Halo Experts,
I have pulled 3 drug molecules with good dipole moment 2.68 D from aqueous phase to the centre of lipid bilayer mimic blood brain barrier. I attach here two plots for probability distribution histograms and free energy per mole after 45 nanosecond restrained run for each of 24 configurations generated after umbrella sampling. Is it correct ? Please help me out with your kind suggestions.
As far as I can say, it looks good.
One thing that is sometimes overlooked with umbrella sampling is whether the sampling time is long enough and if the simulations are properly equilibrated. Does the PMF look (more or less) the same with
-b 0 (which I assume that you have now)
and
-b 10000
?
The problem with that check is that the sampling time will be different. What might be more informative is to see if the results would be different with a sliding window of sampling time, e.g.:
-b 0 -e 20000
-b 5000 -e 25000
-b 10000 -e 30000
-b 15000 -e 35000
-b 20000 -e 40000
-b 25000 -e 45000
If there is a systematic shift in the PMF with increased equilibration time (discarding more time in the beginning of the simulations), you might want to consider extending the simulations and discarding as much data as needed to get only small systematic shifts when sliding the window of sampling further ahead in time.
1 Like
Thank you for your valuable suggestions.