On Monday, December 21, 2015 at 2:52:33 AM UTC+3:30, Carmelo Fruciano wrote: > > Elahep <[email protected] <javascript:>> ha scritto: > > > > > > > Hi morphometricians, > > > > > > I have a data set consists of 7 landmarks and 7 semi-landmarks on the > > carapace of a crab and I want to control my data for the effect of > > allometry with tpsRegr. I am not sure which dataset should I use as > > dependent variable (presence of semi-landmarks confused me in choosing > the > > correct dataset as dependent variable!): > > > > > > 1. 1- TPS file consists of landmarks and semi-landmarks *before* > > removing the control lines > > 2. 2- TPS file consists of landmarks and semi-landmarks *after* > removing > > the control lines > > 3. 3- TPS file consists of superimposed landmarks and > semi-landmarks > > (= *Procrustes coordinates*) > > > > > > Would you please help me with this question? > > Dear Elahep, > not sure what you mean with "control lines". > > The procedure one generally uses when dealing with semilandmarks is to > perform a generalized Procrustes analysis with sliding of > semilandmarks (typically in tpsRelw), export the aligned coordinates > (and the centroid size, if you want to use that as a measure of body > size) and then treat the exported coordinates as if they were landmarks. > In your case, then, I would imagine you would use in tpsRegr the > aligned coordinates and the centroid size previously exported from > tpsRelw. > > Another thing you might want to consider in the process is if your > configuration of landmarks (or, rather, the biological structure > you're trying to capture with them) is symmetric and, if so, if > asymmetry is of interest for you... > > Best, > Carmelo > > Dear Carmelo,
I meant "background curves" by control lines, because the program doesn't support background curves. As you mentioned I'll use the aligned coordinates of landmarks and semilandmarks as dependant variables. My data doesn't have asymmetry since I put landmarks on only left side of the specimens. Thanks a lot for your help. It was great. Elahe -- MORPHMET may be accessed via its webpage at http://www.morphometrics.org --- You received this message because you are subscribed to the Google Groups "MORPHMET" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
