Two Picoseconds Simulation of Protein

B

BaSuP

Active Member
Joined
Jul 24, 2013
Messages
46
Reaction score
0
I have finally managed to get gromacs 4.6.3 to work on a workstation that can simulate the molecular dynamics of ligand bound to a protein so I just have to share it :D.

[video=youtube;TlBi64K-l8M]http://www.youtube.com/watch?v=TlBi64K-l8M&feature=youtu.be[/video]

It was run on the following system:
Xeon 1230 v2
EVGA 660ti with CUDA 5.5 and 320.49 drivers
8GB RAM
120GB SSD
Ubuntu Linux 12.04 and took about 10 seconds to calculate.

What is nice about this is that this system allows me to calculate 1-10 nanosecond simulations in less than a day. Nvidia have developed CUDA and the developers that made gromacs are making use of this utility to run many of the calculations on the GPU and thereby speeding up these calculations quite substantially.

The protein is the 3MXF protein from the rcsb.org databank. Our research group is developing new drugs to bind to this and other proteins. Nanosecond simulations are quite valuable to see how these ligands interact with proteins. So this software together with readily available docking software will allow drug design groups to identify lead hits at a better rate (i.e. less development costs).

Are there any people on MyBroadband interested in these kinds of calculations or would like to get involved in projects like these?
 
I don't get what you mean by 2 picoseconds. Does it take just two picoseconds to do the simulation? Accuracy is the real standard.
 
porchrat said:
Very cool. What are you visualising it with?... VMD?
Click to expand...
I used Chimera for this visualization. VMD is also good. I use these two the most.

Swa said:
I don't get what you mean by 2 picoseconds. Does it take just two picoseconds to do the simulation? Accuracy is the real standard.
Click to expand...
It took about 10-20 seconds to calculate 2 picoseconds worth of the above protein simulation. I.e. the above video is what happens in 2 picoseconds, just stretched over 11 seconds in the video.
 
Last edited:
I have no idea what you are talking about but you seem chuffed that you managed to do it.Well done.
 
Nerhzelok said:
Seconded.

What exactly is it "doing"? I'm stupid, please help me. lol
Click to expand...

Basically he is using computers to simulate the movement of atoms in 3D space using our pre-existing understanding of the physics of atoms and molecules. This particular simulation shows the atoms moving over a duration of 2 picoseconds. A picosecond of course being 10 to the power of -12 of a second (so to us seemingly very short but to atoms that is a reasonable length of time :p). This is his starter simulation though, as he goes on he will likely start running simulations that are measured in nanoseconds.

At regular intervals the programs that calculates the movements of the atoms take a snapshot of where the atoms are in 3D space at a point in time. When you string all those snapshots together in chronological order it creates a movie you can watch to see how the atoms moved and interacted. That is what you guys are seeing here.

This particular simulation involves a protein that you can see as the blue ribbon-like thing (this is a way of visualising proteins that focusses on structure and not individual atoms). While that might seem a little new to you guys if you look at the little red and white things you will notice something that is far more familiar. The red and white objects you can see floating around are water molecules. You can see the red single oxygen bound to the 2 white hydrogens. Hopefully that helps you put this into perspective.

It is a relatively new form of scientific investigation but one that is rapidly gaining popularity, accuracy and power. The rise of GPU computing is really going to move this sort of thing forward (whereby instead of using a CPU like most programs these simulations would make use of graphics processors... GPUs).
 
Last edited:
porchrat said:
Basically he is using computers to simulate the movement of atoms in 3D space using our pre-existing understanding of the physics of atoms and molecules. This particular simulation shows the atoms moving over a duration of 2 picoseconds. <snip>.
Click to expand...

Very interesting, thanks for the explanation!
 
Nerhzelok said:
Very interesting, thanks for the explanation!
Click to expand...
You are most welcome. (in case you, or anyone else reading this, is interested in reading further, this sort of thing is called Molecular Dynamics)

I reckon that in the coming years these sorts of simulations will become a routine part of how those performing chemistry based research in both the private and academic world will decide on which research paths hold the most promise. It is so much quicker and cheaper than trying to get similar information through physical experimentation. Especially nowadays with cheap and powerful hardware.

Please BaSuP keep us up to date on your work. I know I would love to watch it progress. Extremely interesting and thanks for showing us. :)
 
porchrat said:
Haven't used Chimera myself. The movies it makes are lovely though. Good work. :)
Click to expand...
Thanks :). It's a pretty user-friendly piece of software and its free :D.

Finally figured out how to get the correct protonation states for histidines. Made a 20 picosecond simulation with one of our inhibitors bound to a protein it was designed to inhibit. And to my [pleasant] surprise, it seems to corroborate some of our previous docking studies :D. Now I just need to figure out how to calculate Gibbs free energy of binding for this compound. Here is the simulation:
[video=youtube;17Q0X0lfXdo]http://www.youtube.com/watch?v=17Q0X0lfXdo&feature=youtu.be[/video]
 
Interesting. There isn't an awful lot of water there, did you select it out when doing the sim?

Most of the free energy calculations I've seen involve defining a set of variables that accurately describe the shift that is going to occur and setting increments for movements within those variables using a biasing force.

I'm not sure how you would do it in your situation. Good luck though.
 
It was conducted in a box of roughly 5x5x5 nm filled with solvent (water) and NaCl. It looks like this:
image.png


The waters around the ligand were selected otherwise they clutter the movie.

I will try to adopt this protocol for the energy calculations:
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin/gmx-tutorials/free_energy/index.html

Sounds similar to what you are saying.
 
BaSuP said:
It was conducted in a box of roughly 5x5x5 nm filled with solvent (water) and NaCl. It looks like this:
<snip>

The waters around the ligand were selected otherwise they clutter the movie.
Click to expand...
Ah I see. Lekker. Sorry didn't mean selecting out when you did the sim, I meant selecting out when you visualised. Silly me. :p


I will try to adopt this protocol for the energy calculations:
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin/gmx-tutorials/free_energy/index.html

Sounds similar to what you are saying.
Click to expand...
Oh yea looks pretty similar to what I was talking about. Handy that gromacs has it built in. It really is a great looking Molecular Dynamics application that. It can run so many different file formats too without a lot of messing about. NAMD and AMBER inputs for example.
 
porchrat said:
Oh yea looks pretty similar to what I was talking about. Handy that gromacs has it built in. It really is a great looking Molecular Dynamics application that. It can run so many different file formats too without a lot of messing about. NAMD and AMBER inputs for example.
Click to expand...
Yeah, it has a great community as well, meaning that a lot of the errors one may experience have been addressed somewhere (google is handy as always), and there are many very handy scripts to use.

/off prepping for lectures, will have to do the energy calculations much later :(.

See later.
 
BaSuP said:
Yeah, it has a great community as well, meaning that a lot of the errors one may experience have been addressed somewhere (google is handy as always), and there are many very handy scripts to use.

/off prepping for lectures, will have to do the energy calculations much later :(.

See later.
Click to expand...
I was on the gmx-users mailing list for a while. There is a staggering amount of activity there. That community is very healthy. :D

Well enjoy your lectures and feel free to keep these updates coming. Some interesting work you are doing there.
 
So just to be clear, are you predicting protein structure or simulating existing structure?
 
[video=youtube;5db1JK6f2LQ]http://www.youtube.com/watch?v=5db1JK6f2LQ[/video]
5 nanosecond simulation of GABAA alpha1 in a POPC membrane with physiologically correct ion concentrations (intracellular and extracellular) for sodium, chloride and potassium.
 
Awesome seeing the membrane sitting there. I haven't seen many simulations involving membranes before. How many atoms in that system and how long did it take?

Also what hardware are you running?
 
porchrat said:
Awesome seeing the membrane sitting there. I haven't seen many simulations involving membranes before. How many atoms in that system and how long did it take?

Also what hardware are you running?
Click to expand...
+-135000 atoms in a 10x10x15 box including POPC (about 130), Cl, Na, K, protein and GABA.
I used my desktop PC. Xeon 1230v2 with a 660ti running on Ubuntu 12.04. I got about 8ns/day (Gromacs 4.6.3 and Verlet cut-off) so it ran for a little more than a day.

I used the AMBER forcefield for this simulation. VMD was used to insert the protein in the lipid bilayer. Afterwards the the system was minimized and equilibrated (pressure and temperature) before doing the simulation.

The crystal structure for the GABA receptor does not exist so I used the Bergmann homology model (published in PloS ONE). The problem with this is that the intracellular loop domain is missing so I am now buzy trying to make good homology models of these domains. So far I-Tasser seem to do a reasonable job for this.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X