[3dem] 3D printing

[David Bhella]

Hi Cindi,

I too have used shapeways to print viral capsids solved by cryo-EM - you can see the results and representative costs here:
https://www.shapeways.com/shops/viral_art

There are two basic issues to overcome when preparing a model for 3D printing. Firstly if you create an isosurface using structural biology visualisation software, there will typically be vast numbers of polygons (sometimes many millions), this is way too complex for the software that runs most 3D printers. Shapeways has an upper limit on the number of polys in a model, when I last prepared models the limit was 500k but according to their website the limit is now 1 million. Secondly the surfaces created in visualisation software can have problems that mean that they couldn’t (can’t) exist in the real world e.g. they may not be manifold i.e. the volume basically isn’t water tight. Sometimes face normals which define inside and outside are pointing the wrong way.

Shapeways have tutorials on how to fix these problems using free software that I found really helpful:
https://www.shapeways.com/tutorials

In summary:
I surface render my map in UCSF Chimera and export it as VRML (although I noticed that there is now an option to export STL directly)
This is then imported in Meshlab (http://meshlab.sourceforge.net)
To reduce the polycount use:
Filters > Remeshing, simplification and construction > Quadratic Edge Collapse Detection
Shapeways recommend selecting all options except 'simplify only selected faces’ and a quality threshold of 1, but I find the defaults work quite well.  The routine opens with a suggested target polygon count of half the current number. I think it is better to stick with this and iterate the process until you reach an acceptable number rather than asking it to go straight from 5 million to 500k.

When you are happy export the surface as .STL format.

The STL file is then ‘repaired' using NetFabb studio basic (http://www.netfabb.com)
This has routines that will analyse and repair your model to make it suitable for printing:
First use Part > Scale to set the size of the printed model.
Then select ‘Repair’ (a red cross) and press automatic repair (use default), finally apply the repair and remove the old part.
The final step is to make sure you only have one shell for each part (usually only 1 per model). This gets rid of any flying density etc.
With the repair menu still visible use the select shells tool (a green cube button) to select your model.
Toggle selection (the button is a green triangle in front of a blue triangle) to select all the other shells. Then Remove selected triangles (a grey triangle button with a small cross on). Now in the information panel at the bottom right it should say that you have 1 shell only.

Export your fixed part as .STL and it should be suitable for 3D printing. If you do decide to print using shapeways they run checks on your model and will tell you if there are still problems.

Hope this helps!

D.

 
Dr David Bhella
MRC - University of Glasgow Centre for Virus Research 
8 Church Street
GLASGOW
G11 5JR

Telephone:  0141-330-3685
Skype: d.bhella

Virus structure group on Facebook: https://www.facebook.com/CVRstructure
Molecular Machines - Images from Virus Research: http://www.molecularmachines.org.uk

CVR website: http://www.cvr.ac.uk
CVR on Facebook: https://www.facebook.com/centreforvirusresearch 



















On 16 Jan 2014, at 20:17, Andrey Malyutin <agmalyut at umail.iu.edu> wrote:

> Hi all,
> 
> I've used shapeways.com to print several 2 inch models of viral particles. I've started with .mrc files and ended with uploading .stl files to their servers (converted in Chimera as described by Gabe). I've had to 'repair' my initial models to be acceptable for their printers, which required a few additional pieces of free software and just a bit of time. All of the models so far that I've printed were single piece and single colour. They can, however, print coloured (depending on input model format), multi-piece models, and in a large range of materials. The prices have been reasonable. They are dependant on the volume of the model and can be made cheaper by removing internal density where viable. Additionally, educational discount is available. 
> 
> Andrey
> 
>                                                                       
> Andrey Malyutin
> Graduate Student
> Indiana University,
> Department of Chemistry
> 800 E. Kirkwood Av.
> Bloomington, IN 47405
> Phone: (269) 352-6880
> 
> 
> 
> On Thu, Jan 16, 2014 at 2:10 PM, Cindi Schwartz <cindi.schwartz at colorado.edu> wrote:
> Hi all,
> 
> Has anyone out there taken an IMOD model or similar 3D model from their averaging data and printed it to 3D printer? Can you give details from the starting model format (imod?) to the final model format (???)? What kind of printer you used among the many manufacturers out there. How many colors, was it a single piece or multi-piece, etc? Was it owned by your University or did you do it through a 3rd party printer?
> 
> Thanks,
> Cindi
> 
> ****************************
> Cindi L. Schwartz
> University of Colorado
> Dept. MCDB
> 347 UCB
> Boulder, CO  80309
> O: 303-492-7980
> F: 303-735-0770
> http://bio3d.colorado.edu
> ****************************
> 
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
> 
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem