GRNPY - Gene Regulatory Network .py

This python package aims to be a Gene Regulatory Network Graph generator that is fully customizable parameters such as sparsity, directed graphs, weight of edges, type of system (binary or spin) and etc.

Aug 23: Work in progress

• take a dict or list of lists as input and converting them to graphs
• take a graph as an input to the system

Usage

In the terminal, type:

pip install grnpy

In the code, import grnpy and the class BoltzmannGRN

from grnpy import BoltzmannGRN

Functions

Generate_Gene_Network(self, n, sparsity, on_p, directed):

Generates a graph based on the parameters :

• n - int - number of nodes
• sparsity - float - define the sparsity of the graph (1 being sparse, and 0 being dense) [0,1]
• on_p - float - the probability of a node being on at the beginning [0,1]
• directed - bool - if the graph is directed

returns graph

make_wmatrix(self, G, size, weight):

Generates an edge weight matrix

• size - int - size of the graph
• weight - defined set weight of the graph (going to be removed or chagned in later versions)

returns np array

getHit(self, G, gene_i, field)

Generates the h value at timestep t for node i

returns H_{i,t}

probability(self, G, gene, hit, field)

returns probability of a node being at state "1"

returns p

change_state(self, G, field)

Updates the state of all nodes in G

returns NONE

state(self, G, node)

returns the state of node n in graph G returns state(node)

update(self, size, G, timesteps, field)

For each timestep, iterate through all nodes in G and update their state

• timestep - int - number of timesteps to be updated
• size - int - size of the graph
• G - nx.Graph - Graph of nodes
• field - float - constant that is added as a probablity
• time - list - list of timesteps
• states_mean - list - mean state value across timesteps
• list_states - list - list of list containing the specific states of each node for each timestep

returns time, states_mean, list_states

show_plot(self, time, states_mean)

shows plot of state transitions

Define and Use

Define the model

BoltzmannModel = BoltzmannGRN(size, gene_start_probability, weight, sparsity, directed, 'binary')

Define the graph

BoltzmannGraph = BoltzmannModel.graph

Drawing the graph

nx.draw(BoltzmannGraph)
plt.show()

Run the state transition model

time, states_mean, states = BoltzmannModel.update(size, BoltzmannGraph, 100, field)