I am a MSc Computing Science student. As a project for an computational science course we (I and another student) have chosen to implement a molecular dynamics simulation program. We would like to have some real data to test the application, however we know almost nothing about molecular biology etc.
Our M.D. simulation program supports the following interaction types between molecules:
- Non-bonded Lennard-Jones.
- Non-bonded Coulomb.
- Bonded covalent interaction.
- Bonded bond-angle interaction.
- Bonded dihedral interaction.
We would like to have two proteins to test the simulation program. Both proteins have to be so 'simple' that they can completely be described as a set of particles (molecules) with a three-dimensional position and properties like mass, electric charge etc., and a list of all the (2/3/4)-tuples for the bonded interations.
Preferably, both proteins should have a known structure so we can check if the structure as computed by our simulation program is correct.
The first protein should be a very small (toy) protein, so we can run the simulation for this protein on a simple desktop machine without having to wait very long.
The second protein should be larger. We have access to a computer cluster with up to 24 cores and 128 GB of memory, so we have a nice amount of computing power available for the simulation of this large protein.
I know there are protein databanks available on the internet, where I can just download protein descriptions. But I don't know which one to take, and in the PDB file I don't quickly spot a list of all molecules and the bonded interaction combinations.
We would be very happy if someone can help us in the right direction.