-------- Original Message --------
Subject: Re: tpsRegr independent variable file
Date: Tue, 8 Dec 2009 14:33:38 GMT
From: Thomas DeWitt <[email protected]>
To: <[email protected]>
References: <[email protected]>
Germán,
You can do your statistics (MANCOVA) in a stats program, possibly with
only the major principle components, and use tpsRegr for the
visualization part. For tpsRegr you would enter all your raw or aligned
coordinates ad dependent variables and enter your independent variables
(IVs) as a design matrix. Lets say you have shape as your dependent
variable suite, centroid size as a covariate, and three (nominal) levels
of a treatment. You need an IV matrix that includes centsize as one
column, and dummy variables for each level of the treatment. It might
look something like:
'tps file of independent variables for tadpoles
1 6L 4L 0
tad1 tad2 tad3 tad4 tad5 tad6
centsize trt1ornot trt2ornot trt3ornot
1.234 1 0 0
1.456 1 0 0
1.789 0 1 0
1.246 0 1 0
1.369 0 0 1
1.333 0 0 1
This would be an overspecified (SAS) design matrix. You could separate
3 groups in two dimensions if you design the dummy variables
differently. But for complicated designs compressed design matrices get
tricky. In any case it doesn't matter, because from tpsRegr all you
want is a visualization--you are not hypothesis testing or using the
p-values from the overspecified design. You get your stats from a stats
program:
RW1, RW2... RW5 = centsize + trt, and get a viz for each level of the
treatment effect (relative to global consensus) for each treatment
class.
Hope this helps,
-Thom DeWitt
<[email protected]> 12/8/2009 7:17 AM >>>
----- Forwarded message from -----
Date: Tue, 8 Dec 2009 10:39:49 GMT
From: German Orizaola
Reply-To: German Orizaola
Subject: tpsRegr independent variable file
To: [email protected]
Hi,
My name is Germán Orizaola, post-doc at the Uppsala University (Sweden)
and I am currently starting to conduct geometric morphometric analyses
on tadpoles. So far, I have been able to work with my data on tpsDig and
tpsRelw, obtaining and analysing relative warp scores for my
individuals. Now, I am trying to visualise my results on tpsRegr, and I
do not really know how to create a file including my individuals and the
associated independent variable data, to start working with the program.
Could someone, please, help me on this point……..
Thanks in advance,
Germán Orizaola
Germán Orizaola
Population and Conservation Biology
Department of Ecology and Evolution
Evolutionary Biology Centre
Uppsala University
Norbyvägen 18D
75236 Uppsala
Sweden
www.popbiol.ebc.uu.se/people
- ----- End forwarded message -----
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
