This manuscript describes in detail the GPA functionality of the SlicerMorph package implemented in the free software Slicer and discusses the broader implications of applying a common workflow in studies involving 3D geometric morphometric analyses. The manuscript is well written, and I find the workflow in Figure 4 particularly useful.
I recommend minor revisions especially regarding the figures. I also provide some general thoughts on the SlicerMorph package that I hope can contribute to its future development.
See attached file.

I was pleased to see this manuscript being resubmitted as I think SlicerMorph is an important tool that deserves to be published in a high-profile journal.
Overall, the authors have addressed the comments provided by me and the other reviewers on the text in a satisfactory way, even if some issues still remain. The additional SOM and in the Github are very useful and clearly written. The figures and captions although have not changed at all and I think they need to be improved before publication.

Of the two main issues I had with the previous version of the manuscript, the first one – importing the
the landmark files created in SlicerMorph in R – is now resolved with two new functions as detailed in the SOM.
The GPA problems have been addressed, even if a better Figure 3 would help the reader a lot in navigating the module.
I am still confused on the PC deformation visualization. In lines 274-275 the authors state “The scaling coefficient associated with each PC is arbitrary and aims to provide a noticeable deformation for visualization purposes.”. To me this means that the deformation visible in the 3D model is not correlated with the actual PC axis variation. Even if it is small this should not be magnified (or at least it should only be magnified using a known factor that can be changed) or it becomes difficult to interpret the results.
Reading the SOM though, the authors state that the deformation is correlated with the PC axis variation and that therefore similar to the geomorph/plotRefToTarget function that they mention in their response. I think maybe this tool was changed while the manuscript was being written, but please make sure what this tool does is clear and the information in the manuscript are not at odds with the SOM.

The comments I had on the figures have not been addressed at all, so I repeat them here:

Figures

Figure 1 – This figure is confusing. It is difficult to keep track of what each box represents, especially since the first box has 3 letters in it (A-B-C) while the others have only 1 letter. I think it would be best to simplify the figure by putting only 1 example for each landmarking method per box, similar to D in the current figure, and to substitute the letters with the names of the different landmarking methods. The ALPACA workflow should go in a different figure or in the Supplemental Information to make the figure more uniform.

Figure 2 – While testing this dataset in SlicerMorph, the fact that landmark 25 is problematic is very clear. In this figure, the visualization does not allow to easily see this. I recommend changing the graphics, especially B that is too zoomed in, so that the issues with landmark 25 are well visible and that this can serve as an example for future users of SlicerMorph.

Figure 3 – The figure does not match the layout of SlicerMorph, which only has 5 boxes, and it is difficult to follow. It would be best to redesign it so that it can serve as an example for future users of SlicerMorph. The caption should explain more in detail what each box represents and what information does it convey.

Figure 4 – I really like the idea of having this workflow presented in the paper. I would just prefer it was more colorful so that it is easier to follow visually. It would also be useful to re-define “morphotype” in the caption, as its definition can be lost in the text.