I would like to know if it is possible to do MD calculations for protein on a regular PC? If it is possible, then what are some of the software that can be used and if there is anything i need to take in consideration before running the simulation.
Hi Firdaus Abdullah
I'll just add some extra information to what Adron Ung said.
Keep in mind that to perform MD simulations you will need some minimum requirements. Although computing time is directly linked to the size of your system (e.g., protein size, water box, and etc), I'll say that based on my experience (I have performed simulation on all sort of PC and servers), you will need at least an 8 core, 16 GB computer, with a high storage HD, at least 1TB. Here is why.
Firstly, MD is not really based on RAM, they are linked to your processing power, thus, the more cores the better, and if you get it running on GPUs, it would be marvelous. However, RAM will always come in handy when dealing with other visualization programs, or some other software you might need to perform extra analysis. Second, depending on your system size, you will have massive outputs. For example, running a ~400aa protein in a dodecahedric box, using a united atom force field (like Gromos), for 50ns, only the trajectory final file might have around 20GB, and the .trr file (optional) will probably have around 80GB. After performing multiple simulations you will need a lot of HD space. If during the MD you run out of scape, your simulation will crash. You can perform the MD with fewer requirements if you want, or need, but it will demand your full attention on manipulating the outputs, saving them on external HDs, and having just being patient to wait for results.
Just some basic tips, if you have more questions I'll be happy to answer,
Cheers,
Bruno
Yeah, molecular dynamics simulations can totally be done on a personal computer. If you want some popular good free software I recommend either Nanoscale Molecular Dynamics (NAMD) or Gromacs. Gromacs is more command line based. NAMD is more you write out the special custom file specifying all the constraints of your simulation, attach your molecule and attach the file and run the program. Both of these software have very detailed tutorials. There are other software but they cost money. To visualize the final product I recommend using either VMD or Pymol. Whatever suits you best. If you can't get the tutorials to work, make sure you are in the right directory with all the files on your computer and make sure the software is in your PATH? If you download VMD, open command prompt and type "VMD" and it doesn't open up the software then VMD isn't in your PATH. Path is basically a Bash term for terminals. Put the software in your terminal PATH.
This is where you can safely download the original NAMD software and where all the tutorials are. https://www.ks.uiuc.edu/Research/namd/
If you are wondering, there is such a term as "All atom simulations" and "Course grained simulations". In a Course grained simulation, basically amino acid residues are treated like balls connected to springs which are the bonds. In an all atom simulation, every atom in every residue is a ball and every chemical bond stretches like a spring in the simulation. All atom simulations especially for protein complexes take all days/weeks/whatever especially if you are simulating mostly a whole virus or a protein moving through a nuclear pore. All atom simulations are technically more accurate "if you know what you are doing". If you don't know what you are doing and are just using the default parameters assuming they are the best parameters then you are wasting your time generating inaccurate data.
Hi Firdaus Abdullah
I'll just add some extra information to what Adron Ung said.
Keep in mind that to perform MD simulations you will need some minimum requirements. Although computing time is directly linked to the size of your system (e.g., protein size, water box, and etc), I'll say that based on my experience (I have performed simulation on all sort of PC and servers), you will need at least an 8 core, 16 GB computer, with a high storage HD, at least 1TB. Here is why.
Firstly, MD is not really based on RAM, they are linked to your processing power, thus, the more cores the better, and if you get it running on GPUs, it would be marvelous. However, RAM will always come in handy when dealing with other visualization programs, or some other software you might need to perform extra analysis. Second, depending on your system size, you will have massive outputs. For example, running a ~400aa protein in a dodecahedric box, using a united atom force field (like Gromos), for 50ns, only the trajectory final file might have around 20GB, and the .trr file (optional) will probably have around 80GB. After performing multiple simulations you will need a lot of HD space. If during the MD you run out of scape, your simulation will crash. You can perform the MD with fewer requirements if you want, or need, but it will demand your full attention on manipulating the outputs, saving them on external HDs, and having just being patient to wait for results.
Just some basic tips, if you have more questions I'll be happy to answer,
Cheers,
Bruno
Yes, you have to take the processor of the PC into consideration. The protein size will determine the computational speed of the simulation. With a PC that has high processor, the calculation can be carried out.
Totally, in addition to what was said before, you risk damaging your pc in case of the surcharge or an overheat (in case you have not enough ram/GPU or even memory space) but you can do MD simulations (that will take more time depending on the size of your system).
Some examples of free to use MD programs: DL poly or Gromacs
Best regards;
Joseph
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics