AWH FEP mdp options

User discussions
, ,
1

GROMACS version: 2021.1
GROMACS modification: No
Here post your question

Hi,

I’m keen to try out the new AWH for FEP code in gmx 2021, however I am having difficulties working out what would be reasonable starting mdp options for the awh parameters.

Yesterday’s webinar on the subject detailed some specific recommendations for this, but I was unable to make note of them on time. I unfortunately couldn’t seem to be able to find them in documentation either.

I was therefore wondering if anyone would have example AWH FEP mdp settings they could share?

Many thanks,

Irfan

2

Here are the mdp options from the webinar. Note that these are not the only “right” ones and might need to be slightly modified for different use cases, but they should mostly work:

free-energy = yes
couple-lambda0 = none
couple-lambda1 = vdwq
couple-moltype = ethanol
couple-intramol = no
init-lambda-state = 15
vdw-lambdas = 1.0 1.00 1.00 1.00 1.00 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00
coul-lambdas = 1.0 0.80 0.60 0.40 0.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
calc-lambda-neighbors = -1
separate-dhdl-file = no
sc_alpha = 0.5
sc_sigma = 0.3
sc_power = 1
sc_coul = no

awh = yes
awh-potential = umbrella
awh-nstout = 500000
awh-nbias = 1
awh-nstsample = 10
awh-nsamples-update = 10
awh1-error-init = 10
awh1-equilibrate-histogram = no
awh1-target = constant
awh1-growth = exp-linear
awh1-ndim = 1
awh1-dim1-coord-provider = fep-lambda
awh1-dim1-coord-index = 1
awh1-dim1-start = 0
awh1-dim1-end = 15
awh1-dim1-diffusion = 0.002

2 Likes
3

Thank you so much for your response!

I was wondering if I could also ask about analysis. Simply running gmx awh seems to be giving approximately the right dG between the two end states, but I was wondering if there was some additional awh-specific correction I was missing?

4

No, you get the free-energy difference, or better whole profile along lambda, directly.

gmx awh -more gives more information, about sampling and the metric.

5

I wanted to try AWH FEP method with system I am currently working, and got grompp warning:

WARNING 1 [file …/mdp/2_nvt.mdp]:
Masses may not be perturbed when using the free energy lambda state as
AWH coordinate provider. If you are using fep-lambdas to specify lambda
states make sure that you also specify mass-lambdas without perturbation.

I wanted to morph one ligand to another by switching on/of several atoms.
Are such calculations supported by this method?

The mdp file I used was:

free-energy = yes
couple-lambda0 = vdw-q
couple-lambda1 = vdw-q
couple-intramol = no
init-lambda-state = 0
calc-lambda-neighbors = -1
separate-dhdl-file = no
sc_alpha = 0.5
sc_sigma = 0.3
sc_power = 1
sc_coul = no

awh = yes
awh-potential = umbrella
awh-nstout = 500000
awh-nbias = 1
awh-nstsample = 10
awh-nsamples-update = 10
awh1-error-init = 10
awh1-equilibrate-histogram = no
awh1-target = constant
awh1-growth = exp-linear
awh1-ndim = 1
awh1-dim1-coord-provider = fep-lambda
awh1-dim1-coord-index = 1
awh1-dim1-start = 0
awh1-dim1-end = 15
awh1-dim1-diffusion = 0.002

; lambda states 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
vdw-lambdas = 1.0 1.00 1.00 1.00 1.00 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00
coul-lambdas = 1.0 0.80 0.60 0.40 0.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
;mass-lambdas = 0.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Thanks,
Raitis

6

Anything is supported by AWH. It operates on lambda and the energy differences, it doesn’t know at all what lambda actually does with the system.