BioBlender: Blender for Biologists

BioBlender: Blender for Biologists(bioblender.eu)

109 points by rocco 13 years ago | 28 comments

clarle 13 years ago |

Sounds really cool!

We usually just use PyMOL [1] in our lab, but it looks like BioBlender takes PyMOL, Blender, and a few other structural biology and builds on top of both of them.

It looks like the biggest contribution from BioBlender is that it takes properties such as electrostatic potential and combines it with Blender's physics engine to better simulate movement when creating molecular animations. Most of our work usually involves just examining the 3D structure of a protein, so I don't usually create animations, but this is definitely a really useful tool to do so if any lab might ever need to do something like that.

We really do need more people dedicated to just purely better visualization software in the biological community. A lot of interesting hypotheses often start from just rotating around 3D structures, and making it easier to do so can only help.

[1]: http://pymol.org/

pedrobeltrao 13 years ago | |

I agree with the need for better visualization tools in biology. There is an interesting group/project started at UCSF and headed by Graham Johnson (www.grahamjlab.org/). Some of these tools are still closer to art than science but I think it is really worth the investment. It is clear that these will help us think about biological objects.

thundara 13 years ago | |

What worth would the 3d simulations produced by Blender's physics engine be? I'd imagine that they wouldn't even compare to those produced by full MD simulations, which are already of debatable correctness depending on the size of the system, force field + parameters, time step, solvent, etc...

kayhi 13 years ago | |

Great to see some other crystallographers in the crowd, also worth checking out is Discovery Studio Visualizer.

goldfeld 13 years ago |

I have been wanting to get into bioinformatics, but there's something I miss coming from webdev: the ability to create something over a few weeks. I don't mean I want to discover something new, and I'm not sure I'd want to get into research at all. I wanted to be able to experiment and see stuff show up on a screen related to the structures and formulas I'm learning. Some fun and practical, even if not very useful (games?), project I can tackle while learning bioinformatics. Is there such a thing?

clarle 13 years ago | |

Rosalind [1] (previously posted on HN) is pretty good for something like a Project Euler for bioinformatics. The problems are pretty similar to the introductory computational biology course at my school for upper-year undergrads and first-year grad students.

It's a really good way to get a good overview of the field, and what some of the problems have been already solved in bioinformatics. It's definitely best to combine it with a more traditional textbook such as "Introduction to Computational Biology" (Haubold and Wiehe) or "Biological Sequence Analysis" (Mitchison, Krogh, Durbin, and Eddy).

Hope that helps!

[1]: http://rosalind.info/

archgoon 13 years ago | | |

That is absolutely fantastic! Kudos to UCSD and Saint Petersburg Academic University for creating such an awesome educational resource! And thank you for pointing me to it!

daniel-levin 13 years ago | | |

+1 for this.

Rosalind is a great way of improving your knowledge of biology and CS. It is quite cool to see computer science and mathematics problems and techniques applied to bioinformatics.

Evgeny 13 years ago | | |

Rosalind [1] (previously posted on HN) is pretty good for something like a Project Euler for bioinformatics

I think I have a new hobby!

timr 13 years ago | |

If all you want to do is play around with molecules on a screen, check out PyMol.

But here's the rub: if you want to learn "bioinformatics", you're going to have to accept that most of the work is inherently not visual. I worked in about the most shiny, graphical part of biology possible (protein structure), and maybe only 20% of the work involved spinning molecules on a screen.

The real work of bioinformatics involves thinking very hard about problems that are deeply mathematical. The visualization part is strictly secondary -- though it gets the most attention, for obvious reasons.

anigbrowl 13 years ago |

(Applause)

I have always admired Blender in general, and interfacing it it to load molecular structures like this is genius. I hope it's the first of many such explorations. Although I'm not a biologist this is cool enough that I would install blender again just to have fun with making neat images.

shiven 13 years ago |

Great to see so many structural biologists and crystallographers on HN!

PyMOL has been progressing leaps and bounds over the last few years under Jason Vertrees guidance at Schrodinger LLC. So do give it a look if you are excited about this stuff. The development model is also quite nice, in that the most avant-garde features are released in the 'incentive' version (costs money) but then are released into the open source version after the next version update, sometimes even earlier.

It has been a personal interest of mine to use Blender and PyMOL together for visualization and graphics[0].

Good luck to BioBlender and thanks for sharing on HN!

[0] http://pymolwiki.org/index.php/Blender

epenn 13 years ago |

This looks interesting but doesn't appear to have been updated in a year and a half unfortunately.

cing 13 years ago | |

It's my understanding that this project has been superseded by ePMV (http://epmv.scripps.edu/).

_ZeD_ 13 years ago |

Cool!

how is this comparable with avogadro?

http://avogadro.openmolecules.net/wiki/Main_Page

car 13 years ago |

Does anyone know how to 3D-print protein structures?

archgoon 13 years ago | |

You do not 3D-print protein structures (unless you mean 3D models of proteins in which case ignore the rest of this post). Not in the sense of "here's an arbitrary position of atoms, give me a molecule". There's a reason why Folding@Home takes a lot of processing power. It's very hard to figure out how chains of amino acids will interact with themselves.

The best you can do is do gene synthesis using a method called Oligonucleotide Synthesis

http://en.wikipedia.org/wiki/Oligonucleotide_synthesis

Using this technique, you can take an arbitrary gene sequence (ATGC sequence) and get a plasmid out of it. This will specify a sequence of amino acids that will fold into a protein. Then you just have to get a bacterial host to take it up into it's DNA and then make sure the chemical conditions of the host are right to generate your protein (you can't synthesize a protein if there aren't the right types of Amino Acids floating around, for example).

Companies like Genscript (which I found by randomly googling) will manufacture a specified gene, and send you back frozen plasmids, which you can then inject into your favorite bacterial host.

http://www.genscript.com/gene_synthesis.html

jamessb 13 years ago | | |

Depending on what exactly you want, you may be able to use a cell free translation system rather than transfecting a bacterial host (or yeast, or whatevever else).

Short peptides can also be produced by solid-phase synthesis: there are automated machines (peptide synthesisers) that produce a peptide with any arbitrary sequence from the constituent amino acids.

car 13 years ago | | |

Thanks for the reply, but of course this question was in the context of the parent.

Just fyi, mammalian genes are usually expressed in mammalian cells, not bacteria, because of posttranslational modification.

jostmey 13 years ago |

It looks really cool, but I cannot figure out how to use the program to open up crystal structures (PDB file format) on my desktop.

pvaldes 13 years ago | |

You can glue this with a python script probably. Then, call this script from blender converting your data to a readable result. Untested, only an idea.