I am currently using gmxMMPBSA tool for MMPBSA analysis. I have a 100ns gromacs trajectory for protein-ligand complex which contains 10000 frames. Even if I run the MMPBSA analysis with interval of 5, it takes one day to complete analyzing complex contribution alone. Please suggest me a solution to get result in half day.
Between can I take the last 10ns of the trajectory and apply interval of 5 for analysis. will it give be good result in small duration.
Hi Sree Haryini,
I often ran into such problems during my PhD. The Poisson-Boltzmann approach for calculating the surface is probably what causes your issues. Have you considered using the Generalised-Born approach (MM/GBSA)? It is much faster and gives comparable results.
If you are more interested in what frames to use for such mmpbsa/mmgbsa calculations you could check out my thesis. In figure 2.3, I explain the protocol we employ in the lab where I was based.
best Martin
Dear Sree Haryini ,
Here is suggestion from me:
1. Reduce the number of frames: As you suggested, using a smaller portion of the trajectory can significantly reduce computation time. However, make sure that the selected portion is representative of the system's behavior.
2. Increase the interval: Instead of analyzing every frame, you can increase the interval between frames. For example, you can analyze every 10th or 20th frame instead of every 5th frame. This will reduce the number of calculations needed.
3. Parallelize the computation: Calculate by using multiple core, you can parallelize the computation to speed up the analysis. Many MMPBSA tools support parallel computing. If you use the tool by Valdes you can add mpirun -np X and X can be replace with number of processors. Example mpirun -np 16. But you also need to consider the RAM available.
4. Calculate MMGBSA - As suggested by Martin Rosellen calculate GBSA is indeed will reduce the workload for calculation. I have try before that calculate the GBSA take 2 hours while PBSA take almost 1 days in the same trajectory.
- Badges
- Science topic
- Similar topics
- Psychology