Different RMSD results in re-run MD simulations of a ligand

User discussions
1

GROMACS version: 2023.3

short story is when i used MD two times for a same protein-ligand complex, i got very different results. but the long story is, i have 4 ligands with same chemical core but slightly different substituents along their in-vitro IC50s. based on their laboratory IC50s, ligand 1,2 and 4 are potent inhibitors but ligand 3 is weak. i used docking to investigate their orientations inside of a protein active site. in the next step i used 100ns MD to study their docking pose stability. Ligands RMSD results showed ligand 1,2 and 4 are stable and ligand 3 is unstable which is compatible with their laboratoy IC50s. next, i used 100ns MD with same parameters again for ligand 3 but this time ligand RMSD plot showed stable orientation which is in contrast of IC50. i really appreciated it if someone explain me why a ligand has different behavior during several re-run MD simulation and in this case how should i validate my MD simulation. thank you for your time.

2

MD simulations are chaotic. Thus one can expect (very) different results in repeats of simulations, depending on how the free-energy landscape looks like. You need to sample long enough to cover all relevant timescales, which can often be beyond what can be reached with MD.