-------- Original Message --------
Subject: INTRO PAPER FOR BEGINNERS IN GEOMETRIC MORPHOMETRICS Re: help
with classifiers and assessment of ME greatly appreciated
Date: Tue, 25 Oct 2011 04:53:57 -0400
From: andrea cardini <[email protected]>
To: [email protected]
Dear Fidi,
we've just published an introductory paper for beginners where we try to
suggest a simple and in some aspects simplistic protocol for taxonomic
comparisons using landmarks and Procrustes methods. It's on plants but most
of the time it can be used on animals by simply removing the effect of
leaves (i.e., multiple copies of a structure from the same individual) and
assuming that one is working on the same ontogenetic stage and sex. The
effect of sex could be easily tested and taken into account.
Most of the analyses are performed step by step in MorphoJ and PAST. There
are simple guidelines, explanations on how to extract classifiers from
image names and also some considerations about potential pitfalls
(including why after a superimposition one is most unlikely to be able to
test errors one landmark at a time). The original data used in the study
are also available.
Viscosi V., Cardini A., 2011 - Leaves, taxonomy and geometric
morphometrics: a simplified protocol for beginners. PLoS ONE 6(10): e25630.
doi:10.1371/journal.pone.0025630:
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0025630
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0025630#s5
There are some clear limits in this protocol and they're mentioned but it's
easy and, we hope, it may help beginners. For more sophisticated and
accurate methods, you will find plenty of papers and some books (the Green
Book by Zelditch et al., 2004, and few others).
NaN I guess stands for non-available or non-applicable and it is probably
because sample size may be not adequate to get a parametric test.
Remember also that in the Procrustes ANOVA the way and also the order by
which you specify factors may be important.
Cheers
Andrea
At 15:45 21/10/2011 -0400, you wrote:
-------- Original Message --------
Subject: help with classifiers and assessment of ME greatly appreciated
Date: Tue, 18 Oct 2011 18:57:42 -0400
From: Jordan, Friederike <[email protected]>
To: [email protected] <[email protected]>
Hello everybody,
I am a Master student conducting a morphometric study on New Zealand
common dolphin skulls. In addition to manual measurements taken with
calipers ee have also aquired 98 landmarks from 20 skulls with a
microscribe and I am planning to analyse the shape with MorphoJ.
I have read through the help manual of the software, but I am stuck and
I was hoping that someone could help me. Sorry, for asking really basic
questions, but I have never used the software before and nobody else
here has who I could ask for help.
We have taken repeated measurements from a number of skulls and I would
like to assess the measurement error of each the 98 landmarks for each
repeat and. I have trouble uploading classifiers for the files. The
names of two repeats are KS09-13Dd_test1 and KS09-13Dd_test2. I
created a text file with two columns (test: should be the classifier):
ID test
KS09-13Dd_test1 test1
KS09-13Dd_test2 test2
First, I created an identifier from this text fiele and afterwards I
tried to create the classifiers from the same text file. However, when I
upload the file I get an error message saying that the second classifier
could not be matched.
Does anybod yhave an idea what I might be oing wrong?
I read that Procrustes ANOVA can be used to assess measurement error, is
that correct? However, am I correct that it is not possible to assess
the measurement error for each individual landmark in MorphoJ? Does
anybody know a software where the individual assessment is possible?
If it is possible in morphoJ, do the data points for the 2 skulls (the
repeats) have to be combined in a single data set, or should the data
points for each skulls be uploaded as separate data sets?
When selecting the classifiers in the Procrustes ANOVA window could I
just select one classifier for individual, or do I have to input
something for side or error1 and error2?
I tried to carry out a Procrustes ANOVA on one repeat (with the error
message of the classifiers) and I got the output below:
Centroid size:
Effect SS MS df F P (param.)
Individual 0,000441 0,000441 1
Shape, Procrustes ANOVA:
Effect SS MS df F P (param.)
Pillai tr.
P (param.)
Individual 0,00001974 0,0000000688 287 NaN NaN
I was wondering what NaN stands for? I guess the analysis did not work?
Thank you very much in advance!! My email address is : [email protected]
Thanks a lot!
Fidi
Friederike Jordan
MSc Student
Coastal-Marine Research Group
Institute of Natural Sciences
Massey University
Private Bag 102 904
Auckland 0745
New Zealand
Phone: + ext.41520 Fax: +64 (0) 9 443 9790
Email: [email protected]
Courier address: Massey University, Building 5, Gate 4, Old Albany
Highway, Albany, Auckland
