12
$\begingroup$

Is there a tool that would help in creating rather intricate line drawings (not pre-rendered balls-and-sticks/wireframe/sticks/ORTEP diagrams) for a skeletal formula in 3D space?

To illustrate the question a bit further, here are two examples:

  1. Organic molecules — An all-benzene catenane and a trefoil knot [1]:

    Structures of all-benzene catenane 1a and all-benzene trefoil knot 2

  2. Inorganic molecules — proton transfer in polyoxoniobate $\ce{[H_xNb10O28]^{(6-x)-}}$ [2, p. 111]:

    Polyoxoniobate

I also saw a much larger line-art structure of an entire peptide drawn the similar way, but unfortunately I cannot find the reference at the moment.

Line-drawn structures like these are usually presented alongside with the determined crystal structures, so I naturally assume that there is a tool that would allow to convert a CIF file to an illustration like the ones shown above. However, I wasn't able to find anything relevant by screening the IUCr software list, and this already concerns me for several years.

Do people really use molecular editors like ChemDraw for this, or there is a more optimal way? Also, if ChemDraw is the tool for the job, are there any general recommendations as to how to efficiently draw structures like this in 3D space without loosing one's mind on estimating z-buffer for proper bonds' greyscaling/thickness/type based on the distance from the viewer as well as overlapping atoms and bonds?

I would really appreciate if someone who published similar work would share some tips regarding their workflow.

References

  1. Segawa, Y.; Kuwayama, M.; Hijikata, Y.; Fushimi, M.; Nishihara, T.; Pirillo, J.; Shirasaki, J.; Kubota, N.; Itami, K. Topological Molecular Nanocarbons: All-Benzene Catenane and Trefoil Knot. Science 2019, 365 (6450), 272–276. DOI: 10.1126/science.aav5021.
  2. Polyoxometalate Chemistry, 1st ed.; van Eldik, R., Cronin, L., Eds.; Advances in Inorganic Chemistry; AP, Academic Press, an imprint of Elsevier: Cambridge, MA San Diego, CA Oxford, 2017; Vol. 69.
$\endgroup$
8
  • 1
    $\begingroup$ If you ask this question at Mathematica stackexchange, the wizards there will come up with something. However, their solutions may be ad hoc, whereas you seem to searching for more generic software. $\endgroup$ Dec 15, 2019 at 13:33
  • $\begingroup$ @JosephO'Rourke I never thought this question would attract a math scientist:) Unfortunately, you are right, I'm looking for a more generic software, preferably an addon for a common chemical app or a free program, as sorting out licensing for Mathematica is a pain. $\endgroup$
    – andselisk
    Dec 15, 2019 at 13:40
  • 2
    $\begingroup$ There are editors that have 3D mode and they can display a 3D molecule in 2D manner. MarvinSketch is first that comes to my mind, but there are probably more. $\endgroup$
    – vapid
    Dec 15, 2019 at 14:33
  • 1
    $\begingroup$ This sounds like asking for a new (I think, then sixth) installment of ChemDraw Magick by ChemDraw Wizard Pierre Morieux (a chemist by training, working at PerkinElmer around ChemDraw). The address of the channel in question: youtube.com/user/PierreMx3/videos. $\endgroup$
    – Buttonwood
    Dec 15, 2019 at 18:43
  • 1
    $\begingroup$ @KarstenTheis Thank you, but the structures from 10.1038/scientificamerican1166-78 appear to be sticks diagrams: even though they were hand-drawn at that time, these days they can be replicated with crystallographic software. The only skeletal formula there I found was for polysaccharide chains, but that's not really what I mean. $\endgroup$
    – andselisk
    Dec 20, 2019 at 5:43

1 Answer 1

6
$\begingroup$

Thanks to vapid and their comment

There are editors that have 3D mode and they can display a 3D molecule in 2D manner. MarvinSketch is first that comes to my mind, but there are probably more.

I figured a crude solution, which I hope might as well serve as a starting point for other answers.

First of all, I completely forgot that MarvinSketch (free (registration required), available for Windows, Linux, macOS) has support for 3D. Drawing in 3D from scratch with MarvinSketch is probably not the best idea as even though it has a support for optimizing (cleaning) 3D structures, it sometimes gives weird results for organics and practically not suitable for inorganic complexes and especially clusters.

Arguably, the best and the most affordable third-party tool for creating such structures is Avogadro (open source, available for Windows, Linux, macOS).

If there is a crystal structure at hand (e.g. CIF data), then one must do two things first:

  1. crop/grow the fragment of interest;
  2. assign proper bond orders and export the structure in one of the formats that preserve bond orders (such as MOL, MOL2, PDB or CML).

I believe Avogadro can do both, but I personally like to do part 1 in Olex2 (free (registration required), available for Windows, Linux, macOS) by force of habit.

As a proof of concept, I took the crystal structure for the trefoil knot, allowed it to grow from assymetric unit to a complete molecule as well as cut out solvent and guest molecules, so I was left with the initial shape:

Olex2 screenshot

which I immediately exported to the MOL file by running file trefoil-knot.mol.

Next, I opened this MOL file with Avogadro, which automatically assigned bond orders:

Avogadro screenshot

and the file was exported to MOL once again, and subsequently opened in MarvinSketch (I reduced fog in settings, made implicit hydrogens hidden and used F7 button to align the molecule on screen):

MarvinSketch screenshot

And the final result came out to look like this:

Final view

The image is, however, not quite as perfect as the one in the paper. For instance, there are some artifacts about inflection points as if the program was confused how to place double bonds. Also, if you look closer, the grayed out part which was supposed to be further from the viewer, actually appears in front, which I attribute to the fact that Marvin was designed as a 2D molecular editor first and foremost, so it doesn't always handles depth (or the aforementioned z-buffer) adequately.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.