Dear All,
Recently, we have used cgenff Parachem web server "https://cgenff.paramchem.org/" to generate parameters for glucosylated aspartyl residue that exist in high energy non-native sugar (x) puckering conformation. Further, we used "cgenff_charmm2anton.py" script to convert CHARMM stream file to a GROMACS.itp for simulation purpose. After all the necessary modifications, we performed 10ns simulation following nvt, npt with 500 PS and 1ns respectively using our modified residue parameters (CHARMM36.ff and GROMACS.5.1.2)
In order to validate the parametrization, further we extracted 1000 snapshots for every 10 PS and calculated their puckering state. Surprisingly though our parameters and initial structure taken from high energy no-native (x) puckering state, most of the snapshots (95%) were found in 4C1 puckering state not even one (x) non-native state. These results lead us here to ask the following questions:
- Is our given parameters are correct?
because our parameters supposed to produce high no of non-native (x) sugar puckering state. Instead we gets more of native 4C1 puckering conformation.
- So, Is it normal to have more native puckering conformation? if yes kindly give us some comments.
During our cgenff parametrization step, we got the following dihedral "O5,C1,OD2,CG" parameter with high penalty score (45) and indeed this DIHR cause the non-native puckering state.
Thanking you in advance
Stéphane Abel,
Many more thanks for your answer
- The simulations are too short, especially if you perform simple MD
Indeed you are true but as I mentioned in the query though our initial structure was in non-native conformation with high energy, simulation directly jumps in to the less energetically favourable conformation . So we felt that even increasing the duration of simulation might behave same. Still we have plan to run long simulation.
-Indeed the high parameters means that the O5,C1,OD2,CG dihedrals need to be refined (re-optimized) against QM data and after re validated with MD.
can you recommend some tutorials or paper about the same
-Moreover why did you Parachem ? Keep it mind that Parachem should be only used for ligand and not for protein and sugar.
Actually this one of our try among GLYCAM ,GAFF. Among these three Glycam parameters seems better but still it produce the non-native pucker state with only 0.4 percentage so we plan to increase the simulation time.
- Are you sure that your modified residue can not be modelled with the existing parameters for protein and sugar?
Most of the parameters I have taken from CGenFF alone: is only from existing carbohydrate parameters but all the existing parameters standardized for classical most energetically favour 4C1 conformation.
- Badges
- Science topic