[Vegan-commits] r1338 - in pkg/vegan: R inst man
noreply at r-forge.r-project.org
noreply at r-forge.r-project.org
Tue Oct 19 10:15:02 CEST 2010
Author: jarioksa
Date: 2010-10-19 10:15:01 +0200 (Tue, 19 Oct 2010)
New Revision: 1338
Modified:
pkg/vegan/R/treedist.R
pkg/vegan/inst/ChangeLog
pkg/vegan/man/treedive.Rd
Log:
treedist works with zero height trees and man tells that dissimilarity range is 0..2
Modified: pkg/vegan/R/treedist.R
===================================================================
--- pkg/vegan/R/treedist.R 2010-10-18 12:35:34 UTC (rev 1337)
+++ pkg/vegan/R/treedist.R 2010-10-19 08:15:01 UTC (rev 1338)
@@ -4,16 +4,18 @@
n <- nrow(x)
ABJ <- matrix(0, n , n)
dmat <- as.matrix(cophenetic(tree))
- for(j in 1:n)
+ for(j in 1:n) {
for (k in j:n) {
jk <- x[j,] > 0 | x[k,] > 0
if (sum(jk) > 1)
ABJ[k, j] <- treeheight(update(tree, d = as.dist(dmat[jk, jk])))
}
+ }
A <- diag(ABJ)
AB <- as.dist(outer(A, A, "+"))
ABJ <- as.dist(ABJ)
out <- (2 * ABJ - AB)/ABJ
+ out[ABJ==0] <- 0
attr(out, "method") <- "treedist"
attr(out, "call") <- match.call()
attr(out, "Labels") <- row.names(x)
Modified: pkg/vegan/inst/ChangeLog
===================================================================
--- pkg/vegan/inst/ChangeLog 2010-10-18 12:35:34 UTC (rev 1337)
+++ pkg/vegan/inst/ChangeLog 2010-10-19 08:15:01 UTC (rev 1338)
@@ -4,6 +4,11 @@
Versin 1.18-14 (opened October 15, 2010)
+ * treedist: works now with zero-height trees (for instance, only
+ one item in a tree). Help page now tells that tree distances are
+ in range 0..2 instead of 0..1, since combining two trees may add a
+ new common root.
+
* isomap: returns only axes associated with positive eigenvalues.
* wcmdscale: give two goodness of fit statistics for 'k' axes: one
Modified: pkg/vegan/man/treedive.Rd
===================================================================
--- pkg/vegan/man/treedive.Rd 2010-10-18 12:35:34 UTC (rev 1337)
+++ pkg/vegan/man/treedive.Rd 2010-10-19 08:15:01 UTC (rev 1338)
@@ -50,10 +50,12 @@
species in a common tree and seeing how much of the tree height is
shared and how much is unique. The current function uses a Jaccard
like approach and defines dissimilarity as the total unshared height
- proportional of the total height of the combined tree. The idea is
- similar as in the UniFrac distance (Lozupone and Knight 2005), but
- the implementation is completely different and only works in the
- presence absence framework.
+ proportional of the total height of the combined tree. However, the
+ index is not limited to the maximum of one, since combining two trees
+ may add a new root, and therefore the maximum dissimilarity is two.
+ The idea is similar as in the UniFrac distance (Lozupone and Knight
+ 2005), but the implementation is completely different and only works
+ in the presence absence framework.
The functions need a dendrogram of species traits as an input. If
species traits contain \code{\link{factor}} or \code{\link{ordered}}
More information about the Vegan-commits
mailing list