I am currently working on non-equilibrium alchemical transformation simulations involving multiple residue mutations, which may include several charge changes. I am aware of the “two systems in a single box” approach for handling charge-changing mutations; however, incorporating a counter-tripeptide for each mutation could significantly increase the box size, making it impractical.
In my research, I have primarily encountered point mutations in the literature. I have a few questions regarding the use of fast thermodynamic integration (FTI) for multiple mutations:
Are there any known issues or limitations when using FTI for multiple simultaneous mutations?
For multiple charge-conserving mutations, should I sum the corresponding tripeptide ΔG values for the unfolded branch of the thermodynamic cycle?
I have come across post-simulation correction schemes for charge-changing mutations, but I am uncertain if they are applicable to the FTI approach. Can anyone provide clarification on this?
I would greatly appreciate any guidance or suggestions on how to address these concerns.
no principal issues, just be aware that errors propagate, so the uncertainty will be larger for, say, a triple mutation than for a single mutation
yes
yes, we have a paper in press at JCTC on this. Correction schemes can be applied. Additional calculations are required for these corrections, so it may depend from case to case if it’s more convenient to neutralize the box or apply such a correction.
Thank you for your insightful response. I would be grateful if you could provide more details regarding the correction schemes you are utilizing. Are they similar to the ones described in this article? I would be really happy if there was a preprint of your paper that I could review.
Beside the “two systems in a single box” approach and post-simulation correction schemes, I saw people using alchemical counter ions to neutralize their system. Can you comment on the counter-ion approach?
I am eager to use your excellent pmx package for my research; however, at the moment I am stuck because of my issue with the net charge change for multiple mutations.
yes, the corrections are as in the Oehlknecht article (note in that work not all boxes were neutral, in our follow-up we show that neutral boxes w/o corrections yield similar results to non-neutral boxes w/ corrections).
The alchemical counter ion may also work, but you have to make sure that it cancels out in the thermodynamic cycle. Also note that solvation free energies of ions in water are several hundred kJ/mol. Such large numbers become part of the computed dG (twice), which may contribute to larger error bars. Also one needs to think about the placement/sampling of that/those ion(s).