sampling_utils

Auxiliary functions for setting up the conformational sampling process.

Functions

setup_sampling_logging(sampling_log)

Setup logging handlers and files for PyRosetta sampling and Ray.

setup_sampling_parameters(parameters_file)

Update the parameters dictionary before sampling.

setup_sampling_initial_pose(params, sampling_log)

Create initial Pose for sampling and apply any required constraints.

sample_pdb(ppose, databases, targets, output_path, ...)

Sample dihedral angles from a database into target regions of a given structure.

Module Contents

ensemblify.generation.ensemble_utils.sampling_utils.setup_sampling_logging(sampling_log)

Setup logging handlers and files for PyRosetta sampling and Ray.

Parameters:

sampling_log (str) – Path to sampling .log file.

Returns:

logger (logging.Logger):

The Logger object associated with the sampling .log file

ray_log (str):

Filepath to .log file with Ray log messages.

pyrosetta_log (str):

Filepath to .log file with PyRosetta log messages.

Return type:

tuple[logging.Logger,str,str]

ensemblify.generation.ensemble_utils.sampling_utils.setup_sampling_parameters(parameters_file)

Update the parameters dictionary before sampling.

In the ‘targets’ parameter, change a target from e.g. [1,54] to (range(1,55)). If using an AlphaFold model as a starting structure, keep in the sampling ranges only regions of at least a certain contiguous size where each residue’s pLDDT is below a threshold. Targets, secondary structure biases and contacts are also updated to tuples instead of lists.

Parameters:

parameters_file (str) – Path to parameters file following the Ensemblify template.

Returns:

The updated params dictionary.

Return type:

dict

ensemblify.generation.ensemble_utils.sampling_utils.setup_sampling_initial_pose(params, sampling_log)

Create initial Pose for sampling and apply any required constraints.

Constraints applied to the pose are stored in a constraints.cst file stored in the same directory as the sampling_log file.

Parameters:

  • params (dict) – Parameters following the Ensemblify template.

  • sampling_log (str) – Path to the sampling .log file.

Returns:

Object to be used as the starting structure for the sampling process.

Return type:

pyrosetta.rosetta.core.pose.Pose

ensemblify.generation.ensemble_utils.sampling_utils.sample_pdb(ppose, databases, targets, output_path, job_name, decoy_num='', log_file=None, ss_bias=None, variance=0.1, sampler_params={'MC': {'temperature': 200, 'max_loops': 200}}, scorefxn_id='score0', scorefxn_weight=1.0, minimizer_id='dfpmin_armijo_nonmonotone', minimizer_tolerance=0.001, minimizer_maxiters=5000, minimizer_finalcycles=5, cst_weight=1, cstviolation_threshold=0.015, cstviolation_maxres=20)

Sample dihedral angles from a database into target regions of a given structure.

Parameters:

  • ppose (pyrosetta.distributed.packed_pose.core.PackedPose) – Reference to the initial structure.

  • databases (dict[str,dict[str,pd.DataFrame]]) – Reference to the databases dictionary.

  • targets (dict[str,tuple[tuple[str,tuple[int,...],str,str]]]) – Dictionary detailing the target regions for sampling in each chain.

  • output_path (str) – Path to directory where sampled structures will be written to.

  • job_name (str) – Prefix identifier for generated structures.

  • decoy_num (str, optional) – Identifier to differentiate between different decoys of the same batch in a multiprocessing context. Defaults to ‘’.

  • log_file (str, optional) – Path to the PyRosetta .log file. Defaults to ‘pyrosetta.log’ in current working directory.

  • ss_bias (tuple[tuple[tuple[str,tuple[int,int],str],...],int], optional) – Secondary Structure Bias with the desired percentage of total structures to respect this bias. Defaults to None.

  • variance (float, optional) – New dihedral angle values inserted into sampling regions are sampled from a Gaussian distribution centered on the value found in database and percentage variance equal to this value. Defaults to 0.10 (10%).

  • sampler_params (dict[str,dict[str,int]], optional) – Parameters for the used sampler, assumes MonteCarloSampler is used. Defaults to {‘MC’:{‘temperature’:200,’max_loops’:200}}.

  • scorefxn_id (str, optional) – PyRosetta ScoreFunction identifier. Must pertain to a .wst weights file present in /…/pyrosetta/database/scoring/weights/ . Defaults to ‘score0’.

  • scorefxn_weight (float, optional) – Weight for the repulsive Van der Waals term in the ScoreFunction. Will only have an effect if the ScoreFunction has a repulsive Van der Waals term. Defaults to 1.0.

  • minimizer_id (str, optional) – PyRosetta minimization algorithm identifier used in MinMover. Defaults to ‘dfpmin_armijo_nonmonotone’.

  • minimizer_tolerance (float, optional) – Tolerance value for the PyRosetta MinMover object. Defaults to 0.001.

  • minimizer_maxiters (int, optional) – Maximum iterations value for the PyRosetta MinMover object. Defaults to 5000.

  • minimizer_finalcycles (int, optional) – Number of times to apply the MinMover to our final structure. Defaults to 5.

  • cst_weight (int, optional) – Weight of the AtomPairConstraint term in the ScoreFunction. Defaults to 1.

  • cstviolation_threshold (float, optional) – Any residue with AtomPairConstraint score term value above this threshold is considered in violation of the applied constraints. Defaults to 0.015.

  • cstviolation_maxres (int, optional) – Number of residues allowed to be above the constraint violation threshold. Defaults to 20.

Returns:

Path to the sampled .pdb structure. Only written and returned if the sampled structure is valid (does not violate the applied constraints). Otherwise, return None.

Return type:

str | None