[MORPHMET] Re: testing for convergence in 3D shape space
Thanks everyone! On Thursday, January 28, 2016 at 1:39:43 AM UTC+11, Christy Hipsley wrote: > > Dear All, > > I'm working with a 3D GM data set and am looking for a specific test of > convergence that also accounts for phylogeny. After some searching it's > still not clear to me what is appropriate - so far I only see programs that > use reduced PC axes as continuous characters, but nothing that uses the > full Procrustes coordinates. I've applied a phylogenetic ANOVA in Geomorph > to at least show that after accounting for phylogeny, morphological shape > is significantly different among ecological groups (in this clade several > unrelated lineages occur in the same ecological niche). > > Does anyone know of a method for testing for convergence across a > phylogeny using the full shape data or is it always using PC scores? What > about using scores from a CVA instead, since they are specifically > addressing the question of differences among the a priori ecological groups > (and PCA does not)? Could those be mapped onto a phylogeny and modelled in > terms of BM vs OU, to test if that morphological change is adaptive? > > Any advice on these analyses would be appreciated. I'm aware of the > package Surface but I'm not convinced that is right for my system. > > My shape data has a significant phylogenetic signal and I suspect that one > ecological group in particular is highly convergent. > > > Thanks for any help! > > Christy > > School of BioSciences > University of Melbourne > Parkville VIC 3010, Australia > Email: chips...@unimelb.edu.au > -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] comparing ontogenetic trajectories in a phylogenetic framework
Dear GM community, I'd like to get some opinions on methods for comparing ontogenetic trajectories of extant taxa within a phylogenetic framework. I've read several papers and spoken to colleagues about this and there appears to be no satisfactory conclusion. It seems that one can either describe trajectories as vectors (usually based on PC axes) and compare angles between them against a phylogeny using ancestral state reconstruction/PIC, or use the Procrustes data themselves, divided into age bins, to reconstruct ancestral morphologies of juveniles and adults, use those to build an ancestral ontogeny, and then compare that to the tip taxa. Both of these methods seem unsatisfactory to me, in that they rely on observed data to reconstruct ancestral states, which by definition will not fall outside of the current range. I know ASR is implicit and necessary in most phylogenetic tests, but I wonder if there is another way, or at least a complementary way to e.g., test for heterochronic shifts in ontogenetic allometry. Thanks for any thoughts on this, and I realize it may be a difficult subject. References to any papers with relevant methods would also be appreciated! Christy School of BioSciences University of Melbourne Parkville VIC 3010, Australia Email: chips...@unimelb.edu.au Museum Victoria GPO Box 666 Melbourne VIC 3001, Australia Ph: +61 (0)3 8341 7423 Email: chips...@museum.vic.gov.au -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] testing for convergence in 3D shape space
Dear All, I'm working with a 3D GM data set and am looking for a specific test of convergence that also accounts for phylogeny. After some searching it's still not clear to me what is appropriate - so far I only see programs that use reduced PC axes as continuous characters, but nothing that uses the full Procrustes coordinates. I've applied a phylogenetic ANOVA in Geomorph to at least show that after accounting for phylogeny, morphological shape is significantly different among ecological groups (in this clade several unrelated lineages occur in the same ecological niche). Does anyone know of a method for testing for convergence across a phylogeny using the full shape data or is it always using PC scores? What about using scores from a CVA instead, since they are specifically addressing the question of differences among the a priori ecological groups (and PCA does not)? Could those be mapped onto a phylogeny and modelled in terms of BM vs OU, to test if that morphological change is adaptive? Any advice on these analyses would be appreciated. I'm aware of the package Surface but I'm not convinced that is right for my system. My shape data has a significant phylogenetic signal and I suspect that one ecological group in particular is highly convergent. Thanks for any help! Christy School of BioSciences University of Melbourne Parkville VIC 3010, Australia Email: chips...@unimelb.edu.au -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] MeshLab - setting default FOV as Ortho?
Hi all, Does anyone know if it's possible to set the Field of View (FOV) in MeshLab to Ortho as a default setting? It drives me nuts each time I open a file to have to change it, and I don't know why anyone would work in MeshLab without automatically changing this anyway. I've asked a couple of other colleagues but no happy solution so far - I'm hoping someone here might have the magic answer! Thanks, Christy -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] Re: allometry and phylogeny
To follow up on Damien's question and Dean's response, which values from the advanced.procD.lm results would one use for reconstructing species ontogenetic slopes? There are several outputs relating to slope from that test, like "slopes", "slope.lengths", "slopes.dist", "slopes.angles". I understand these each describe different parameters of the slope vector like length and orientation - basically magnitude and direction of shape change per unit size. So should one reconstruct ancestral states of each of these each separately to compare with tip taxa? Thanks for any advice, Christy On Friday, July 1, 2016 at 12:31:37 AM UTC+10, dcadams wrote: > > Damien, > > > > Yes, one can take the regression parameters for each species and use these > as ‘tips’ data to obtain ancestral estimates of the regression slope. In > this case, the species’ data is the multivariate set of slope coefficients > for each multivariate regression. Thus, the ancestral state of the > allometry is also a multivariate set of parameters. > > > > Since you are using geomorph the slopes for each species’ allometry > trajectory may be found in the output list from advanced.procD.lm. > > > Best, > > > > Dean > > > > > > Dr. Dean C. Adams > > Professor > > Department of Ecology, Evolution, and Organismal Biology > >Department of Statistics > > Iowa State University > > www.public.iastate.edu/~dcadams/ > > phone: 515-294-3834 > > > > *From:* Damien Esquerre Gheur [mailto:damien@anu.edu.au ] > > *Sent:* Thursday, June 30, 2016 6:50 AM > *To:* morp...@morphometrics.org > *Subject:* [MORPHMET] allometry and phylogeny > > > > Dear morphometricians, > > > > I have a pretty big dataset of specimens for most species of pythons > encompassing most of the size range for every species. I have been > projecting the allometric trajectories and doing pairwise slope comparisons > and things like that in Geomorph. However, it seems very obvious that the > is a very strong phylogenetic signal in the data, as species within the > same clade tend to have parallel trajectories. > > I have been trying to find a way to analyse this data in a phylogenetic > framework, and so far I know there is no straightforward solution. I know > in Adams and Nistri 2010 the trajectories where coded ad isometric and > allometric and then reconstructed the ancestral states but the case it not > so binary with my data. > > Would there be a way, for example, to extract a value for the slope and do > an ancestral state reconstruction on that? Also, is there a way to > incorporate phylogeny in slope comparison models? > > Any help would be greatly appreciated > > > > Damien Esquerré > > PhD Student, Keogh Lab > > Division of Evolution, Ecology and Genetics, Research School of Biology > > The Australian National University > > 44 Daley Road, ACTON ACT 2601, Australia > > > > -- > 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 morphmet+u...@morphometrics.org . > -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] the problem with CVA... or is it?
Dear Morphmet-ers, I'm seeking advice on methods for visualizing shape features that distinguish multiple groups using GM. I know CVA has fallen out of favor for a number of reasons discussed here - e.g., more variables than groups, nonisotropic variation: Mitteroecker, P., and Bookstein, F. 2011. Linear discrimination, ordination, and the visualization of selection gradients in modern morphometrics. Evol. Biol. 38:100–114. Klingenberg, C. P., and Monteiro, L. R. 2005. Distances and directions in multidimensional shape spaces: Implications for morphometric applications. Syst. Biol. 54:678–688. Although given these limitations, is it really expected to give completely false results regarding the visualization of shape changes? In my study sytem, I show that ecological groups have statistically different cranial shapes, using both Procrustes ANOVA and PGLS. Now I simply want to visualize what the main features are that distinguish them, preferably using warps or wireframes, so that those changes must be directly relateable to the original landmark coordinates. I did that using individual specimens instead of species means, so I have 161 individuals vs 144 variables (48 landmarks*3D). I also did a between-group PCA on the species means which shows the same pattern, so is it technically "wrong" to show both? Thanks for any feedback on this issue, and I would appreciate to hear any alternative methods that people might use. I use MorphoJ and Geomorph for analyses. Best, Christy -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] Re: procD.allometry with group inclusion
Hi Mike, I have a follow-up to Tsung's procD.allometry question - If the initial HOS test is nonsignificant for (shape~size, ~group), but the following ANOVA table is significant for size and group effects, can one interpret that as differences in y-intercept but not slope? So basically the HOS test determined that the size:group interaction did not improve the model, and so removed it from the ANOVA formula? In that case, the ANOVA model is (shape~size+group), correct? That makes perfect sense looking at the output graphs, but I just want to be sure. And if one wanted to then determine which groups differed in y-intercept, would one set up the advanced.procD.lm model like this and compare the pairwise LS means? advanced.procD.lm(Y ~ size, ~ size+group, groups = ~group, slope = NULL, iter=1) Thanks if you can confirm that I'm setting this up correctly, Christy On Thursday, December 8, 2016 at 7:37:33 PM UTC+11, Tsung Fei Khang wrote: > > Hi all, > > I would like to use procD.allometry to study allometry in two species. > > I understand that the function returns the regression score for each > specimen as Reg.proj, and that the calculation is obtained as: > s = Xa, where X is the nxp matrix of Procrustes shape variables, and a is > the px1 vector of regression coefficients normalized to 1. I am able to > verify this computation from first principles when all samples are presumed > to come from the same species. > > However, what happens when we are interested in more than 1 species (say > 2)? I could run procD.allometry by including the species labels via > f2=~gps, where gps gives the species labels. Is there just 1 regression > vector (which feels weird, since this should be species-specific), or 2? If > so, how can I recover both vectors? What is the difference of including > f2=~gps using all data, compared to if we make two separate runs of > procD.allometry, one for samples from species 1, and another for samples > from species 2? > > Thanks for any help. > > Rgds, > > TF > > > > > > -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] fitting lines to allometric plots?
Hi all, This is a fairly simple question, but one that I am unsure of - is it appropriate to fit lines to plots of allometric scores, for example regression score or common allometric component (CAC) vs size, or the first residual shape component (RSC) vs CAC? If the scores represent individuals from different groups or species, does it make sense to add a line of best fit to each group to visualize trends? I've noticed in other papers that lines are not used, so I'm wondering if this is somehow "wrong". I realize the line itself is not the regression line, but if it's only used for visual comparison, is that technically okay? Thanks for any feedback, Christy -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] creating mirror reflection of one-side landmarks
Dear all, I'm working with a 2D landmark data set of vertebrae in anterior view in which one side (right) has been landmarked, with 5 landmarks along the midplane axis for a total of 12 points. I would like to generate a mirror reflection of the paired landmarks, and wonder if there is an easy way to do this. I naively tried multiplying the x coordinate dimensions *-1 and keeping y the same, but since the midplane x coordinates are not at 0 this obviously doesn't work. Any advice or code would be greatly appreciated! Thanks, Christy -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] modularity tests - effect of negative value in correlation matrix?
Dear all, I'm seeking help with modularity tests using the EMMLi package in R. I have developmental data for different species of mammals, and my goal was to group specimens in each species by age classes (e.g., neonate, juvenile, adult based on tooth eruption) and then test different hypotheses of cranial modularity following Goswami & Finarelli 2016 EMMLi: A maximum likelihood approach to the analysis of modularity. Evolution 70:1622–1637. Some of those groupings are rather small, <10 individuals. The first step is to generate a correlation matrix from the 3D array of landmark coordinates (Y), but when I do I get the message Warning message: In dotcorr(Y) : CVM has negative eigenvalue -1.1286357464565e-15 Will this will cause problems for the EMMLi model test, or I can I ignore it and move on? Also, does anyone have thoughts of the minimum number of samples one can use to compare modularity hypotheses? In Goswami & Finarelli, they show that reducing the dataset down to 10 individuals from 48 still performs very well, but is there a true minimum? For example could I take the 3 largest individuals from each species for the test? Thanks for any advice, Christy -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] new article out on digital morphology data openness based on Morphmet survey
Dear Morphmet community, Almost 2 years ago Emma Sherratt and I posted a survey here asking your feedback on generation and use of digital morphology data, and we are happy to announce that those results have now been published in a Comment article in Scientific Data titled '*Psychology, not technology, is our biggest challenge to open digital morphology data*', available here -> https://www.nature.com/articles/s41597-019-0047-0. We thank everyone again who participated, and many others who have engaged in similar conversations over the years. Best, Christy + Emma -- 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 morphmet+unsubscr...@morphometrics.org.
[MORPHMET] papers with 1000+ 3D models?
Hi all, I'm looking for GM or other similar papers that include high numbers of 3D models, like several hundreds or over a thousand. Particularly if these were generated using microCT. Thank you for any help, Christy -- MORPHMET may be accessed via its webpage at http://www.morphometrics.org --- To unsubscribe from this group and stop receiving emails from it, send an email to morphmet+unsubscr...@morphometrics.org.