findsub: Find Highly Connected Subnetworks after Eliminating 'Weak'...

Description Usage Arguments Details Value


Find Highly Connected Subnetworks after Eliminating 'Weak' Edges, and calculating aggregate 'heat' scores for these subnetworks


findSub(method = "Leis", EGG, minOrd, noLog = FALSE, silent = FALSE)



method of edge removal. Currently only "Leis" (Leiseron et al. 2015) method of edge removal available. See details for more information


an igraph object of format AGG with score annotations


minimum number of vertices per generated subgraph


if FALSE, save process to an .Rdata log file


if FALSE, print process to console in real time


Depending on method used, edges are removed from EGG based upon the assigned 'delta' attribute and their 'Influence' scores(see next paragraph). In the "Leis" method (Leiseron et al. 2015), edges with Influence scores below the value of 'delta' are eliminated. Then, highly connected components of the EGG graph made of the remaining edges are found using igraph::components(), and subgraphs generated for these components. The vertices in these subgraphs come with the Gene Scores (score can differ depending on method), which for each subgraph's vertices are summed to find an aggregate heat score for each individual graph. E.g., if a subgraph's vertices had scores of 5,8,11 the aggregate heat score would be 24. The aggregate heat score is then assigned as an attribute aggHeatScore. Subgraphs are stored in a list of subgraphs sN, which is returned. sN is ordered from highest aggHeatScore to lowest aggHeatScore of the subgraph. All subgraphs in sN contain same attributes as EGG and have aggHeatScore (aggregate heat score) added as an attribute.

EGG must have the following: vertices: genes with HGNC symbol and Gene Score ('heat', see documentation for score function for more details) as the minimal vertex attributes; edges: directed edges from 1 gene to the next, derived from protein protein interaction data. Edges' 'from' and 'to' columns (edge vertices) named with HGNC symbols. Minimum attributes for edges include 'Influence', a score weighting edges depending on weighting method (for our purposes this will likely ). EGG must also come with a delta attribute (integer or of type double) (assigned threshold for edge removal regardless of method used in generating edge scores) and an EGGversion, a graph attribute with a character vector of length 1. Unless further changes are made, EGGversion must be "1.0".


A list of igraph objects, with "Influence" edge attributes, "Gene Score" vertex attributes, "EGGversion" graph Attribute, "Aggregate Heat score graph attribute

hyginn/ekplektoR documentation built on May 17, 2017, 12:08 a.m.