I would like to apply the phylogenetic generalized least-squares
analysis implemented in NTSYS. I have tried with the three species
example from Rohlf (2001, Evolution,55:2143-2160), but I am unable to
reproduce the values for the intercept (0.6698; pag 2147) and the slope
(0.0617).
I would appreciate any suggestion.
Miquel Palmer
=======================
Here you are the files I am using:
1) Tree file
#NEXUS
Begin trees;
Translate 1 sp1, 2 sp2, 3 sp3;
tree ((1:1,2:4):5,3:6);
End;
2) The y file
1 1 3 0
1.0 1.25 0.5
2) The x file
1 1 3 0
1.5 1.0 0.75
3) ...and the output
MulRegr: NTSYSpc 2.10e, (C) 2000, Applied Biostatistics Inc.
Date & time: 25/06/2003 10:17:59
----------------------------------------
Input parameters
Read Y matrix from file: C:\Y.NTS
Variables are: rows
Use all Y variables
Read X matrix from file: C:\X.NTS
Variables are: rows
Use all X variables
Read covariance matrix from file: C:\PHYLOCOV.NTS
Details of computations will be displayed
Regressions will be through the means
Y-matrix:
Matrix type = 1, size = 1 by 3, missing value code = "none"
(rectangular)
X-matrix:
Matrix type = 1, size = 1 by 3, missing value code = "none"
(rectangular)
V-matrix:
Comments:
COPH: tree=C:\tree.nex, method=PhyloCov
Matrix type = 3, size = 3 by 3, missing value code = "none"
(similarity)
=== Generalized least-squares analysis (cov. among obs.) ===
Y Variables Mean Std.Dev.
1 0.77966 0.13809
X Variables Mean Std.Dev.
1 1.08051 0.20713
Partial regression coefficients:
Y variable 1:
Intercept 0.8985
X variables b SE ts df Prob
1 -0.1100 0.6857 -0.1604 1 0.89872
Variance inflation factors:
Singular variable that has been removed
Summary results for each Y variable:
Variable R2 Fs df1 df2 Prob
1 -0.0578 -0.0273 2 1 99.00000
Multivariate tests not performed because there was only one Y variable.
Ending date & time: 25/06/2003 10:18:00
==
Replies will be sent to list.
For more information see http://life.bio.sunysb.edu/morph/morphmet.html.
