MOLGRAPH3D

MGLTools PMV - Gelato

2009-08-05T17:12:00.001-07:00

I have recently discover the MGLTools-PMV software and its amazing number of possibility...(see http://mgltools.scripps.edu/ from http://mgl.scripps.edu/ ). Playing with, its seems really easy to add plug-in and new functionality. Like I did here to integrate the nvidia rendering engine gelato as a plug-in accessible for the file menu of pmv :

A simple capture from pmv (with antialiasing) look like this (and this is a realy simple one...):

Using gelato, I can select which object inside Pmv I want to render with gelato, and choose an apropriate shader. Currently, the plug-in only handle CPK and ball&stick style.
The ambiant occlusion is still a good choice to increase the quality of picture :

Then I just have to combine this two picture using imagemagic, gimp or either the vision software (part of the mgl tools... http://mgltools.scripps.edu/packages/vision ), i will try to show something later with this.

IBISC gift

2009-07-14T07:24:00.000-07:00

From skecth to final crystal printing picture. The idea was to unfold the dna double strand around a gene netowrk ....

metaBALL blender

2009-03-05T19:04:00.001-08:00

PDB -> metaball blender -> Kerkythea

home protein contemporary art

2008-07-23T03:19:00.000-07:00

Same protein as previous, but this time put in a room with daylight (the scene comes from http://www.mrcad.com/?p=166)...I don't know if I would like such art in my appartment ...

protein breakfast & lunch

2008-07-21T08:24:00.000-07:00

I dedicated this picture for all people who think, live, wakeup etc... with protein structure. The scene is come from http://www.3dm3.com/tutorials/mr/. The protein come from the PDB (http://www.rcsb.org/)...

better test (SSS)

2008-07-18T08:13:00.000-07:00

a new test of this green material. Less rendering time on a more powerfull computer...(25min)...
This can be for sure optimised...but I like the way the atom appear behind the surface...(clic to enlarge).

rendering test

2008-07-18T01:41:00.000-07:00

Its just a test, but it took more than 4hours on my poor computer...(click to enlarge the picture)

small heat-shock protein

2008-06-25T06:44:00.000-07:00

During my last post-doc i had work on small heat shock protien assembly using some nice tools for doing symmetrical restrained computational docking. Thus, as usual i try to produce some nice graphism, and here is a exemple of the HSP16.5 24mer respresented as cartoon ribbon fitted in the molecular shape computed in the molsoft icm package.

Pushing the Boundaries of Biomolecular Simulation

2008-06-07T01:16:00.000-07:00

I am really happy to show the poster that I will present during the Pushing the Boundaries of Biomolecular Simulation meeting (http://web.mac.com/laveryr/ISQBP08/Welcome.html). Many important people from the molecular modeling fields will be present and I hope they will enjoy my litlle poster...

Proteins Blocks

2008-05-27T07:42:00.000-07:00

A litle attempt to get nices pictures illustrating proteins blocks. You can find more details on the web site of Alexandre De Brevern.

hand shadow

2008-04-29T14:20:00.000-07:00

For me, beautifull results happens accidently.Here is an exemple of hasard. Thats fun this face appearing in the background, like hand shadow coming from the "Nightmare Before christmas" movie.

2008-04-24T09:32:00.000-07:00

Nice video that i have found on a friends web site (volta.fr):

"The video was created by me (Michael Wesch), working alone from my house in St. George, Kansas. I used CamStudio for the screen captures and Sony Vegas for the panning/cropping/zooming animations. Someday I might make a video tutorial for those who are interested."

Protein-Protein Interaction Website

2008-04-22T00:11:00.000-07:00

The most relevant place to find information about protein-protein is the JBC Protein-Protein Interaction Website : http://www.imb-jena.de/jcb/ppi/

You should find all the litterature, software, database, company, meeting, course, tutorial etc...about protein-protein interaction. Put it on your favorites if it is not the case!

bezier curve for protein ruban

2008-04-18T05:49:00.000-07:00

Thats dones! I am quite happy to finally get a nice protein worm representation directly in blender. But there is still the coloring method problem...I have to associate vertex with material according physical property...And try to display arrow and helix for secondary structure type...
(thanks to the blender tutorial here :http://jmsoler.free.fr/didacticiel/blender/tutor/cpl_celticknot.htm and the API help)

Can a Biologist Fix a Radio?

2008-04-18T01:19:00.000-07:00

This is not news, and I am not the first to post this, but as much it is diffuse the better it is!
here it is the : Can a Biologist Fix a Radio? (PDF - 469 KB), by Yuri Lazebnik. It’s all about how experimental biologists would try to find out how a radio works. The article suggests that an engineering-like approach would lead to a deeper understanding of how cells work (last time i have saw it reasonable deviation, also on Freelancing Science and God Plays Dice).

blender protein ribbon

2008-04-15T15:08:00.000-07:00

I continue to play around python and blender API. Thats not easy, but copy/past technics gives some result...Thus, i currently explore the curve system. Each point have the coordinate of an atom, and define a nurb curve. Thats not realistic, I had to improve it to use bezier curve (BezTriple class).

The result (simple, not sophiticated as it is just modeling test, and I have no idea of what color is good or not...dyschromatopsy!):

dsimb

2008-04-08T08:18:00.001-07:00

DSIMB - Dynamique des Structures et Interactions des Macromolécules Biologiques

2008-03-29T01:35:00.000-07:00

My laboratory have his new web site on-line at this url : http://www.dsimb.inserm.fr/.
Currently under construction, this web site gives some information on the different research thematics handle by the lab. Except mine as i already didn't send my contribution to the webmaster!!.

blender armatures and protein's cube

2008-03-26T05:01:00.000-07:00

As I work around molecular graphism, i am trying all methods that can be available. This time it is the powerfull blender and his python API. Generally, I first generated the protein object using an ancilliary software like VMD or ICM, but i was thinking about a direct way to use PDB files in 3d modeling package. There is some plug-in for blender, 3DSmax or maya, but they just get atom as sphere. In the other hand you can easily find on the web some python API dedicated for molecular modeling. Thus, i am currently trying this path. And here are my first result. All sphere or cube are group in amino acid. And you can select any subselection available for protein (amino acid, atom, secondary structure, etc..).
The script is at his begining so i cannot put it here, but when or if I have a clean version i will post it for sure. Because, for instance i need to add material for each type of atom or amino acid.
The first one is just an idea i have with a friend (richard eudes), about why always visualize protein as sphere and not CUBE...

This second exemple illustrate the carbon alpha (Cube) connected using blender armature's bones. The bones are rendering using the script from http://www.malefico3d.org/blog-en/?p=13.

I think about trying metaball of type Cube, and also some collision in blender.....

Docking contact pairs occurence

2008-03-25T05:53:00.000-07:00

Docking or montecarlo procedure produce a huge number of conformation that have to be analysed. As bioinformatician we generally wrote some script to help in such task. In the present exemple, I explain how to calculate/analyse the amino acid occurence of presence at the interface, and furthermore, the occurence of particular pairs of residues in contact. All is done in the ICM package (http://www.molsoft.com/), and lead to produce 3D plot of matrice of pair contact.

The icm code (for example, not complete for clarity purpose):

#All this script is place in the loop exploring all the conformation (ie stack or trajectories)
load conf i OR load frame i
#method1 get the residues from each partnairs at the interface for a particular pose
show surface area a_1.// a_1.//
inter1=Res(Sphere(a_2.//* a_1.//* 4.) & Acc(a_1./*))
show surface area a_2.// a_2.//
inter2=Res(Sphere(a_1.//* a_2.//* 4.) & Acc(a_2./*))

#method2 get the pairs residues at the interface for a particular pose
t=Area(a_2./A, a_1./A)
read column group name="aa" input=s_out

#given the list of pair, creating the matrix of occurence
RI12=Matrix(72,50,0)
for j=1,Nof(aa)
y=Integer(Name(aa.2[j])[3:Sz])
x=Integer(Name(aa.3[j])[3:Sz])
RI12[x,y]=RI12[x,y]+1
endfor

#Zscore/normalisation of the matrix
Ntot=Sum(Sum(RI12))
for i=1,Nof(RI12[1,?])
for j=2,Nof(RI12[?,1])
nobsi=Sum(RI12[?,i])
fj=Sum(RI12[j,?])/Ntot
ntheoij=nobsi*fj
nobsvij=RI12[j,i]
zij=(nobsvij-ntheoij)/Sqrt(ntheoij)
ZRI12[j,i]=zij
endfor
endfor
#normal plot
plot area RI12 grid link ds

#3D plot
x=Rarray(50,Min(ZRI12),Max(ZRI12))
y=Rarray(72,Min(ZRI12),Max(ZRI12))
make grob solid "3dp_plot_name" ZRI12 x y 0.5 # scales for x,y and z

logo DSIMB

2008-03-20T07:14:00.000-07:00

Continuation of the logo's work. This time i have tried the svg import option of blender. Namely, you can export spline made in inkscape or gimp to blender in order to extrude them.
Here is an exemple, the text comes from Aurelie Bornot and are then convert to path in inkscape and save as svg file. Loaded in blender to extrude and modify (blend) to follow the sphere shape.
There is still work to do, get plain text, modify the D color, modify the M form, and change the background to transparent...

structural alphabet soupe

2008-03-12T03:53:00.000-07:00

I like this animation. For the details it is a molecular dynamic (gromacs) play in reverse mode. I have produce the letter using http://bioserv.rpbs.jussieu.fr/~autin/cgi-bin/ABCgen ... that I have made for some reason..
The rendering was done using VMD and gelato

molecular gift

2008-03-12T03:51:00.000-07:00

I have found this web site, I think they propose origninal gift for you entourage!

http://www.madewithmolecules.com/

(i found the link on the Inkling Magazine - On the hunch that science rocks web blog:

http://www.inklingmagazine.com/)

Chemical sunflow

2008-02-28T05:47:00.000-08:00

A new test of the sunflow rendering engine. Just to see how render the metal material in a simple scene...

DARC docking

2008-02-28T04:21:00.000-08:00

I am actually working on some report on my research. And for that purpose I have this several images representing my preliminar results which comes from rigid-body and flexible docking.

DSIMB logo

2008-02-26T04:18:00.001-08:00

I have play around the DSIMB motif, which is the new name for my actual laboratory (EBGM).
DSIMB is for "Dynamique et Structure des interactions des macromolécules biologiques".
The first one were using my ABCgen web server, which have produce the pdb file for the text "DSIMB", and then using VMD exported to gelato. Layer were add in the gimp.
The second serie comes from Blender.
The last one correspond to an idea from Aurélie Bornot which is a phd in the lab.

Sunflow

2008-02-15T05:55:00.000-08:00

In my search for nice molecular graphism, i am trying a lot of 3d render engine. I would like to put here some picture done with the last I have found on the web : Sunflow (http://sunflow.sourceforge.net/). It is a java based rendering engine, and he works really fine. It is really easy to use, and the API seems well documented.
The following image was done from their aliens_shiny exemple merge with the cornell_box exemple.
I have to tried the ambiant occlusion shader, and see if it simple to automatize sunflow in a pipeline.

My first test:

DARC-Docking

2008-02-15T03:14:00.000-08:00

I have made a litle script in the ICM (molsoft) package. It permit me to observe the position of the docking search by sphere (one sphere one ligand position), color by the number of visit. The representation is load in VMD and render to Gelato. The different pass were merged in gimp using the "value" mode.
For interested people, here is the litle icm script:

#generate a polyA sequence string
s1=""
for i=1,Nof(conf)
s1=s1+"A"
endfor

#build the polyA peptide and make only the Ca
build string IcmSequence( s1 "nter" "cooh" ) name= Name( Name( "dot" simple ), object )
makeCaTrace a_dot.
delete a_dot.

#for every conformation generatd during the docking, assign to the Ca atom the position of the Center of Mas of the ligand, and assign the number of visit to the atom's Bfactor

for i=1,Nof(conf)
load conf i
t=i+1
set a_dot_ca.m/t/ca Mean(Xyz(a_2.a))
set bfactor a_dot_ca.m/t/ca Real(tm.nvis [i] )
endfor

#then color by bfactor
color a_dot_ca.//* Bfactor(a_dot_ca.//*)

and the resulting figure :

ADN images

2008-01-29T06:43:00.001-08:00

Still playing with gelato. In this case i have use ray-traced shadow and indirect light in addition to the ambiant occlusion pass. Namely:

Light("light2", "distantlight", "float intensity", 1, "color lightcolor", (1, 1, 1), "point from", (1, 2, 0.5), "point to", (0, 0, 0),"string shadowname", "shadow","float shadowblur", 0,"float shadowsamples", 16,"float __nonspecular",0)
Light("light4", "indirectlight", "string indirectname", "indirect","float intensity", 0.5, "string sdbmode", "","float samples" ,64,"float bias" , 0.1)

Then all is loaded as layer in the gimp, and i use the grain merge mode with 50% opacity of the beauty pass.

The scene have a transparent background. (click to enlarge the picture)

new gelato image test

2008-01-28T02:37:00.000-08:00

Some other exemples of my gelato test. I have tried some differents shader in addition to the ambiant occlusion. There is transparancy, plastic postgloss pretexture shader test. Then I have used the Gimp to combine the different layer...
In this case the protein object comes from the ICM package ((c)Molsoft http://www.molsoft.com/). The latest icm version permit to export in clean wavefront object (thanks to Levon Budagyan). But the software is not free, as he includes molecular modeling algorithm like protein docking, and the free version doesn't permit to export.
For include this wavefront object in gelato, i have used the obj importer from Bill Robertson (http://www.tekbot.com/gelobj/).

DARC molecular dynamics

2008-01-23T04:35:00.000-08:00

This a first test on my research project on the DARC protein. This is the result of a molecular dynamics used in order to refine a complex previously obtain by combination of flexible and rigid-body docking. The animated gif was generated using my local pmg script which export the trajectory from VMD to gelato, and then create the gif or a avi movie (combining ambiant occlusion and diffuse pass). Click on the image below to see the animated gif (be carefull the gif is huge and can slow down your computer if you havn't enought memory).

VMD-gelato suite

2008-01-21T08:59:00.000-08:00

The same as before, but in 1500x1500 and then upload on youtube....

vmd-gelato script for Molecular movie trajectories movie

2008-01-18T03:28:00.001-08:00

I have adapted some part of the PMG code to get a litle script which permit to get automatically gelato movie from vmd. Here are exemples using 20 frames from a molecular dynamic. computation of ambiant occlusion take 3 min for each frame.(clic on the image to see the animation)

PMG vmd-gelato t_brown_sky

2008-01-10T06:51:00.001-08:00

gelato ambiant occlusion	gelato shader dynamic
gelato multiply	PMG megapov

gelato-vmd scene test 01

2007-12-17T05:05:00.000-08:00

occlusion pass and beauty pass (with indirect light)

Then different compose method:
Multiply

Lighten

Dissolve 50%

Overlay

I like the dissolve and lighten ones. They are automatically generated using ImageMagick package using linux command lines.

rpbs logo - gelato multipass test

2007-12-09T04:06:00.000-08:00

Ambiant occlusion pass
Normal pass (diffuse)
Multiply and crop

RPBS logo work part 2

2007-09-12T00:20:00.001-07:00

My last work on the rpbs logo, i show one of the rendered image and a final animated gif.

PMG last development

2007-09-04T05:55:00.000-07:00

t_shower

t_solvate

PMG new effect, protein's shower

2007-08-30T07:07:00.000-07:00

In the beta version of PMG-2.0, I have add a new scene type call "t_shower", here some snapshot of the development and final rendering example (http://bioserv.rpbs.jussieu.fr/~autin/cgi-bin/PMGdev)

test1

test2

final result in PMG

molecule men dancing 2

2007-08-30T07:01:00.000-07:00

that's done, with the Everbody was kungfu fighting song. But it's not the final.
Need to put some background, refine the enveloppe as some bug appear in the shape...

molecule men dancing !

2007-08-22T10:46:00.000-07:00

ok now he is alive, and he is dancing. I will improve the animation and then ad some music...

RPBS logo work

2007-08-22T10:34:00.000-07:00

I currently work on the rpbs logo, the idea is to produce an animated logo

new molecule men

2007-06-30T12:28:00.001-07:00

molecule guy

2007-06-18T12:04:00.000-07:00

i have play with my alphabet in order to produce a guy. Don't know now if is usefull...but it is done and seem nice.

annexin docking poster

2007-06-14T12:00:00.001-07:00

PMG in science

2007-06-09T18:09:00.000-07:00

NETWATCH: Molecular Home Movies

Press article in science magazine about PMG :

http://www.sciencemag.org/content/vol316/issue5830/r-samples.dtl?rss=1#316/5830/1401d

"The new Protein Movie Generator (PMG) provides an online studio for budding scientific Walt Disneys. Produced by two researchers at the University of Paris, the site makes it easy to create animations that put molecules in motion.

PMG starts with files from the Protein Data Bank or trajectories from molecular simulations. Users can then script simple scenarios, such as an enzyme pirouetting to display its active site, or more complex maneuvers, such as a ligand gliding in to dock with its receptor. If you prefer stills, you can use PMG to craft molecular graphics. This illustration, for instance, bares some of the internal structure of triosephosphate isomerase, one of the sugar-slicing enzymes of glycolysis."

bioserv.rpbs.jussieu.fr/~autin/help/PMGtuto.html