PANDA (version 0.9.9)

SignificantPairs: Identify functionally associated protein pairs

Description

This function uses the PAND distribution to calculate p-values (or probabilities) for each pair of proteins with at least one common neighbor in a protein-protein interaction network. It returns protein pairs with significant p-values (or probabilities).

Usage

SignificantPairs(PPIdb, Lambda=2, pvalue=FALSE)

Arguments

PPIdb

A two-column data frame consisting of binary interactions where each row represents an undirected edge (interaction) between two nodes (proteins) from two columns.

Lambda

Weight of direct interactions in the PAND algorithm. Lamda has different biological meanings with different values: "0" indicates that a direct link gives no information on the functional association; "1" indicates that a direct link is as informative as sharing one common neighbor (defined as an indirect link) on the functional association; "2" (or greater integer) indicates that a direct link is more informative than an indirect link. Since direct links should represent stronger evidence of functional associations than indirect links, we recommend using "2" as Lamda.

pvalue

logical; if TRUE, p-values for protein pairs will be calculated using PAND; if FALSE, probabilities will be calculated.

Value

This function returns a data frame with column names: "Sym_A", "Sym_B", "p_value" and "CommonNeighbor". "Sym_A" and "Sym_B" are a pair of nodes that share a significant functional linkage. "p_value" or "Probability" (calculated by the PAND algorithm) measures the significance of the linkage. "CommonNeighbor" is the number of shared nodes.

See Also

ProteinCluster, KEGGpredict, GOpredict, SignificantSubcluster

Examples

## not run
## data(dfPPI)
## OrderAll=SignificantPairs(dfPPI)