ProteinPool

The protein pool is generated by creating protein shapes using input parameter distributions for alpha (the concavity of the shapes) and num_points (the number of points used in shape hull generation). This file is saved as .pkl and loaded in DatasetGenerator to compute interactions.

class Dock2D.DatasetGeneration.ProteinPool.ParamDistribution(**kwargs)
__init__(**kwargs)

Unzip alpha (shape concavity) and number of points (number of points used to generate shape hulls) parameter distributions, then normalize probabilities.

Parameters

kwargs (list of tuples [(alpha, prob), ...] or [(num_points, prob), ...]) – parameter distributions for alpha and num_points

extract_params(param_name)

Extracts parameters and associated probabilities e.g. [(alpha, prob),…] or [(num_points, prob),…] from parameter distributions used in protein shape pool generation.

Parameters

param_name ([str](alpha or num_points)) – alpha or num_points.

Returns

two separate tuples of values and probabilities

normalize(param_name)

Normalize probabilities from parameter distribution.

Parameters

param_name – name of parameter to normalize

Returns

alpha and num_points normalized probabilities

class Dock2D.DatasetGeneration.ProteinPool.ProteinPool(proteins)
__init__(proteins)
Parameters

proteins – protein shape pools generated

classmethod generate(num_proteins, params, size=50)

Generate protein shapes to be used in protein pool by randomly sampling alpha and num_points using their associated probabilities from param parameter distributions.

Parameters
  • num_proteins – number of proteins to generate in the pool

  • params – parameters used in shape generation, list of tuples [(alpha, prob),…] or [(num_points, prob),…]

  • size – size of the box to generate a shape within.

Returns

protein pool shapes and corresponding individual shape parameters

classmethod load(filename)

Load protein pool .pkl

Parameters

filename – protein pool filename.pkl

Returns

protein pool

save(filename)

Save protein pool shapes and corresponding params to .pkl

Parameters

filename – protein pool filename.pkl