GROMACS version:
GROMACS modification: Yes/No
Here post your question
These are umbrella sampling graphs performed on 3 proteins with same ligand. I am unexperienced and unable to interpret and different differences in these graphs. It would be really helpful if I can know the interpretation between three and any errors if I should correct in WHAM/PMF.
your sampling need more windows or more time.
Sir, i made sampling windows at spacing of 0.2nm ,which gave me 23 windows from summary.dat file… So do i need to again perform the whole umbrella sampling by reducing this spacing to 0.1nm which will increase my window numbers or I can just randomly increase the number windows in between the gaps.
You don’t have to rerun your current simulations. Regenerate new starting configurations to cover regions with little overlap in your histogram plots. You probably just need a few more.
As suggested above, it might also be good to increase the sampling time per umbrella window.
When making comparisons between simulations you should also use the bootstrap error estimation options in `gmx wham´.
Hello Manish.
For an interpretation of what those graphs mean, I think that literature already covers that. See for example Lemkul and Bevan (https://pubs.acs.org/doi/10.1021/jp9110794).
Regarding the soundness of your calculation, I think that you need further sampling. To be specific, the analyses show above show that there is more sampling needed at the following reaction coordinates:
First experiment:
0.75; 1.35; 3.6.
Second experiment:
0.7; 1.15; 1.8; 2.8.
Third experiment:
0.95; 4.6.
All values above are in nanometres.
Also, it might be appropriate to revisit the umbrella sampling tutorial. It would be important in your case because it addresses the choice of window spacing which in your case is 0.2 nm.
I hope this helps.
Ivan
Thank you sir it worked for me and I have increased the window sampling. But as it is seen that the PMF graph shows multiple trunks and vallys, than how can i calculate deta G from this. Please if you can help.
