rw — random walk

Random-walk simulation over the network.

bobaT.rw.random_walks(attractor_dict, rules, regulators_dict, nodes, save_dir, radius=2, perturbations=False, iters=1000, max_steps=500, stability=False, reach_or_leave='leave', random_start=0, on_nodes=[], off_nodes=[], basin=0, overwrite_walks=True, overwrite_perturbations=True, verbose=2)

Wrapper function to perform random walks.

Parameters:

• attractor_dict (dictionary) – Dictionary of attractors used to get to steady states

• rules (dictionary) – Dictionary of probabilistic rules for the regulators

• regulators_dict (dictionary) – Dictionary of relevant regulators

• nodes (list) – List of nodes in the transcription factor network

• save_dir (string or path-like object) – Path/directory to save output

• radius (int or list) – Single or multiple radii values for random walks

• perturbations (bool) – Whether to perform random walks with perturbations or not

• iters (int) – Number of iterations of random walks

• max_steps (int) – Max number of steps to take in random walks

• stability (bool) – Whether to perform stability testing with multiple radii or not

• reach_or_leave (string) – Define what type of random walk to perform

• random_start (int) – Whether to perform walks with a random list of start states. If >0, run that many random starts

• on_nodes (list) – Define ON nodes of a perturbation

• off_nodes (list) – Define OFF nodes of a perturbation

• basin (int) – Define a basin for random_walk_until_reach_basin

• overwrite_walks (bool, default True) – If False and walks/{start_idx} already exists, do not overwrite the results. Instead move on to the next start_idx. Note that if this is set to False and walks/{start_idx} already exists, the code skips all random walks for this start_idx (including any perturbations and stability testing)

• overwrite_perturbations (bool, default = True) – If False and perturbations/{start_idx} already exists, do not overwrite the results. Instead move on to the next start_idx.

Return type:

None

bobaT.rw.random_walks_parallel(attractor_dict, rules, regulators_dict, nodes, save_dir, radius=2, perturbations=False, iters=1000, max_steps=500, stability=False, reach_or_leave='leave', random_start=0, on_nodes=[], off_nodes=[], basin=0, overwrite_walks=True, overwrite_perturbations=True, cpu_usage=0.5, cpus=None, verbose=1)

Wrapper function to perform random walks.

Parameters:

• attractor_dict (dictionary) – Dictionary of attractors used to get to steady states

• rules (dictionary) – Dictionary of probabilistic rules for the regulators

• regulators_dict (dictionary) – Dictionary of relevant regulators

• nodes (list) – List of nodes in the transcription factor network

• save_dir (string or path-like object) – Path/directory to save output

• radius (int or list) – Single or multiple radii values for random walks

• perturbations (bool) – Whether to perform random walks with perturbations or not

• iters (int) – Number of iterations of random walks

• max_steps (int) – Max number of steps to take in random walks

• stability (bool) – Whether to perform stability testing with multiple radii or not

• reach_or_leave (string) – Define what type of random walk to perform

• random_start (int) – Whether to perform walks with a random list of start states. If >0, run that many random starts

• on_nodes (list) – Define ON nodes of a perturbation

• off_nodes (list) – Define OFF nodes of a perturbation

• basin (int) – Define a basin for random_walk_until_reach_basin

• overwrite_walks (bool, default True) – If False and walks/{start_idx} already exists, do not overwrite the results. Instead move on to the next start_idx. Note that if this is set to False and walks/{start_idx} already exists, the code skips all random walks for this start_idx (including any perturbations and stability testing)

• overwrite_perturbations (bool, default = True) – If False and perturbations/{start_idx} already exists, do not overwrite the results. Instead move on to the next start_idx.

Return type:

None

bobaT.rw.simple_random_walk(stg, edge_weights, start_idx, steps)

Perform random walk on a state transition graph with known edge weights.

Paramters

stggraph tools Graph() object

State transition graph

start_idxint

Index of the vertext to start the walk

stepsint

Walk length

returns:

verts – Path of vertices taken during random walk

rtype:

list

bobaT.rw.random_walk_until_leave_basin(start_state, rules, regulators_dict, nodes, radius=2, max_steps=10000, on_nodes=[], off_nodes=[])

Parameters:

• start_state (int) – Index of attractor to start walk from

• rules (dictionary) – Dictionary of probabilistic rules for the regulators

• regulators_dict (dictionary) – Dictionary of relevant regulators

• nodes (list) – List of nodes in the transcription factor network

• radius (int) – Radius to stay within during walk

• max_steps (int) – Max number of steps to take in random walks

• on_nodes (list) – Define ON nodes of a perturbation

• off_nodes (list) – Define OFF nodes of a perturbation

Returns:

• walk (list) – Path of vertices taken during random walk

• Counter(walk) – Histogram of walk

• flipped_nodes (list) – Transcription factors that flipped during walk

• distances (list) – Starting state to next step in walk

bobaT.rw.random_walk_until_reach_basin(start_state, rules, regulators_dict, nodes, radius=2, max_steps=10000, on_nodes=[], off_nodes=[], basin=1)

Parameters:

• start_state –

  .

• rules (dictionary) – Dictionary of probabilistic rules for the regulators

• regulators_dict (dictionary) – Dictionary of relevant regulators

• nodes (list) – List of nodes in the transcription factor network

• radius (int) – Radius to stay within during walk

• max_steps (int) – Max number of steps to take in random walks

• on_nodes (list) – Define ON nodes of a perturbation

• off_nodes (list) – Define OFF nodes of a perturbation

• basin (int or list) – List of attractors to reach (or a single attractor) by state index (integer)

Returns:

• walk (list) – Path of vertices taken during random walk

• Counter(walk)

• flipped_nodes (list)

• distances (list) – All distances to basin

bobaT.rw.long_random_walks(starting_attractors, attractor_dict, rules, regulators_dict, nodes, save_dir, on_nodes=[], off_nodes=[], max_steps=2000, iters=100, overwrite_walks=False)

Alternative way to do random walks (until reach basin). Instead of looking for a specific basin, keep walking some length of steps. Can be used for visualizing effects of perturbations using bb.plot.plot_random_walks.

Parameters:

• starting_attractors (list) – name of the attractors to start the walk from (key in attractor_dict)

• attractor_dict (dict()) – Dictionary of attractors

• rules (dict()) – Rules from BooleaBayes rule fitting

• regulators_dict (dict()) – Dictionary of regulators from rule fitting

• nodes – list of nodes in the network

• save_dir (str) – Directory to save output

• on_nodes (list, optional) – activating perturbations to run simulations for, defaults to []

• off_nodes (list, optional) – knockdown perturbations to run simulations for, defaults to []

• max_steps (int, optional) – Length of random walks, defaults to 2000

• iters (int, optional) – Number of iterations to run, defaults to 100

• overwrite_walks (bool, optional) – If false, don’t rewrite walks if the folder already exists, defaults to False