pyNBS.network_propagation

network_propagation is a module that contains functions for performing the random-walk based network propagation method used in the NBS algorithm. For more details specifically about how the random-walk network propagation algorithm works, please see the documentation for the network_propagation function.

Functions in this module

This module contains the following functions (bolded functions are key pyNBS functions):

• fast_random_walk: This is a helper function called inside the network_propagation specifically used to perform the network smoothing for a connected component in the molecular network.

• network_kernel_propagation: This is a modified version of the network_propagation function where a pre-computed "kernel" or random-walk propagation influence matrix is given to speed up the process of performing network propagation over the same network many times. This function can be called within the NBS_single function.

• network_propagation: This is the main function called to perform network-smoothing of a matrix of binary somatic mutations over a molecular network. It implements a closed-form version of the random-walk-based network propagation method. This function can be called within the NBS_single function.

• normalize_network: This is a helper function called inside the network_propagation function that performs degree normalization for the adjacency matrix of the molecular network.