AlchemicalSamplerSettings

pydantic model openfe.protocols.settings.AlchemicalSamplerSettings

Settings for the Equilibrium Alchemical sampler, currently supporting either MultistateSampler, SAMSSampler or ReplicaExchangeSampler.

Create a new model by parsing and validating input data from keyword arguments.

Raises ValidationError if the input data cannot be parsed to form a valid model.

field sampler_method = 'repex'

Alchemical sampling method, must be one of; repex (Hamiltonian Replica Exchange), sams (Self-Adjusted Mixture Sampling), or independent (independently sampled lambda windows). Default repex.

field online_analysis_interval: int | None = 250

MCMC steps (i.e. IntegratorSettings.n_steps) interval at which to perform an analysis of the free energies.

At each interval, real time analysis data (e.g. current free energy estimate and timing data) will be written to a yaml file named <SimulationSettings.output_name>_real_time_analysis.yaml. The current error in the estimate will also be assed and if it drops below AlchemicalSamplerSettings.online_analysis_target_error the simulation will be terminated.

If None, no real time analysis will be performed and the yaml file will not be written.

Must be a multiple of SimulationSettings.checkpoint_interval

Default 250.

field online_analysis_target_error = <Quantity(0.0, 'boltzmann_constant * kelvin')>

Target error for the online analysis measured in kT. Once the free energy is at or below this value, the simulation will be considered complete. A suggested value of 0.2 * unit.boltzmann_constant * unit.kelvin has shown to be effective in both hydration and binding free energy benchmarks. Default 0.0 * unit.boltzmann_constant * unit.kelvin, i.e. no early termination will occur.

field online_analysis_minimum_iterations = 500

Number of iterations which must pass before online analysis is carried out. Default 500.

field n_repeats: int = 3

Number of independent repeats to run. Default 3

field flatness_criteria = 'logZ-flatness'

SAMS only. Method for assessing when to switch to asymptomatically optimal scheme. One of [‘logZ-flatness’, ‘minimum-visits’, ‘histogram-flatness’]. Default ‘logZ-flatness’.

field gamma0 = 1.0

SAMS only. Initial weight adaptation rate. Default 1.0.

field n_replicas = 11

Number of replicas to use. Default 11.