GROMACS version: 2023.3-Homebrew
GROMACS modification: No
Hi all,
I was wondering what is regarded as the best/most popular force field for dealing with miRNA (or RNA in general). I have read in this forum that the AMBER ones are quite outdated and should not be used, so are CHARMM and GROMOS my only options? And between the various CHARMM and GROMOS force fields, which is best suited for nucleic acids?
Thanks for your help.
Hi,
All RNA force fields have their issues, and there is an excellent review paper detailing the main challenges in simulating RNAs (well, with 120+ pages it’s more a book than a paper). Since then, though, a new and quite successful force field from DE Shaw’s group/DESRES was released.
In general, different questions might require different approaches, and so e.g. capturing a spontaneous binding event required rescaling LJ interactions, some stability issues could be fixed by adjusting h-bond strengths, and when e.g. NMR data is available, orthogonal observables can be improved by enforcing agreement with experimental data.
But for generic out-of-the-box applications, I honestly never saw an advantage in CHARMM or GROMOS over the two cutting-edge Amber family FFs that are used as baseline for most benchmarks, chiOL3 (that tends to over-collapse otherwise flexible structures but fares well in general) and DESRES (that mostly solves the hydrophobic collapse problem but hasn’t been that well explored by other groups yet). While there are chiOL3 versions for Gromacs, for Gromacs-compatible DESRES files you might need to directly contact the company (although the parameters are also publicly available in recent Amber distributions and for Desmond).
Thanks a lot! Your advice was very helpful.