Dear All,
I used mm/pbsa to calculate the binding energy of protein-ligand complex. However, the errors for binding energy seem very hugh (mostly contributed by polar energy). The MD is running for 10 ns. May I please know how I'm able to reduce errors? Thank you very much!
I recommend you to do more calculations. Correct analysis is needed, you may get correct results with low errors
Title: "The MM/PBSA and MM/GBSA methods to estimate ligand-binding affinities"
Samuel Genheden & Ulf Ryde
Section 3 about the precision. "3. The precision"
I recommend you to use the CaFE tool for binding affinity prediction using end-point free energy methods.
it's easy to use.
here is the paper for CaFE:
https://academic.oup.com/bioinformatics/article/32/14/2216/1743371
best regards.
Yiran Qu , Please check if these ligands are moving. Since, you said that these results belong to the first 10 ns, I am assuming that you had sampled frames pre-convergence. Please check the RMSD plots of 1) protein, 2) ligand backbone w.r.t protein backbone to select the span of trajectory that you can extract the frames from. Ideally it is better to take these snapshots from the post convergence (trajectory). Ideally at least 50 to 100 frames between 8 to10 ns might give you a better result (Provided convergence is achieved).
All the best.
You may crosscheck your calculation by using CaEF tool, which is inbuilt in the latest version of VMD. This tool is very helpful and straight forward, here you may make different selections as per your requirement. In addition, you may also try AMBERTOOL for the MMPBSA , which is freely available, you need to convert your file according to the MMPBSA input.
Again, it depends on what kind of converged values you are looking for, you cannot rely on one software to predict the energies, always cross-check them there is a huge possibility of wrong output.
- Badges
- Science topic
- Similar topics
- Names