For high-dimensional parameter spaces, methods such as Approximate Bayesian Computation are popular for uncertainty quantification in place of more expensive, exact methods like MH-MCMC.
Evaluate the accuracy of such methods, and consider whether one could be added to PyBNF.