

## plot
plot(dune.pca, display = c("sites","species"), type = "n",
     scaling = 3)
points(dune.site)
arrows(0, 0, dune.spp[,1], dune.spp[,2], col = "red", length = 0.05)
lab.xloc <- dune.spp[,1] * 1.2
lab.yloc <- dune.spp[,2] * 1.2
text(lab.xloc, lab.yloc, rownames(dune.spp), col = "red", cex = 0.8)

The placement of the labels is crude and can be improved, but not in a
mailing list post.

HTH

G
On Sun, May 22, 2011 at 1:19 AM, Dave Roberts <
dvrbts at ecology.msu.montana.edu> wrote:

On 05/20/2011 11:54 PM, Andrew Halford wrote:

Hi All,

I have done a IndVal analysis on a dataset which has identified 27 species
out of 57 as being significant arbiters of site differences. What I want
to
do is a PCA biplot (data is hellinger transformed prior to this) with only
these 27 species on the ordination, not all 57, How can I control the
species I want displayed?

cheers

Andy

--
Andrew,

  We need a little more information.  Which IndVal function did you use?
 It should return a vector of p-values you can use as a mask [p-val <=
0.05,] on your taxon matrix.  But are you saying you want to use all 57 in
constructing the PCA and only identify 27, or that you only want to use
those 27 in the construction of the PCA?

Dave Roberts

_______________________________________________
R-sig-ecology mailing list
R-sig-ecology at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology

%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Dr. Gavin Simpson             [t] +44 (0)20 7679 0522
 ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
 UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
Thanks Gavin.  That's actually very nice and I suspect just what Andy 
was looking for.

Dave
require(vegan)
require(labdsv)
data(dune)
data(dune.env)

## Indicators of Management type
set.seed(38)
inds<- with(dune.env, indval(dune, Management))

## ordinate
dune.pca<- rda(dune, scale = TRUE)

## exract scores you want to plot presume species and sites
dune.site<- scores(dune.pca, display = "sites", scaling = 3)
dune.spp<- scores(dune.pca, display = "species",
                     scaling = 3)[inds$pval<= 0.05, ]

## plot
plot(dune.pca, display = c("sites","species"), type = "n",
      scaling = 3)
points(dune.site)
arrows(0, 0, dune.spp[,1], dune.spp[,2], col = "red", length = 0.05)
lab.xloc<- dune.spp[,1] * 1.2
lab.yloc<- dune.spp[,2] * 1.2
text(lab.xloc, lab.yloc, rownames(dune.spp), col = "red", cex = 0.8)

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
David W. Roberts                                     office 406-994-4548
Professor and Head                                      FAX 406-994-3190
Department of Ecology                         email droberts at montana.edu
Montana State University
Bozeman, MT 59717-3460