Hii,
I have a beta-sheet protein of 128 residues, of which 8 are tyrosines. I have mutated all 8 TYR simultaneously to PHE. I used the same mdp files, except the temperature was changed to 300K. I followed the PMX protein mutation tutorial, created 50 windows for forward and reverse, and performed the final 80 ps equilibration. I got the result as shown. The number seems too high after I divide it by 8. What could be the issue? Should I increase the equilibration run length, or perform mutations sequentially rather than simultaneously?
Hi,
you don’t mention if you subtracted a reference case (such as the TRP->PHE mutation in the unfolded state, for which you can use a tripeptide as an approximation; in this case you should apply this correction 8 times). Without that, the raw free energy numbers are meaningless. Also, dividing by 8 might not be so informative, as it will only give you the average effect per mutation, but likely some of the mutations contribute a lot more than others, depending on the position.
Hi, thank you for the response. The purpose of this calculation is to calculate the effect of mutation on protein dimerization. As TYR residues are a dimerization site, I am mutating them. I am using the thermodynamic cycle shown in the attached figure. till now I simulated only the lower leg, where I mutated TYR of monomer to PHE of monomer, so this value of ~568kJ/mol corresponds to ΔGmonomermutation, which I think is very high. What correction factor should I use here?
the thermodynamic cycle looks fine. The number you get only makes sense relative to the upper leg (dimer). So if you subtract the monomer value from the dimer value, you know how much the mutations affect the dimerization. The individual monomer and dimer mutation dG values have no physical meaning.
Thanks for the response. I ran the dimer leg and calculated the final values
for monomer
Bootstrap (Std Err): iteration 100/100
Bootstrap (Conv): iteration 100/100
BAR: dG = 568.22 kJ/mol
BAR: Std Err (analytical) = 1.56 kJ/mol
BAR: Std Err (bootstrap) = 1.25 kJ/mol
BAR: Conv = 0.25
BAR: Conv Std Err (bootstrap) = 0.15
for dimer
Running Nelder-Mead Simplex algorithm…
Bootstrap (Std Err): iteration 100/100
Bootstrap (Conv): iteration 100/100
BAR: dG = 581.80 kJ/mol
BAR: Std Err (analytical) = 10.20 kJ/mol
BAR: Std Err (bootstrap) = 1.69 kJ/mol
BAR: Conv = 0.97
BAR: Conv Std Err (bootstrap) = 0.00
The gap between forward and backward leg for dimer is quite high, and distributions are not overlapping, and I think that’s why the BAR: Std Err (analytical) = 10.20 kJ/mol is coming high. Should I worry about this? As the free energy value seems to be in the range, but I don’t have experimental validation.
the lack of overlap is likely because you’re doing 8 mutations at once, which is a big perturbation. This can be remedied by making slower (2 fold or 5 fold) transitions.
Thank you for the help. Now I used 500 ps equilibration per frame. And I am getting a low error bar.
Thanks again!
Running Nelder-Mead Simplex algorithm…
Bootstrap (Std Err): iteration 100/100
Bootstrap (Conv): iteration 100/100
BAR: dG = 578.12 kJ/mol
BAR: Std Err (analytical) = 1.01 kJ/mol
BAR: Std Err (bootstrap) = 0.89 kJ/mol
BAR: Conv = 0.00
BAR: Conv Std Err (bootstrap) = 0.13
I
