Programmatic conversion of crystal structures into 3D printable files using Jmol

Abstract

Background

Three-dimensional (3D) printed crystal structures are useful for chemistry teaching and research. Current manual methods of converting crystal structures into 3D printable files are time-consuming and tedious. To overcome this limitation, we developed a programmatic method that allows for facile conversion of thousands of crystal structures directly into 3D printable files.

Results

A collection of over 30,000 crystal structures in crystallographic information file (CIF) format from the Crystallography Open Database (COD) were programmatically converted into 3D printable files (VRML format) using Jmol scripting. The resulting data file conversion of the 30,000 CIFs proceeded as expected, however some inconsistencies and unintended results were observed with co-crystallized structures, racemic mixtures, and structures with large counterions that led to 3D printable files not containing the desired chemical structure. Potential solutions to these challenges are considered and discussed. Further, a searchable Jmol 3D Print website was created that allows users to both discover the 3D file dataset created in this work and create custom 3D printable files for any structure in the COD.

Conclusions

Over 30,000 crystal structures were programmatically converted into 3D printable files, allowing users to have quick access to a sizable collection of 3D printable crystal structures. Further, any crystal structure (>350,000) in the COD can now be conveniently converted into 3D printable file formats using the Jmol 3D Print website created in this work. The 3D Print website also allows users to convert their own CIFs into 3D printable files. 3D file data, scripts, and the Jmol 3D Print website are provided openly to the community in an effort to promote discovery and use of 3D printable crystal structures. The 3D file dataset and Jmol 3D Print website will find wide use with researchers and educators seeking to 3D print chemical structures, while the scripts will be useful for programmatically converting large database collections of crystal structures into 3D printable files.

Scalfani et al. J Cheminform (2016) 8:66 DOI 10.1186/s13321-016-0181-z

Programmatic conversion ofcrystal structures into3D printable les using Jmol

Vincent F. Scalfani1*, Antony J. Williams2*, Valery Tkachenko3, Karen Karapetyan3, Alexey Pshenichnov3, Robert M. Hanson4, Jahred M. Liddie5 and Jason E. Bara5

Background

Three-dimensional (3D) printed crystal structures oer several advantages over traditional molecular model constructs (e.g. plastic molecular model kits, Styrofoam balls, beads). Perhaps the greatest advantage is that 3D printing is capable of fabricating extremely complex molecular

structures that would be difficult or impossible to create with traditional molecular model fabrication techniques [1]. As such, numerous researchers have recently used 3D printed crystal structures to advance their teaching and research [17].

Digital 3D printable crystal structure les can be manually created from standard crystallographic information les (CIF) using a variety of open source, free, and commercial software packages [1, 2, 46]. In general, the crystal structure is rst opened in a crystallographic viewing program [812] where the structure representation is adjusted to preference. For example, it may be necessary

*Correspondence: [email protected]; [email protected]

1 University Libraries, Rodgers Library for Science and Engineering, The University of Alabama, Tuscaloosa, AL 35487, USA

2 ChemConnector, 513 Chestnut Grove Court, Wake Forest, NC 27587, USAFull list of author information is available at the end of the article

The Author(s) 2016. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/

Web End =http://creativecommons.org/ http://creativecommons.org/publicdomain/zero/1.0/

Web End =publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated.

Scalfani et al. J Cheminform (2016) 8:66

Page 2 of 8

to display the atoms as spheres, delete unwanted molecules, or generate a unit cell. Once the desired crystal structure representation is achieved, the structure is typically exported as either a stereolithography le (STL) for single color 3D printing or a virtual reality modeling language le (VRML) for multi-color printing. Lastly, les may need to be repaired of any inconsistencies in the surface of the 3D model (e.g. holes, overlaps) and veried for compatibility with 3D printer software before printing. In short, manual conversion of crystal structures into 3D printable le formats can be a tedious and time consuming process. As there are hundreds of thousands of crystal structures currently available in open access databases such as the Crystallography Open Database (COD) [13, 14], it would be ideal to automate the conversion of crystal structures (CIFs) from these databases into 3D printable le formats. Providing quick access to 3D printable crystal structure les removes the barriers associated with le manipulation and conversion, which makes it easier for educators and researchers to 3D print crystal structures.

There are currently only two methods available for pro-grammatic conversion of a crystal structure into a 3D printable le including the NIH 3D Print Exchange web tools developed by Coakley and coworkers at the NIH [1517] and the stand-alone Cif2VRML Windows application developed by Kaminsky [18]. The NIH 3D Print Exchange web tool uses Chimera [12] and Blender [19] scripts that automatically process crystal structures into 3D printable les. The Cif2VRML application is written in the Delphi programming language and allows for one-click creation of 3D printable les from CIFs. The NIH 3D Print Exchange web tool and Cif2VRML application have greatly advanced the process and efficiency of creating 3D printable crystal structures. However, crystal structures are still converted on an individual basis; users must upload or process each crystal structure individually. The NIH 3D Print Exchange Team has made all scripting code available in the public domain, so researchers could adapt these scripts to program-matically process multiple les locally in the future [16]. Multiple le processing is also not currently available in the Cif2VRML application, but is being considered in a future release [18].

In this article, we report our progress to overcome the current limitations with 3D printable crystal structures; that is, the lack of programmatic multiple le conversion methods. We used a Jmol [11, 20] script to program-matically convert over 30,000 crystal structures into 3D printable les. The 3D printable le datasets and pro-grammatic scripts are provided openly to the community on gshare. A Jmol 3D Print website was created based on the COD search interface that allows users to discover

the 3D printable crystal structure le dataset via text, SMILES, reference, elements, and cell parameters. In addition, users can create a custom 3D printable le (STL or VRML) from their own CIFs and for any structure in the COD using the JSmol implementation within the 3D Print website.

Methods

Programmatic processing of crystal structures into 3D printable les presents several challenges. One such challenge is that many crystal structures contain counterions and solvent molecules. These companion molecules would ideally be removed from the 3D printable le for two reasons: (1) most often the counterions and solvent molecules are not of direct interest for 3D printing; and (2) 3D les with multiple independent objects can be difcult for 3D printers to fabricate. While removal of counter ions and solvent molecules can be readily achieved with cheminformatics toolkits [2123], this procedure has not yet been applied in the preparation of small molecule and extended solid 3D printable crystal structure les.

Another challenge with processing crystal structures into 3D printable les arises with the initial representation of the crystal structure within crystallographic viewers. By default, many crystallographic viewing programs only load the unique atom positions in the crystallographic data, namely the asymmetric unit. This can lead to displaying only a fraction of a molecule or a very small segment of an extended solid. To gain a complete representation of a molecule or a complete unit cell of an extended solid, the software must be instructed to apply the appropriate space group symmetry operations to the asymmetric unit [24]. While the asymmetric unit is a useful representation, we suspect in most cases, those interested in 3D printing crystal structures would prefer the symmetry operations applied to the crystallo-graphic data. Applying the symmetry operations ensures that the representation will present either one or more whole molecules or a complete unit cell fragment of an extended solid. Moreover, with extended solids it is often useful to pack the unit cell; that is, display all atoms and bonds that t only within one or more unit cells.

To overcome the aforementioned challenges of processing crystal structures into 3D printable les, we developed a Jmol script, Jmol3DP (Fig.1). When the Jmol3DP script is executed with a batch le from a DOS command line, the script will automatically process any number of CIFs into VRML les using the following method:

1. A CIF crystal structure from a text le list is loaded in Jmol as a ball and stick representation and the symmetry elements appropriate to the space group

Scalfani et al. J Cheminform (2016) 8:66

Page 3 of 8

are applied. Note: For extended solids, the Jmol loading option PACKED was added (i.e. load @ {alist[i]} {1 1 1} PACKED). If processing biomolecules, the Jmol STRUTS command should be used to add necessary mechanical support to the model.2. The molecule (or extended solid network) with the greatest number of atoms within the displayed structure(s) is selected.

3. All other molecules are deleted (i.e. duplicate molecules with same number of atoms and molecules with fewer atoms).

4. All axes and unit cell guides are deleted.5. The largest molecule (or extended solid network) that was selected in step (2) is scaled ten times its original size and saved as a VRML le.

6. Steps (1)(5) are repeated for the next CIF.

The ball and stick representation was selected as it one of the most popular visual representations of chemical structures. Scaling was necessary to generate structures

with dimensions on the order of 110cm length scales (in each x, y, z, axes), which are appropriate for most 3D printers.

To test our methods, a subset of CIFs within the COD were selected. This included 31,239 CIFs originating from a dataset of the COD that was also incorporated into ChemSpider [25]. This collection of CIFs was a convenient and sizable dataset composed of small organic molecules and inorganic extended solids. All CIF data within the COD is in the public domain and therefore allowed us to reuse, adapt, and share the 3D printable les created from COD data [13].

After processing the CIFs into VRML les using Jmol scripting, additional le preparation steps were added to increase the compatibility of the 3D les across numerous 3D printer software applications and 3D printing services. AccuTrans 3D (v2014-1-0) [26] was used to remove high level objects (cone, sphere, and cylinders) from the VRML les and to convert a subset of the les (11,732) into STL format. Lastly, the STL les were run through an extended batch repair process using Netfabb

Scalfani et al. J Cheminform (2016) 8:66

Page 4 of 8

Professional [27]. Both AccuTrans 3D and Netfabb Professional have intuitive graphical user interfaces for batch processing.

For the 3D test prints, we used a Stratasys uPrint SE with ABS P430 XL Ivory model material and SR-30 XL soluble support material. Catalyst EX software was used to create the toolpath. Settings included SMART support structure and high density ll. Approximate print time was 4h for each 3D printed crystal structure. The SR-30 XL soluble support material was removed with a solution of P400-SC Waterworks Soluble Concentrate (a caustic basic solution). The 3D printed crystal structures containing the associated support material were submerged in the P400-SC basic solution for 3h at 80C to selectively dissolve the support material. The completed 3D printed crystal structures absent of any support material were then rinsed three times with water.

Results anddiscussion

The Jmol3DP script described here was successfully used to programmatically convert CIFs from the COD into 31,239 printable VRML les and 11,732 printable STL les. Our original process involved post-processing of VRML les created by Jmol 14.1 using AccuTrans 3D and Netfabb Professional. This post-processing was necessary in order to remove the high level VRML 2.0 objects (cone, cylinder, sphere) [28] that Jmol 14.1 created and to convert the VRML to STL format, which many 3D printers require. In the process of review of this manuscript, Jmol was developed further, and Jmol 14.6.4 now exports VRML, X3D (the XML version of VRML), and STL les that can be submitted directly to 3D print services and 3D printer software without the need for any additional post-processing.

We note that success of STL and VRML les for 3D printing relies upon careful attention to details. All rendered parts must be closed surfacesno open ends of cylinders used for bonds; no missing triangles in the geodesics used for atoms. The counter-clockwise winding of all triangles must be designed to project all triangle normals in the proper direction. Innitely thin surfaces (which appear properly on screens and in images) must be given some thickness. Jmols (v14.6.4) STL and VRML exporters take care of all of this automatically. In addition, in Jmol, care has been taken to ensure that no two identical objects are placed in the same location. For example, where two bonds come together, the geodesics used to form the rounded endcaps of the cylinders are only generated once, not twice. Some more advanced features of Jmol, such as slabbed (unclosed) isosurfaces, dot surfaces, and labels, are not supported due to these restrictions.

Two representative examples of CIFs processed with the Jmol3DP script are shown in Fig.2. The crystal structure of 4-[(2E)-2-(4-Pyridinylmethylene)hydrazino]-2,2:6,2-terpyridine viewed in Jmol, as loaded with the symmetry operations applied, contains two identical molecules of the pyridinyl terpyridine derivative (diering in orientation), two chloroform molecules, and two water molecules (Fig. 2a). After running the Jmol3DP script, only one pyridinyl terpyridine molecule is selected, scaled, and exported as a VRML le. All other molecules and axes were automatically deleted (Fig.2b). In the extended solid example, Pb3O4, the crystal structure is loaded in Jmol with symmetry and packing applied (Fig. 2c). After running the Jmol3DP script, the entire Pb3O4 packed structure is selected, scaled, and exported as a VRML le; the axes and unit cells were automatically deleted (Fig.2d). In both examples, a VRML 3D model le was created containing only the single desired chemical structure.

The quality of the programmatically converted 3D printable crystal data is satisfactory. We manually veried a random set of 115 3D printable les and found 90% accuracy. Accuracy here is being dened as producing the expected digital 3D printable molecule or extended solid from the original CIF. The Jmol3DP script automatically deleted small solvent and counterions, and ultimately produced a 3D printable le containing a single molecule or extended solid with high delity to the original selected structure within the crystallographic data. Inaccuracies and undesired results in the data mostly arose from co-crystallized structures where both molecules could be of interest for 3D printing. With our current 3D printing conversion method, only the larger molecule would be available as a 3D printable le, and as a result, the smaller co-crystallized molecule is deleted. A similar situation occurred with large or complex counterions. The molecule of interest contained fewer atoms than its counterion, and as a result was mistakenly deleted. Lastly, racemic mixtures present a further complication for processing. Enantiomers contain the same number of atoms. With the current Jmol3DP script, only one of the enantiomers would be selected at random. At this time it is not clear how to circumvent these challenges with programmatic multi-le processing of crystal structures for 3D printing. One method may be to write a Jmol script that produces a separate 3D printable le for each molecule in (or partly in) the crystal structure unit cell. This would preserve all structures in the original CIF and allow the end-user to select the molecule/extended solid of most interest. Until such a method is developed, it will be necessary for the community to help manually curate 3D printable crystal structure data.

Scalfani et al. J Cheminform (2016) 8:66

Page 5 of 8

Tests with the NIH 3D Print Exchange web tools [15] conrmed that its process, in contrast to ours, retains all molecular structures in the CIF data, makes no attempt to remove solvent molecules or counterions from small molecules and extended solids, does not apply symmetry, and does not select a dened subset for representation. If a user wishes to selectively 3D print molecules or atoms, that selection must be done manually before uploading to the system. Similarly, the Cif2VRML application does not remove solvent molecules (except for biomolecules) or counterions, does not apply symmetry, and does not automatically select a particular structure for representation. Several of these options are reported to be included in a future release of the Cif2VRML application [18].

The NIH 3D Print Exchange web tool, the Cif2VRML application, and the Jmol3DP script reported herein all have utility and unique benets that help to advance the creation of 3D printable crystal structures. Looking toward the future, we reiterate, that the ideal solution to programmatically convert crystal structures into 3D printable les may be to produce separate 3D printable les for each molecule/extended solid in the CIF le so that no loss of structural representation occurs and users can select the structure of interest to 3D print.

The available 3D printable crystal structures created in this work will allow researchers and educators to quickly locate a large collection (>30,000) of crystal structures that are 3D print ready; no further modications to the crystal structure or le format should be needed. In

Scalfani et al. J Cheminform (2016) 8:66

Page 6 of 8

addition to providing the entire collection on gshare, a Jmol 3D Print website was created that allows for discovery of the entire collection of 3D printable les created in this work. The STL and VRML les for the compounds were placed on the chemapps.stolaf.edu server in a set of zip les based on COD id 1000000-1999999, 2000000-2999999, etc. The list of ID numbers for all models was also included. The website works by making two calls to COD from the Figshare Only button. It rst requests ID numbers for a users search involving SMILES, text, reference, elements, or cell parameters. The returned list of hits from COD is then ltered using the ID list for the 31,239 crystal structures, and this ltered list is sent back to COD, now requesting a full report for only these specic structures. Up to 100 structures are returned. The results are then displayed using a modied form of the COD search results page where users can download available VRML or STL les (Fig.3).

If users wish to create a custom 3D printable le that is not available in the predened le dataset created in this work, an option is available to search the entire COD and design custom 3D print les using JSmol, the non-Java HTML5 version of Jmol. When a search is executed on the Jmol 3D Print website using the All Structures button, hits from the COD are returned in a list with both

the predened VRML and STL les, if available, and a 3D link that opens a JSmol Crystal Symmetry Explorer website to allow custom creation of 3D print les. This JSmol Crystal Symmetry Explorer allows various symmetry load options, molecule selection options, and 3D printable STL and VRML le saving options. Moreover, users can even upload their own CIF data and scripts for custom conversion of their CIFs into 3D printable les (Fig.3).

Creation of VRML and STL les within the JSmol Crystal Symmetry Explorer is a local operation executed in JavaScript within the web browser; there are no server processes necessary for preparing the 3D visualization and subsequently saving a VRML or STL le. As such, if users open the JSmol Crystal Symmetry Explorer from their local machine (from a copy of the JSmol distribution), CIF data can be readily processed into VRML and STL les without any connection to the internet. This process is in contrast to other web-based 3D crystal structure le converters such as the NIH 3D Print Exchange, which do this processing on a server and thus require an internet connection. Furthermore, since the capability of creating VRML and STL les is built into all versions of Jmol starting with version 14.6.4 (both Java and JavaScript), it will be possible to create 3D

Scalfani et al. J Cheminform (2016) 8:66

Page 7 of 8

printable les of models displayed in Jmol at any website that has upgraded their implementation of Jmol to this version or higher (e.g. Crystallography Open Database, Protein Data Bank, Inorganic Crystal Structure Database, Cambridge Structural Database and more). To do so, users need only to click on the JSmol logo or right-click on the model window to call up a pop-up menu and select File > Export > Export VRML 3D Model or File>Export>Export STL 3D Model.

We have 3D printed several STL structures in-house from the produced dataset and a VRML le outsourced to Shapeways, conrming our programmatic methods successfully created 3D printable les (Fig.4) [29]. Users can immediately load the 3D les into their 3D printing tool path software and begin fabricating the crystal models. Because many of the crystal models are complex, the use of a removable/dissolvable support material in the 3D printing process is recommended to achieve optimal results and quality of the printed model.

Conclusions

Jmol scripting was used to programmatically convert 31,239 crystal structures (CIFs) from the COD into 3D printable les. These 3D printable les and Jmol scripts are available openly to the community on gshare and on a custom Jmol 3D Print website based on the COD search interface that allows users to discover the 3D printable crystal structure datasets via text, SMILES, reference, elements, or cell parameters. The Jmol 3D Print website also allows users to convert their own CIFs and structures in the COD (>350,000) into a 3D printable le using JSmol, the non-Java HTML5 version of Jmol. Our

hope is that these 3D printable les will be deposited into popular databases such as PubChem [30, 31], ChemSpider [32, 33], and the Crystallography Open Database [13]. There is also future potential for implementing the JSmol Crystal Symmetry Explorer into the Crystallography Open Database [13] and other crystallography databases, allowing users to easily customize and create 3D printable crystal structure les. Current crystallography databases and websites that use Jmol can upgrade to v14.6.4 or later to allow users the option of exporting 3D printable VRML and STL les. We highly encourage the community to use the 3D printable crystal structure dataset, Jmol scripts, and Jmol 3D Print and Crystal Symmetry Explorer websites to advance their research, education, and outreach eorts with 3D printed molecular models.

Abbreviations

3D: three-dimensional; CIF: crystallographic information le; STL: stereolithography le; VRML: virtual reality modeling language; COD: Crystallography Open Database; NIH: National Institutes of Health.

Authors contributions

VFS, AJW, RMH, and VT conceived the study. VFS, JML and JEB preparedthe 3D printable les and 3D printing tests. RMH and VFS developed the Jmol script. RMH added STL and VRML export capabilities to Jmol 14.6 and developed the Jmol 3D Print website. VT, AP, and KK processed the les from the Crystallography Open Database that were used in this article. VFS, AJW, RMH, and JEB wrote the manuscript. All authors read and approved the nal manuscript.

Author details

1 University Libraries, Rodgers Library for Science and Engineering, The University of Alabama, Tuscaloosa, AL 35487, USA. 2 ChemConnector, 513 Chestnut Grove Court, Wake Forest, NC 27587, USA. 3 Royal Society of Chemistry, 14900 Forest Landing Circle, Rockville, MD 20850, USA. 4 Department of Chemistry, St. Olaf College, Northeld, MN 55057, USA. 5 Department of Chemical

Scalfani et al. J Cheminform (2016) 8:66

Page 8 of 8

and Biological Engineering, The University of Alabama, Tuscaloosa, AL 35487, USA.

Acknowledgements

VFS thanks The University of Alabama Libraries for support of this work and Ms. Rosanne Rumstay for her assistance with 3D printing. JEB and JML thank the National Science Foundation Research Experiences for Undergraduates program (EEC-1358750).

Competing interests

RMH is the lead developer of Jmol.

Availability of data and materials

The datasets and scripts supporting the conclusions of this article are available in the gshare repository, https://dx.doi.org/10.6084/m9.figshare.c.3302859.v6

Web End =https://dx.doi.org/10.6084/m9.gshare.c.3302859. https://dx.doi.org/10.6084/m9.figshare.c.3302859.v6

Web End =v6 . The 3D le dataset can be searched and discovered on the Jmol 3D Print website at https://chemapps.stolaf.edu/jmol/3dprint/

Web End =https://chemapps.stolaf.edu/jmol/3dprint/ . The JSmol Crystal Symmetry Explorer is available at https://chemapps.stolaf.edu/jmol/jsmol/jcse/explore.htm

Web End =https://chemapps.stolaf.edu/jmol/jsmol/jcse/ https://chemapps.stolaf.edu/jmol/jsmol/jcse/explore.htm

Web End =explore.htm .

Dedication

Dedicated to the memory of Prof. Jean-Claude Bradley. We celebrate the future of open science and open data.

Received: 15 August 2016 Accepted: 16 November 2016

References

1. Scalfani VF, Vaid TP (2014) 3D printed molecules and extended solid models for teaching symmetry and point groups. J Chem Educ 91:11741180

2. Chen T-H, Lee S, Flood AH, Miljanic OS (2014) How to print a crystal structure model in 3D. CrystEngComm 16:54885493

3. Herman T, Morris J, Colton S, Batiza A, Patrick M, Franzen M, Goodsell DS (2006) Tactile teaching. Exploring protein structure/function using physical models. Biochem Mol Biol Educ 34:247254

4. Kitson PJ, Macdonell A, Tsuda S, Zang H, Long D-L, Cronin L (2014) Bringing crystal structures to reality by three-dimensional printing. Cryst Growth Des 14:27202724

5. Moeck P, Stone-Sundberg J, Snyder TJ, Kaminsky W (2014) Enlivening 300 level general education classes on nanoscience and nanotechnology with 3D printed crystallographic models. J Mater Educ 36:7796

6. Rodenbough PP, Vanti WB, Chan S-W (2015) 3D-printing crystallographic unit cells for learning materials science and engineering. J Chem Educ 92:19601962

7. Grazulis S, Sarjeant AA, Moeck P, Stone-Sundberg J, Snyder TJ, Kaminsky W, Oliver AG, Stern CL, Dawe LN, Rychkov DA, Losev EA, Boldyreva EV, Tanski JM, Bernstein J, Rabeh WM, Kantardjie KA (2015) Crystallographic education in the 21st century. J Appl Crystallogr 48:19641975

8. Pymol: a molecular visualization system on an open source foundation, maintained and distributed by Schrdinger. https://www.pymol.org/

Web End =https://www.pymol.org/

9. RasMol and OpenRasMol. http://www.openrasmol.org/

Web End =http://www.openrasmol.org/

10. Mercurycrystal structure visualisation, exploration and analysis made easy. http://www.ccdc.cam.ac.uk/solutions/csd-system/components/mercury/

Web End =http://www.ccdc.cam.ac.uk/solutions/csd-system/components/ http://www.ccdc.cam.ac.uk/solutions/csd-system/components/mercury/

Web End =mercury/

11. Jmol: an open-source Java viewer for chemical structures in 3D. http://jmol.sourceforge.net/

Web End =http:// http://jmol.sourceforge.net/

Web End =jmol.sourceforge.net/

12. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE (2004) UCSF Chimeraa visualization system for exploratory research and analysis. J Comput Chem 25:16051612

13. Crystallography Open Database. http://www.crystallography.net/

Web End =http://www.crystallography.net/ 14. Grazulis S, Chateigner D, Downs RT, Yokochi AFT, Quiros M, Lutterotti L, Manakova E, Butkus J, Moeck P, Le Bail A (2009) Crystallography Open Databasean open-access collection of crystal structures. J Appl Crystal-logr 42:726729

15. NIH 3D Print Exchange. http://3dprint.nih.gov/

Web End =http://3dprint.nih.gov/ 16. NIH 3D Print Exchange: 3D model scripts. https://github.com/niaid/3Dmodel_scripts

Web End =https://github.com/ https://github.com/niaid/3Dmodel_scripts

Web End =niaid/3Dmodel_scripts

17. Coakley MF, Hurt DE, Weber N, Mtingwa M, Fincher EC, Alekseyev V, Chen DT, Yun A, Gizaw M, Swan J, Yoo TS, Huyen Y (2014) The NIH 3D print exchange: a public resource for bioscientic and biomedical 3D prints. 3D Print Addit Manuf 1:137140

18. Kaminsky W, Snyder T, Stone-Sundberg J, Moeck P (2014) One-click preparation of 3D print les (*.stl, *.wrl) from *.cif (crystallographic information framework) data using Cif2VRML. Powder Dir 29:S42S47

19. Blender. https://www.blender.org/

Web End =https://www.blender.org/ 20. Hanson RM (2010) Jmola paradigm shift in crystallographic visualization. J Appl Crystallogr 43:12501260

21. Steinbeck C, Han Y, Kuhn S, Horlacher O, Luttmann E, Willighagen E (2003) The Chemistry Development Kit (CDK): an open-source java library for chemo- and bioinformatics. J Chem Inf Comput Sci 43:493500

22. epam Indigo Toolkit. http://lifescience.opensource.epam.com/indigo/

Web End =http://lifescience.opensource.epam.com/indigo/ 23. RDKit: Open-Source Cheminformatics Software. http://www.rdkit.org/

Web End =http://www.rdkit.org/ 24. Massa W (2004) Crystal structure determination. Springer, Berlin25. ChemSpider Search for Data in Crystallography Open Database (COD). http://www.chemspider.com/Search.aspx%3fdsn%3dCrystallography%2bOpen%2bDatabase%2b(COD

Web End =http://www.chemspider.com/Search.aspx?dsn=Crystallography+Open + http://www.chemspider.com/Search.aspx%3fdsn%3dCrystallography%2bOpen%2bDatabase%2b(COD

Web End =Database+(COD )

26. MicroMouse Productions: AccuTrans 3D object conversion. http://www.micromouse.ca/

Web End =http://www. http://www.micromouse.ca/

Web End =micromouse.ca/

27. Netfabb Professional. http://www.netfabb.com/professional.php

Web End =http://www.netfabb.com/professional.php 28. The Virtual Reality Modeling Language Concepts. http://www.web3d.org/documents/specifications/14772/V2.0/part1/concepts.html

Web End =http://www.web3d. http://www.web3d.org/documents/specifications/14772/V2.0/part1/concepts.html

Web End =org/documents/specications/14772/V2.0/part1/concepts.html

29. Scalfani VF, Williams AJ, Hanson RM, Bara JE, Day A, Tkachenko V (2014) Accessing 3D printable chemical structures online. American Chemical Society, Washington, pp CINF-125

30. PubChem. https://pubchem.ncbi.nlm.nih.gov/

Web End =https://pubchem.ncbi.nlm.nih.gov/ 31. Bolton EE, Wang Y, Thiessen PA, Bryant SH (2008) PubChem: integrated platform of small molecules and biological activities, Chap 12, vol 4. In: Ralph AW, David CS (eds) Annual reports in computational chemistry. Elsevier, Amsterdam, pp 217241

32. ChemSpider. http://www.chemspider.com/

Web End =http://www.chemspider.com/ 33. Pence HE, Williams A (2010) ChemSpider: an online chemical information resource. J Chem Educ 87:11231124