Mixed resolution AA/CG MD simulations in Gromacs

Developers discussions
1

Hello Gromacs Community,

I am interested in exploring the mixed-resolution computational methodologies for the modeling of large trans-membrane protein complexes (up to 10,000 AAs) through Gromacs. My approach would be to employ coarse-grained models such as Martini for the majority of the system. For segments of the system where key interactions occur, I would like to use all-atom simulations.

I am considering the ONIOM model as a potentially viable approach, albeit with possible modifications needed to adjust it from the QM/MM to an AA/CG approach. That being said, I am unsure of ONIOM’s availability in the latest versions of Gromacs.

Could you please provide guidance on the feasibility and practicality of this method? Moreover, are there other techniques available or that could be straightforwardly implemented in Gromacs to address similar issues?

For context, I have adequate programming abilities and some comprehension of the Gromacs software. If necessary, I might be able to enlist assistance from our institution’s computer science department.

Looking forward to your insights and suggestions. Thank you.

2

In QMMM parts of the system are either QM or MM and do not change description dynamically. If that’s what you want you don’t need to modify GROMACS at all. You only need to construct suitable topology and coordinate files.