Different masses of dummy atoms in webserver pmx and standalone pmx

pmx
1

Dear Gromacs users,
I am trying to calculate ddG of folding of barnase mutant H18D (barnase ID: 1BNI) using FEP with double system/single box protocol. While generating hybrid structures for barnase and tripeptide I noticed that there is the difference of mass values for dummy atoms in itp files generated in pmx webserver and latest pmx version from GitHub (GitHub - deGrootLab/pmx: Toolkit for free-energy calculation setup/analysis and biomolecular structure handling).

Here are mass values of dummy atoms for 18H2D residue itp generated in pmx webserver:

235 N 18 H2D N 235 -0.415700 14.0100 N -0.516300 14.0100
236 H 18 H2D H 236 0.271900 1.0080 H 0.293600 1.0080
237 CT 18 H2D CA 237 0.018800 12.0100 CT 0.038100 12.0100
238 H1 18 H2D HA 238 0.088100 1.0080 H1 0.088000 1.0080
239 CT 18 H2D CB 239 -0.046200 12.0100 CT -0.030300 12.0100
240 HC 18 H2D HB1 240 0.040200 1.0080 HC -0.012200 1.0080
241 HC 18 H2D HB2 241 0.040200 1.0080 HC -0.012200 1.0080
242 CC 18 H2D CG 242 -0.026600 12.0100 DUM_CC 0.000000 1.0000
** 243 NA 18 H2D ND1 243 -0.381100 14.0100 DUM_NA 0.000000 1.0000**
** 244 H 18 H2D HD1 244 0.364900 1.0080 DUM_H 0.000000 1.0000**
** 245 CV 18 H2D CD2 245 0.129200 12.0100 DUM_CV 0.000000 1.0000**
** 246 H4 18 H2D HD2 246 0.114700 1.0080 DUM_H4 0.000000 1.0000**
** 247 CR 18 H2D CE1 247 0.205700 12.0100 DUM_CR 0.000000 1.0000**
** 248 H5 18 H2D HE1 248 0.139200 1.0080 DUM_H5 0.000000 1.0000**
** 249 NB 18 H2D NE2 249 -0.572700 14.0100 DUM_NB 0.000000 1.0000**
250 C 18 H2D C 250 0.597300 12.0100 C 0.536600 12.0100
251 O 18 H2D O 251 -0.567900 16.0000 O -0.581900 16.0000
252 DUM_C 18 H2D DCG 252 0.000000 1.0000 C 0.799400 12.0100
** 253DUM_O2 18 H2D DOD1 253 0.000000 1.0000 O2 -0.801400 16.0000**
** 254DUM_O2 18 H2D DOD2 254 0.000000 1.0000 O2 -0.801400 16.0000**
For all dummy atoms mass values are equal to 1
Here are mass values of dummy atoms for 18H2D residue itp generated with generate_hybrid_topology.py script from standalone pmx from GitHub:
235 N 18 H2D N 235 -0.415700 14.0100 N -0.516300 14.0100
236 H 18 H2D H 236 0.271900 1.0080 H 0.293600 1.0080
237 CT 18 H2D CA 237 0.018800 12.0100 CT 0.038100 12.0100
238 H1 18 H2D HA 238 0.088100 1.0080 H1 0.088000 1.0080
239 CT 18 H2D CB 239 -0.046200 12.0100 CT -0.030300 12.0100
240 HC 18 H2D HB1 240 0.040200 1.0080 HC -0.012200 1.0080
241 HC 18 H2D HB2 241 0.040200 1.0080 HC -0.012200 1.0080
242 CC 18 H2D CG 242 -0.026600 12.0100 DUM_CC 0.000000 3.9633
** 243 NA 18 H2D ND1 243 -0.381100 14.0100 DUM_NA 0.000000 4.6233**
** 244 H 18 H2D HD1 244 0.364900 1.0080 DUM_H 0.000000 1.0000**
** 245 CV 18 H2D CD2 245 0.129200 12.0100 DUM_CV 0.000000 3.9633**
** 246 H4 18 H2D HD2 246 0.114700 1.0080 DUM_H4 0.000000 1.0000**
** 247 CR 18 H2D CE1 247 0.205700 12.0100 DUM_CR 0.000000 3.9633**
** 248 H5 18 H2D HE1 248 0.139200 1.0080 DUM_H5 0.000000 1.0000**
** 249 NB 18 H2D NE2 249 -0.572700 14.0100 DUM_NB 0.000000 4.6233**
250 C 18 H2D C 250 0.597300 12.0100 C 0.536600 12.0100
251 O 18 H2D O 251 -0.567900 16.0000 O -0.581900 16.0000
252 DUM_C 18 H2D DCG 252 0.000000 3.9633 C 0.799400 12.0100
** 253 DUM_O2 18 H2D DOD1 253 0.000000 5.2800 O2 -0.801400 16.0000**
** 254 DUM_O2 18 H2D DOD2 254 0.000000 5.2800 O2 -0.801400 16.0000**

Here mass values are different from 1 for non-hydrogen atoms. Did you implement some scaling factor for masses for a standalone version of pmx? How to define this scaling factor and which itp should I use? I assume this detail may drastically affect the ddG prediction results.

I am a new user so I cannot upload full itp files in any way, but hopefully I will be able to do that in a discussion.

Thank you all in advance!

2

Hi Mary,

The recent version of pmx (develop branch) applies a scaling factor of 0.33 by default. Hence, the DUM_C has a mass of 12.01*0.33=3.9633. The web server uses an old version of pmx, in which the mass of a dummy is 1.0 by default. The scaling factor is used to maintain the stability of a simulated system.
You can read this thread for more on the discussion: Oscillational period warning during tripeptide A2I forward simulation

Using either of the itp files should be fine provided that your simulation is stable (no oscillation-period-warning). The mass of the dummy particle will not have much effect on the ddG prediction.

Best,
Sudarshan

1 Like