Here post your question
Simulations using different GROMACS versions resulted in noticeable variations in protein structure, suggesting that version-specific algorithms or parameters influence structural dynamics and outcomes. This whole structure is provided after using GROMACS version 2024.4, but using version 2018.1 there is no such changes in protein structure.
nvt.mdp and topology file is also attached here.
The exact same setup for a simulation won’t necessarily produce the exact same results since the velocities are generated stochastically. Also, the visualization on VMD isn’t always perfect, I’ve found that it sometimes displays sheets and sometimes doesn’t for what should be the same atomic coordinates. Difference in versions shouldn’t affect the algorithms used to generate the trajectories
I know that the same simulation setup can produce different outcomes, such as differences in system energy or molecular trajectories. However, the formation of a new bond that did not initially exist in the system (protein) indicates an error. I want to know whether this is a visualisation error or a problem in the simulation? And How can I differentiate between the two so that I can decide whether to continue this simulation or not?
By new bond, do you mean a new hydrogen bond inferred from the visualization? Given the geometric criterion for hydrogen bonding, it’s entirely possible that a different molecular trajectory could result in the atoms shifting to fit that criterion, and is not necessarily an error. The different versions don’t change which algorithms are used to generate the trajectories, those are selected in your mdp file.
I’ve also had instances where the exact same structure seems to be visualized slightly different with newcartoon in VMD. In fact, if you were to try reloading your structure now into VMD, it’s entirely possible the newcartoon representation is slightly different from the pictures you’ve uploaded.
