Hi Tiger,
You can't do a detailed prediction of a mutated structure with simple sidechain substitution in ChimeraX.  The sidechain substitution only allows a simple first-order estimate of whether the substitution is reasonably favorable or not, e.g. whether it causes severe clashes or loss/gain of H-bonds, assuming the rest of the protein stays the same.  Of course, you cannot assume the rest of the protein stays the same... it might not even fold!!

Chimera (but not ChimeraX) has a minimize structure tool allowing a local minimization, and a molecular dynamics tool for short dynamics, but at best, they give only a second-order estimate of the effect of a mutation.  Minimization might relieve local distortions, but major changes in fold are very unlikely to be accessed, even with dynamics.  It will not predict the change in stability in any quantitative sense.

That said:
ChimeraX does have "minimize" and "tug" mouse modes that use OpenMM dynamics (see Right Mouse toolbar) but the entire atomic model must be parametrized in OpenMM. Currently, this requires a fully protonated structure containing only standard residues and/or water. Protons can be added beforehand with the command addh.  Details:
<https://rbvi.ucsf.edu/chimerax/docs/user/tools/mousemodes.html#openmm>

ChimeraX with the ISOLDE plugin can do dynamics, but that is generally in the context of having a cryoEM map to model into, rather than a protein alone.

I hope this helps,
Elaine
-----
Elaine C. Meng, Ph.D.                     
UCSF Chimera(X) team
Department of Pharmaceutical Chemistry
University of California, San Francisco