run_xmwas: run_xmwas

Description Usage Arguments Author(s)

Description

The function uses sPLS or PLS and network functions in mixOmics package to perform pairwise integrative and correlation analysis. The pairwise graphs are merged using igraph and community detection is performed using the Multilevel clustering algorithm. Association networks can be visualized in R or using Cytoscape.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
run_xmwas(xome_fname = NA, yome_fname = NA, zome_fname = NA, wome_fname = NA,
 outloc = NA,
 class_fname = NA, Xome_data = NA, Yome_data = NA, Zome_data = NA, Wome_data = NA, 
 classlabels = NA, xmwasmethod = "spls", plsmode = "canonical", max_xvar = 5000,
  max_yvar = 5000, max_zvar = 5000, max_wvar = 5000, rsd.filt.thresh = 1,
   all.missing.thresh = 0.1, missing.val = 0, corthresh = 0.4, keepX = 1000,
    keepY = 1000, keepZ = 1000, keepW = 1000, pairedanalysis = FALSE,
     optselect = TRUE, rawPthresh = 0.05, numcomps = 10,
      net_edge_colors = c("blue", "red"), net_node_colors = c("orange", "green", "blue", "gold"),
       Xname = "X", Yname = "Y", Zname = "Z", Wname = "W", 
       net_node_shape = c("circle", "rectangle", "triangle", "star"), seednum = 100,
        label.cex = 0.3, vertex.size = 6, max_connections = NA, 
        centrality_method = "eigenvector", use.X.reference = FALSE,
         removeRda = TRUE, compare.classes = TRUE, class.comparison.allvar = TRUE, ...)

Arguments

xome_fname

Full path with filename for dataset A. Default: NA; The software uses the value provided for Xome_data when this is set to NA.

yome_fname

Full path with filename for dataset B Default: NA; The software uses the value provided for Yome_data when this is set to NA.

zome_fname

Full path with filename for dataset C Default: NA; The software uses the value provided for Zome_data when this is set to NA.

wome_fname

Full path with filename for dataset D Default: NA; The software uses the value provided for Wome_data when this is set to NA.

Xome_data

Data matrix for dataset A. Run: data(exnci60); head(exnci60$mrna) to see how to format data matrices.

Yome_data

Data matrix for dataset B

Zome_data

Data matrix for dataset C

Wome_data

Data matrix for dataset D

outloc

Output directory

classlabels

Data matrix with phenotype information. Set to NA if this information is not available. see: data(classlabels_casecontrol) for case vs control design or data(classlabels_repeatmeasures) for repeat measures

class_fname

File with phenotype information. Set to NA if this information is not available. see: data(classlabels_casecontrol) for case vs control design or data(classlabels_repeatmeasures) for repeat measures

xmwasmethod

Method for data integration. eg: "pls": partial least squares regression "spls": sparse partial least squares regression "o1pls": orthogonal partial least squares regression

plsmode

"canonical" for bi-directional relationships; "regression" for regression/predictive relationships

max_xvar

Maximum number of X variables to select based on relative standard deviation (RSD). e.g. 10000

max_yvar

Maximum number of Y variables to select based on relative standard deviation (RSD). e.g. 10000

max_zvar

Maximum number of Z variables to select based on relative standard deviation (RSD). e.g. 10000

max_wvar

Maximum number of W variables to select based on relative standard deviation (RSD). e.g. 10000

rsd.filt.thresh

Relative standard deviation (sd/mean) threshold

all.missing.thresh

Remove variables (rows) that do not meet the minimum threshold for presence of non-missing values. e.g. 0.8

missing.val

How are the missing values represented in the input data files? Default: 0

corthresh

Correlation threshold. eg: 0.7

keepX

Maximum number of X variables to select in sPLS. Note: keepX, keepY, keepZ, and keepW are only used when xmwasmethod is set to "spls"

keepY

Maximum number of Y variables to select in sPLS. Note: keepX, keepY, keepZ, and keepW are only used when xmwasmethod is set to "spls"

keepZ

Maximum number of Z variables to select in sPLS. Note: keepX, keepY, keepZ, and keepW are only used when xmwasmethod is set to "spls"

keepW

Maximum number of W variables to select in sPLS. Note: keepX, keepY, keepZ, and keepW are only used when xmwasmethod is set to "spls"

pairedanalysis

Are their repeated measurements? TRUE or FALSE

optselect

Find optimal number of PLS components. TRUE or FALSE

rawPthresh

p-value threshold calculated using Student's t-test. eg: 0.05

numcomps

Number of components to use in PLS model. eg: 3

net_edge_colors

Colors for edges.

net_node_colors

Colors for nodes.

Xname

Name for X dataset. eg: "Genes"

Yname

Name for Y dataset. eg: "Proteins"

Zname

Name for Z dataset. eg: "Metabolites"

Wname

Name for W dataset. eg: "EnvironmentalExposures"

net_node_shape

Shapes for nodes.

seednum

Seed for random number generator used for plotting the network.

label.cex

Size of the labels. eg: 0.8

vertex.size

Size of the nodes.

max_connections

Maximum number of associations to include in the network. The connections between nodes are ranked based on the strength of association (+ve and -ve). Only the top "max_connections" connections are shown and used for centrality and community detection analyses. Set max_connections=NA if you want to use all connections. e.g. 1e5, 1e6, or NA

centrality_method

Method for centrality analysis. Options: 1) "eigenvector" for eigenvector centrality, which is based on the number and quality of connections - centrality scores range from 0 to 1, where 1 means high centrality and 0 means low or no centrality. Nodes/vertices with high centrality scores are connected to many other nodes, which are in turn connected to many other nodes. Please see igraph::eigen_centrality function for more details.

2) "betweenness" for betweenness centrality, which is based on the number of shortest paths going through a node/vertex - centrality scores are normalized and scaled to 0 to 1 range, where 1 means high betweenness centrality and 0 means low or no betweenness centrality. Please see igraph::betweenness function for more details.

3) "degree.count" based on the number of connections of a node. -centrality scores are normalized and scaled to 0 to 1 range. High centrality means more connections. Please see igraph::degree function for more details.

4) "degree.weight" is based on the sum of absolute weights of connections of a node. -centrality scores are normalized and scaled to 0 to 1 range. High centrality means stronger connections.

5) "closeness" based on the reciprocal of the sum of the distances of a node/vertex to all other nodes. -centrality scores are normalized. High centrality means the node is closer to all other nodes.

use.X.reference

TRUE or FALSE if you want to use Xome_data as reference. If TRUE, only X<->Y, X<->Z, and X<->W pairwise analysis will be performed.

removeRda

TRUE or FALSE; set to TRUE if you want to remove the intermediate files.

compare.classes

TRUE or FALSE; set to TRUE if you want to compare individual classes as provided in class labels file.

class.comparison.allvar

TRUE or FALSE; set to TRUE if all nodes shown

Author(s)

Karan Uppal [email protected]


joneslabemory/xMWAS documentation built on May 21, 2019, 1:44 p.m.