MWAS_network: Visualize MWAS results in a correlation-based metabolic...

In MWASTools: MWASTools: an integrated pipeline to perform metabolome-wide association studies

Description

This function allows visualizing MWAS results in a correlation-based metabolic network. The network is an undirected graph where the nodes represent the metabolites, and the edges represent a co-abundance relationship between pairs of nodes. Different node parameters (e.g. color, size) can be customized based on MWAS results.

Usage

MWAS_network (metabo_SE, MWAS_matrix, alpha_th = 0.05, cor_th = 0.25,
              file_name = "Correlation", res_cor = 2)

Arguments

metabo_SE	SummarizedExperiment object. See "MWAS_SummarizedExperiment()".
MWAS_matrix	numeric matrix generated by the function "MWAS_stats()".
alpha_th	numeric value indicating MWAS significance threshold.
cor_th	numeric value indicating the co-abundance similarity threshold. Thus, two metabolites will be linked in the network if the absolute correlation (Pearson) between them exceeds cor_th.
file_name	character string indicating the name given to the cytoscape files that will be exported to the working directory.
res_cor	numeric value restricting the number of decimals of the correlation of coefficients used to build the edges of the network.

Value

A correlation based-metabolic network formalized as a weigthed igraph object. This igraph object contains two node attributes: "score" and "color". "score" is a vector containing the MWAS score (-log10(pvalue)*estimate sign) of each metabolite. "color" is a vector indicating the color of each node based on MWAS results ("cornflowerblue": "downregulation", "gray":"no change", "firebrick1":"upregulation"). These attributes can be used to customize node parameters based on MWAS results. The function also exports a network file ("Correlation_NetworkFile.txt") and an attribute file ("Correlation_AttributeFile.txt") of MWAS scores, which can be imported into cytoscape to visualize the network.

References

Csardi G, Nepusz T. (2006). The igraph software package for complex network research. InterJournal, Complex Systems, 1695.

Examples

## Load data
data(targetMetabo_SE)

## Test for association between diabetes and target_metabolites
T2D_model <- MWAS_stats (targetMetabo_SE, disease_id = "T2D",
                         confounder_ids = c("Age", "Gender", "BMI"),
                         assoc_method = "logistic")

## Build correlation-based metabolic network
net_T2D <- MWAS_network(targetMetabo_SE, T2D_model, file_name = "MWAS_T2D",
                        cor_th = 0.30)

## Visualize network using the igraph package
# library(igraph)
# plot(net_T2D, vertex.size = abs(V(net_T2D)$score*6)) # node size based on scores

# plot(net_T2D, vertex.size = abs(V(net_T2D)$score*6),
#      edge.label = E(net_T2D)$weight) # show edge labels

Example output

MWASTools documentation built on Nov. 8, 2020, 5:07 p.m.