
Jump to:
Freely distributable documents and computer programs
Le Progiciel R / The R Package
CCA Bibliography 86-96
Lexique anglais-français d'écologie numérique et de statistique


|
|
How to use T-REX
(Windows version)
Using T-REX is easy!
- Start the program by double-clicking its icon.
- Open a dissimilarity matrix file (in ASCII format).
- Look at the resulting window to ensure that this is the matrix you wanted.
- Select Reconstruct a tree, Reconstruct a tree from a partial matrix or Reconstruct a reticulogram in the T-REX menu.
- Choose a tree reconstruction method for your additive tree (left-hand side of the dialog), and select what to output on the right-hand side. T-REX will now compute the fitted distance matrix, tree or reticulogram edges and tree or reticulogram statistics, which will appear in a new text window. A second window will contain the tree or reticulogram drawing. You can copy the information appearing in both windows to the clipboard, to be pasted into your favorite text or picture editor. The information in both windows can also be saved and printed.
T-REX screenshot, showing a phylogenetic tree and reticulation network
inferred from a dissimilarity matrix among nine species of frogs (see Case,
1978 for more information).
General Information about T-REX
This program carries out some algorithms for the reconstruction of additive (phylogenetic) trees
and reticulograms given a dissimilarity matrix. Additive trees can also be inferred
from data matrices containing missing values. An additive tree distance or a
reticulogram distance is fitted to the given dissimilarity.
As far as additive tree reconstruction is concerned, the program carries out five
methods of fitting an additive distance (distance representable by a tree with
non-negative edge lengths) to a given dissimilarity.
The following methods are available:
- ADDTREE by Sattath and Tversky (1977),
- Neighbor-joining (NJ) method by Saitou and Nei (1987),
- Unweighted neighbor-joining method (UNJ) by Gascuel (1997),
- Circular order reconstruction method by Makarenkov and Leclerc (1997), and Yushmanov (1984),
- Weighted least-squares method MW by Makarenkov and Leclerc (1999).
You will find more information about these methods in the Additive Tree
reconstruction section of the T-REX Help manuel.
When reconstructing additive trees from dissimilarity matrices containing
missing values, the following methods are available:
- Triangle method by Guénoche (1999),
- Ultrametric procedure for estimation of missing values by De Soete (1984) and
Landry et al. (1996) followed by the MW method,
- Additive procedure for estimation of missing values by Landry et al. (1996)
followed by the MW method,
- Weighted least-squares method MW by Makarenkov and Leclerc (1999), with weights of 1 to existing trees and 0 to the missing ones.
With reticulogram reconstruction, the program first computes a classical
additive tree using one of the five available tree reconstruction algorithms.
Then, at each step of the reticulogram reconstruction procedure, a reticulation
(i.e., new edge) is chosen to minimize the least-squares or the weighted
least-squares loss function. That edge is then added to the growing reticulogram.
You will find more information about these methods in the Reticulogram reconstruction
section of the T-REX Help manuel.
As results T-REX provides:
- A window with the tree or reticulogram fitting statistics:
- The fitted additive or reticulogram distance matrix,
- The list of the tree or reticulogram edges with their lengths,
- The values of the (weighted) least-squares criterion, the (weighted) average
absolute difference, the (weighted) maximum absolute difference and the total
length of the obtained tree or reticulogram. If reticulogram reconstruction is
performed the program also provides the values of the (weighted) least-
squares criterion, as well as the values of the goodness-of-fit criterion Q1 or Q2
for each reticulation added to the basic additive tree.
- A window with the tree or reticulogram drawing. The tree edges are depicted by full lines
and the supplementary edges (reticulations) are depicted by dashed lines.
Last updated on Sunday, August 01, 2010 by Philippe Casgrain
Created on Thursday, September 14, 2000
|