I'm kinda having a hard time understanding what the step-by-step procedure for the analysis of fluctuating asymmetry is with MorphoJ and thus far, every publication I'm reading doesn't really make it clearer.
I'm comparing different kind of butterflies (basically 2 'larger groups', i.e. forest and grassland butterflies) to basically see whether one or the other group is more exposed to stressors by the sorroundings. This is where my first problem appears: since I'm comparing many different butterfly species with different wing veins, I can't compare them in one run, since the procedure requires the same amount of Landmarks and structure (i.e. I can't directly compare the wings of Sphingidae and Nymphalidae doing only one run in MorphoJ, or can I?).
Benítez et al. 'Breaking Symmetry' Paper for example has a pretty useful flow chart, explaining in 5 short steps how to do it. But if you're new to the field, some crucial information seems to be missing. Like, i get that I'm starting with the Procrustes fit step after i digitalized the left and right side individually. But does the Procrustes fit step happen after i imported the two sides to the MorphoJ Software and combined them? (that's how I'd understand it at least, since the procrustes fit step removes size, location and orientation and just extracts the shape information, so the left and right dataset become one "procrustes fitted" dataset). And then the Procrustes ANOVA analysis is performed on this very combined dataset? If not and the combination is in fact the combination of the repeated measurements of one side, how do I compare the left and right side with one another? And can I compare the results of Individuals using given values from the ANOVA (e.g. the MS value for Ind * side) ?.
I'm sure I've got more questions regarding this topic, but that's it for now.
Thanks in advance!
Manuel
Hi Manuel, I don´t have the answers you need, but I think you can find some of them watching a course by Hugo Benitez. I remember a class about fluctuating asymmetry and other about MorfoJ (https://www.youtube.com/playlist?list=PLRcUGHF9lgaHwRCs2QUN9f97O0WNv-UcL). I also encourage you to write to him directly, is a really nice person, I think he would be happy to help you.
Good luck!
I can't answer all your questions, but the MorphoJ manual will answer some of them. In the section Procrustes Anova in the UserGuide, it says how the analysis is done, and what you need to provide the software. You might ask these questions on [email protected] because that's for people with questions and answers about morphometrics.
Solimary García-Hernández i will definitely have a look at his Videos, thank you! Also, i think contacting him would be the easiest way, I will try this asap.
Miriam Leah Zelditch I already looked at the UserGuide, but I think it's never wrong to give it another try. Maybe I just overlooked it, thanks
If you haven't yet received an answer and the video did not help, the main thing to know is that the analysis is done by the Procrustes Anova. So, the steps that precede that include creating a new project, creating a data set (importing the data and the classifiers, which should contain Individuals, Sides, and Replicates and if you have another factor, that should also be in the Classifiers. If you have object rather than matching symmetry, you don't need to specify "Side"). MorphoJ is case-sensitive so "Left" and "left" are not equivalent.
Once you have the Project, and dataset loaded, do the Procrustes Fit. Then do the Procustes Anova. For each step, you will probably need consult the manual (which will tell you what the program expects of your input files, as well as how to load them and whatever is needed for the particular analysis you are running.
Miriam Leah Zelditch I also found that this is the way to go. Turns out, if one knows what needs to be done, the procedure mentioned by Benitez et al makes much more sense. With a combination of trial and error and contacting Mr. Hugo Benitez himself, I managed to run the FA analysis in MorphoJ successfully. It is also important to note, that the value of interest is the Ind*Side value for MS.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA