Package 'BioNAR'

Description

Copy edge attributes from one graph to another

Usage

addEdgeAtts(GG, gg)

Arguments

Value

annotated version of gg igraph object

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
GG <- igraph::read_graph(file, format="gml")
gg<-findLCC(GG)
gg <- addEdgeAtts(GG, gg)
edge_attr_names(gg)

Description

For the protein-protein interaction (PPI) or disease gene interaction (DGN) graphs that have EntrezID as a vertex name this function extract standard name from org.Hs.eg.db and annotate vertices.

Usage

annotateGeneNames(gg, orgDB = org.Hs.eg.db, keytype = "ENTREZID")

Arguments

Details

If vertex name attrubite stores not EntrezID or network is build not from human genes, other OrgDb-class object could be provided in orgDB and one of keytypes from that object that correspond to the nature of the vertex name attrubite could be provided in the keytype attribute.

If for some vertices name attrubite does not match keys with particular keytypes in the orgDB object, empty string is added as GeneName.

Value

igraph object with new vertex attribute GeneName

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
agg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(agg)$name == '80273')
paste(V(agg)$GeneName[idx], 'GRPEL1')

Description

The function loads an annotation data matrix called annoF, which contains three columns; the first containing gene Entrez IDs, the second gene GO BP ID terms, the third gene GO BP description terms. The function then performs a many-to-one mapping of each matrix row to a network vertex using matching Entrez IDs, filling the vertices attributes GO_BP_ID and GO_BP.

Usage

annotateGoBP(gg, annoF, idatt = "name")

Arguments

Value

annotated igraph object

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
sfile<-system.file("extdata", "flatfile.go.BP.csv", package = "BioNAR")
goBP <- read.table(sfile, sep="\t", skip=1, header=FALSE,
strip.white=TRUE, quote="")
sgg <- annotateGoBP(gg, goBP)

Description

The function loads an annotation data matrix called annoF, which contains three columns; the first containing gene Entrez IDs, the second gene GO ID terms, the third gene GO CC description terms. The function then performs a many-to-one mapping of each matrix row to a network vertex using matching Entrez IDs, filling the vertices attributes GO_CC_ID and GO_CC.

Usage

annotateGoCC(gg, annoF, idatt = "name")

Arguments

Value

annotated igraph object

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
sfile<-system.file("extdata", "flatfile.go.CC.csv", package = "BioNAR")
goCC <- read.table(sfile, sep="\t", skip=1, header=FALSE,
strip.white=TRUE, quote="")
sgg <- annotateGoCC(gg, goCC)

Description

The function loads an annotation data matrix called annoF, which contains three columns; the first containing gene Entrez IDs, the second gene GO MF ID terms, the third gene GO MF description terms. The function then performs a many-to-one mapping of each matrix row to a network vertex using matching Entrez IDs, filling the vertices attributes GO_MF_ID and GO_MF.

Usage

annotateGoMF(gg, annoF, idatt = "name")

Arguments

Value

annotated igraph object

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
sfile<-system.file("extdata", "flatfile.go.MF.csv", package = "BioNAR")
goMF <- read.table(sfile, sep="\t", skip=1, header=FALSE,
strip.white=TRUE, quote="")
sgg <- annotateGoMF(gg, goMF)

Description

For the protein-protein interaction (PPI) or disease gene interaction (DGN) graphs that have EntrezID as a vertex name this function extract GeneOntolgy annotation from orgDB, which should be OrgDb-class, split them into three ontology group (MF,BP,CC) and annotate vertices with .

Usage

annotateGOont(gg, orgDB = org.Hs.eg.db, keytype = "ENTREZID", idatt = "name")

Arguments

Details

If vertex name attrubite stores not EntrezID or network is build not from human genes, other OrgDb-class object could be provided in orgDB and one of keytypes from that object that correspond to the nature of the vertex name attrubite could be provided in the keytype attribute.

If for some vertices name attrubite does not match keys with particular keytypes in the orgDB object, empty string is added as GeneName.

Value

igraph object with new vertex attribute GeneName

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
ggGO <- annotateGOont(gg)

Description

Function takes data from annoF matrix and add them to attributes InterPro_Family for term and InterPro_Family_ID for IDs.

Usage

annotateInterpro(gg, annoF, annoD, idatt = "name")

Arguments

Details

Function takes data from annoD matrix and add them to attributes InterPro_Domain for term and InterPro_Domain_ID for IDs.

Value

annotated igraph object

See Also

getAnnotationVertexList

Description

Function takes from anno matrix manually curated presynaptic genes functional annotation derived from Boyken at al. (2013) doi:10.1016/j.neuron.2013.02.027 and add them to attributes PRESYNAPTIC.

Usage

annotatePresynaptic(gg, anno, idatt = "name")

Arguments

Value

annotated igraph object

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
sfile<-system.file("extdata", "PresynAn.csv", package = "BioNAR")
pres <- read.csv(sfile,skip=1,header=FALSE,strip.white=TRUE,quote="")
gg <- annotatePresynaptic(gg, pres)

Description

The function loads an annotation data matrix of functional groups for schizopherina risk genes (1) called anno, which contains three columns; the first containing gene Entrez IDs, the second gene functional group ID terms, the third gene functional group description terms. The function then performs a many-to-one mapping of each matrix row to a network vertex using matching Entrez IDs, filling the SCHanno vertices attribute.

Usage

annotateSCHanno(gg, anno, idatt = "name")

Arguments

Details

References:

1. Lips E, Cornelisse L, Toonen R, Min J, Hultman C, the International Schizophernia Consortium, Holmans P, Donovan M, Purcell S, Smit A, Verhage M, Sullivan P, Visscher P, D P: Functional gene group analysis identifies synaptic gene groups as risk factor for schizophrenia. Molecular Psychiatry 2012,17:996–1006.

Value

annotated igraph object

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
afile<-system.file("extdata", "SCH_flatfile.csv", package = "BioNAR")
dis    <- read.table(afile, sep="\t", skip=1, header=FALSE,
strip.white=TRUE, quote="")
agg<-annotateSCHanno(gg, dis)

Description

The function loads a human disease annotation matrix called dis, which contains three columns; the first containing gene Entrez IDs, the second gene Human Disease Ontology (HDO) ID terms, the third gene HDO description terms. For human protein-protein interaction (PPI) or disease-gene networks (DGN) that have human Entrez IDs for the igraph vertex name attribute. The function then performs a many-to-one mapping of each matrix row to a network vertex using matching Entrez IDs, filling the vertices attributes TopOnto_OVG_HDO_ID and TopOnto_OVG.

Usage

annotateTopOntoOVG(gg, dis, idatt = "name")

Arguments

Value

annotated igraph object

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
# read HDO data extracted from hxin/topOnto.HDO.db for synaptic network
afile<-system.file("extdata", "flatfile_human_gene2HDO.csv",
package = "BioNAR")
dis    <- read.table(afile, sep="\t", skip=1, header=FALSE,
strip.white=TRUE, quote="")
agg<-annotateTopOntoOVG(gg, dis)

Description

Function to build and fill a vertex attribute given an igraph object. Where parameter 'name' is the new vertex attribute name and values are filled from a two column data.frame supplied to 'value' attribute. The first containing vertex name IDs, and the second the vertex annotation value.

Usage

annotateVertex(gg, name, values, idatt = "name")

Arguments

Details

As a first step all attributes with provided names will be removed.

Value

igraph object where vertex attribute name contains annotation terms separated by semicolon.

See Also

getAnnotationVertexList

Examples

g1 <- make_star(10, mode="undirected")
V(g1)$name <- letters[1:10]
m<-rbind(data.frame(ID=letters[1:10], terms=letters[1:10]),
data.frame(ID=letters[1:10], terms=LETTERS[1:10]))
g2<-annotateVertex(g1, name='cap', values=m)
V(g2)$cap

Description

This function suits more for updating calculated vertex properties rather than node annotation. For the later case use annotateVertex.

Usage

applpMatrixToGraph(gg, m)

Arguments

Details

Unlike annotateVertex, which is able to collapse multiple annotation terms, this function assume that vertex ID values are unique in the m matrix and corresponds to the name vertex attribute. If graph has no name vertex attribute error will be raised.

Value

modified igraph object

See Also

annotateVertex

Examples

g1 <- make_star(10, mode="undirected")
V(g1)$name <- letters[1:10]
m<-cbind(ID=letters[1:10],capital=LETTERS[1:10])
g1<-BioNAR::applpMatrixToGraph(g1,m)
V(g1)$capital

Description

The R package BioNAR, developed to step by step analysis of PPI network. The aim is to quantify and rank each protein’s simultaneous impact into multiple complexes based on network topology and clustering. Package also enables estimating of co-occurrence of diseases across the network and specific clusters pointing towards shared/common mechanisms.

Author(s)

Maintainer: Anatoly Sorokin [email protected]

Authors:

• Colin Mclean [email protected]

• Oksana Sorokina [email protected]

• J. Douglas Armstrong [email protected] [funder]

Other contributors:

• T. Ian Simpson [email protected] [contributor, funder]

See Also

Useful links:

• Report bugs at https://github.com/lptolik/BioNAR/issues/

Description

Wrapper for graph_from_data_frame function which will always return the largest connect component for a given network ff. The function will also annotated the edges in ff with PubMed data from kw if provided.

Usage

buildNetwork(ff, kw = NA, LCC = TRUE, simplify = TRUE)

Arguments

Value

igraph object of the largest connected component

Examples

f<-data.frame(A=c('A', 'A', 'B', 'D'), B=c('B', 'C', 'C', 'E'))
gg<-buildNetwork(f)
V(gg)$name

Description

This function will call calcClustering for each clustering algorithm given in our predefined list. In the event no clustering could be performed, warnings will be issued and no new vertex attribute added to the graph.

Usage

calcAllClustering(gg, weights = NULL)

Arguments

Value

new graph object with all membership results stored as a vertex attribute.

See Also

calcClustering

Examples

g1 <- make_star(10, mode="undirected")
V(g1)$name <- letters[1:10]
g1<-calcAllClustering(g1)
clusteringSummary(g1)

Helper function that uses getBridgeness to calculate graph node bridgeness values for selected algorithm and consensus matrix and save them as a graph attribute BRIDGENESS.<alg> with <alg> replaced by the selected algorithm name.

Description

Helper function that uses getBridgeness to calculate graph node bridgeness values for selected algorithm and consensus matrix and save them as a graph attribute BRIDGENESS.<alg> with <alg> replaced by the selected algorithm name.

Usage

calcBridgeness(gg, alg, conmat)

Arguments

Value

graph with additional attributes to store Bridgeness value

See Also

getBridgeness

Examples

library(BioNAR)
karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
set.seed(100)
gg <- calcClustering(karate, 'louvain')
cnmat <- makeConsensusMatrix(gg, N=10, alg = 'louvain', type = 2, mask = 10)
gg<-calcBridgeness(gg, alg = 'louvain', cnmat)
hist(V(gg)$BRIDGENESS.louvain)

Description

Calculate the vertex centrality measures (degree, betweenness, closeness, semi-local, etc....) for each graph vertex and store each result as new vertex attribute in the graph.

Usage

calcCentrality(gg, weights = NULL)

Arguments

Details

A wrapper function that first calls getCentralityMatrix, to calculate all vertex centrality measures, and then applpMatrixToGraph to store each centrality result as a new vertex attribute in the graph. The use of weights explained in details in getCentralityMatrix.

Value

modified igraph object

See Also

getCentralityMatrix()

Examples

data(karate,package='igraphdata')
ggm<-calcCentrality(karate)
V(ggm)$DEG

Description

Function to calculate a distance matrix between a list of permuted vertex centrality matrices and a unperturbed reference matrix.

Usage

calcCentralityExternalDistances(m, l, keepOrder = FALSE, dist = "euclidean")

Arguments

Value

matrix with seven columns containing distances between each element of l and reference matrix m

See Also

getRandomGraphCentrality

getCentralityMatrix

calcCentralityInternalDistances

Examples

data(karate,package='igraphdata')
m<-getCentralityMatrix(karate)
gnp<-list()
for(i in 1:10){
    gnp[[i]]<-getRandomGraphCentrality(karate,type = 'gnp')
}
gnpEDist<-calcCentralityExternalDistances(m,gnp)
summary(gnpEDist)

Description

Function calculates a set of distance metrics between each vertex pair given a list of vertex centrality matrices

Usage

calcCentralityInternalDistances(l, keepOrder = FALSE, dist = "euclidean")

Arguments

Value

matrix with seven columns containing distances between all pairs of l elements.

See Also

getRandomGraphCentrality

getCentralityMatrix

calcCentralityExternalDistances

Examples

data(karate,package='igraphdata')
m<-getCentralityMatrix(karate)
gnp<-list()
for(i in 1:10){
    gnp[[i]]<-getRandomGraphCentrality(karate,type = 'gnp')
}
gnpIDist<-calcCentralityInternalDistances(gnp)
summary(gnpIDist)

Description

When applying resampling the clustering results of a clustering algorithm applied to a graph can differ due to the stochastic nature of the resampling algorithm. To allow reproducible downstream analysis clustering results are stored as vertex attributes in the graph. This function call getClustering and stores community membership as new vertex attribute in the graph, and Modularity as a new graph attribute prefix with the alg name.

Usage

calcClustering(gg, alg, weights = NULL)

Arguments

Details

NOTE: getClustering verifies algorithm names with match.arg so correct membership will be calculated, but name of the attribute is taken from alg argument, so it is possible that vertex attribute name won't exactly match name of the algorithm from link{getClustering}.

Value

modified igraph object with calculated membership stored as a vertex attribute and modularity as a graph attribute

See Also

getClustering

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
g<-calcClustering(karate, 'louvain')
vertex_attr_names(g)
graph_attr(g, 'louvain')

Description

Calculate each disease-disease pair overlap (or separation) on a given PPI network model, based on analysis described in Menche et al. 2015

Usage

calcDiseasePairs(
  gg,
  name,
  diseases = NULL,
  permute = c("none", "random", "binned")
)

Arguments

Value

list with three matrices:

• disease_separation – Ndisease X Ndisease matrix of separations

• gene_disease_separation – Ngenes X Ndisease+2 matrix of gene-disease separation

• disease_localisation – matrix with diseases in rows and number of genes (N), average and standard deviation of gene-disease separation in columns

References

Menche, J. et al. Uncovering disease-disease relationships through the incomplete interactome. Science, 347, (6224):1257601 (2015).

See Also

degreeBinnedGDAs

sampleDegBinnedGDA

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
agg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(agg)$name == '80273')
paste(V(agg)$GeneName[idx], 'GRPEL1')
p <- calcDiseasePairs(
agg,
name = "TopOntoOVGHDOID",
diseases = c("DOID:10652", "DOID:3312", "DOID:12849"),
permute = "n"
)
p$disease_separation

Description

This function calculate the graph entropy for each perturbed vertex by calling getEntropy, and save the results as new vertex attributes SR_UP and SR_DOWN in the graph.

Usage

calcEntropy(gg, maxSr = NULL, exVal = NULL)

Arguments

Details

According to Teschendorf et al., 2010, network entropy measure quantifies the degree of randomness in the local pattern information flux around single genes. For instance, in metastatic cancer this measure was found significantly higher than in non-metastatic and helped to identify genes and entire pathways involved on metastasis. However, for the assessment of scale-free structure we do not actually require gene expression data as it based solely on the network topology.

Value

graph with SR_UP and SR_DOWN vertex attributes storing the graph entropy values with over- or under-expressing each vertex.

See Also

getEntropy()

Other Entropy Functions: getEntropy(), getEntropyRate(), plotEntropy()

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
gg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(gg)$name == '80273')
paste(V(gg)$GeneName[idx], 'GRPEL1')
gg<- calcEntropy(gg)

Description

Calculates the clustering membership for each of the 10 clustering algorithms defined in function getClustering

Usage

calcMembership(
  gg,
  alg = c("lec", "wt", "fc", "infomap", "louvain", "sgG1", "sgG2", "sgG5", "spectral"),
  weights = NULL
)

Arguments

Value

data.frame with columns names and membership

See Also

getClustering

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
m<-calcMembership(karate, 'lec')
head(m)

Description

This function takes in a gg and initial vertex community membership values mem as returned by calcMembership, and then performs a reclustering of the graph given the clustering algorithm alg to those clusters of size greater than CnMAX

Usage

calcReclusterMatrix(
  gg,
  mem,
  alg,
  CnMAX = 10,
  weights = NULL,
  keepSplit = FALSE
)

Arguments

Value

remembership matrix, that contains vertex ID membership and result of reclustering

Examples

data(karate,package='igraphdata')
alg<-'louvain'
mem<-calcMembership(karate,alg = alg)
remem<-calcReclusterMatrix(karate,mem,alg,10)

Description

For a simple unweighted, undirected graph G(N,E). Network sparseness is defined as the ratio of the actual number of graph edges (E) to the maximum number of edges possible in a graph with same number of vertices (N): E/binom(N,2)

Usage

calcSparsness(gg)

Arguments

Value

sparsness value

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
calcSparsness(gg)

Description

Function to calculate basic summary statistics after apply clustering algorithm:

• N – number of vertices in the graph vcount

• mod – clustering modularity modularity, the ratio of edges found within communities to the number of edges found between communities, relative to a randomised model

• C – number of clusters

• Cn1 – number of singletones (clusters of size 1)

• Cn100 – number of clusters containing more than 100 nodes

• mu – the ratio of edges found within communities to the number of edges found between communities

• Min. C – minimum of the cluster size

• 1st Qu. C – first quartile of the cluster size

• Median C – median of the cluster size

• Mean C – average cluster size

• 3rd Qu. C – third quartile of the cluster size

• Max. C – maximum of the cluster size

Usage

clusteringSummary(
  gg,
  att = c("lec", "wt", "fc", "infomap", "louvain", "sgG1", "sgG2", "sgG5", "spectral")
)

Arguments

Value

matrix of clustering characteristics

Examples

data(karate,package='igraphdata')
g<-calcAllClustering(karate)
clusteringSummary(g)

Description

Calculate the cluster enrichment of a graph given a clustering algorithm alg and vertex annotation attribute 'name'. Function generates an enrichment table, one row for each cluster, containing: size of the cluster (Cn), number of annotated vertices in the graph $F_n$ (Fn), number of annotated vertices in the cluster $\mu$ (Mu), odds ratio (OR) and its 95% Confidence interval $[CI_l,CI_u]$ (CIl and CIu), two fold enrichment values $F_e$ (Fe) and $F_c$ (Fc). We also provide the list of vertices from the cluster that contribute to the annotation term, p.value of enrichment (pval) and depletion (palt) using the Hypergeometric test, adjusted p.values using Benjamini and Yekutieli correction (BY).

Usage

clusterORA(g, alg, name, vid = "name", alpha = 1, col = COLLAPSE)

Arguments

Details

Given the enrichment results, we can calculate the log of the Odds Ratio (OR) as:

$\ln(OR)=\ln(\frac{\mu(N-F_n+\mu-C_n)}{(C_n-\mu)(F_n-\mu)})$

and it’s upper and lower 95% Confidence Interval:

$CI(\ln(OR))=\ln(OR)\pm 1.96\sqrt{\frac{1}{\mu}+\frac{1}{C_n-\mu}+\frac{1}{F_n-\mu}+\frac{1}{N-F_n+\mu-C_n}}$

Using the odds ratio allows us to distinguish functionally enriched communities relative to functionally depleted communities.

Two types of fold enrichment values calculated as follow:

$F_e=\frac{(\frac{\mu}{F_n})}{(\frac{C_n}{N})}$

$F_c=\frac{(\frac{\mu}{C_n})}{(\frac{C_n}{N})}$

Value

A table with overrepresentation results. Each row corresponds to a tested annotation in particular cluster. The columns are the following:

• alg – name of the clustering algorithm;

• cl – cluster ID;

• FL – name of the enriched term;

• N – number vertices in the network;

• Fn – number of vertices in the graph annotated by term Fl ($F_n$);

• Cn – size of the cluster;

• Mu – number of vertices in the cluster annotated by term Fl ($\mu$);

• OR – odds ratio ;

• CIl – odds ratio 95% confidence interval lower bound ($CI_l$);

• CIu – odds ratio 95% confidence interval upper bound($CI_u$);

• Fe – fold enrichment $F_e$;

• Fc – fold enrichment $F_c$;

• pval – an enrichment p-value from hypergeometric test;

• padj – a BY-adjusted p-value;

• palt – an depletion p-value from hypergeometric test;

• paltadj – a BY-adjusted depletion p-value;

• overlapGenes – vector with overlapping genes.

Examples

options("show.error.messages"=TRUE)
file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
g <- igraph::read_graph(file, format="gml")
anL<-getAnnotationVertexList(g, 'TopOntoOVGHDOID')
res<-clusterORA(g, alg='louvain', name='TopOntoOVGHDOID', vid='name')
andf<-unique(data.frame(ID=vertex_attr(g, 'TopOntoOVGHDOID'),
Term=vertex_attr(g, 'TopOntoOVG')))
rr<-merge(andf, res, by.y='FL', by.x='ID')
rr[order(rr$cl), ]

Description

Function to randomly shuffle vertex annotation terms, whilst preserving the vertex degree originally found with that annotation term.

Usage

degreeBinnedGDAs(gg, GDA, dtype)

Arguments

Value

mapping matrix between vertices, vertex-degree groups and annotation terms.

See Also

prepareGDA

getAnnotationList

sampleDegBinnedGDA

Examples

options("show.error.messages"=TRUE)
file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
agg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(agg)$name == '80273')
paste(V(agg)$GeneName[idx], 'GRPEL1')
gda<-prepareGDA(agg, 'TopOntoOVGHDOID')
m<-degreeBinnedGDAs(agg, gda, getAnnotationList(gda))
c(dim(m), vcount(agg), length(getAnnotationList(gda)))
head(m)

Description

In the paper Goh.t al. (2007) doi:10.1073/pnas.0701361104 Barabasi with colleagues published Diseasome: a network of disorders and disease genes linked by known disorder–gene associations. We extract definition of the genes, disorders and interactions from papers supplementary materials and store it as graph object.

Usage

diseasome

Format

A bipartite graph as graph object.

Vertex attributes: ‘name’ for the node ID, ‘Name’ for the human readable node name, ‘Disorder.class’, ‘Type’ for the human readable node type, ‘label’ and ‘shape’ for plotting the graph, ‘type’ the node type for bipartite graph representation.

Details

Diseasesome is a bipartite graph that have nodes of two types gene and disease and links are allowed only between nodes of different types. It could be projected to Human Disease Network (HDN) and Disease Gene Network (DGN).

Source

Goh, K.-I. et al. The human disease network. Proc. Natl. Acad. Sci. U.S.A. 104, 8685–8690 (2007). https://pnas.org/doi/full/10.1073/pnas.0701361104

Description

In situations when a given list of annotation ID terms may not be well formatted, and therefore not be interoperated as unique. For example, given a list of HDO IDs: HDO:14, HDO:143, HDO:1433, and HDO:14330, a grep for the term HDO:14 could return: HDO:143, HDO:1433, HDO:14330. To avoid this all terms should be enclosed in escape characters, which unlikely to find within annotation itself.

Usage

escapeAnnotation(annVec, col = COLLAPSE, esc = ESC)

Arguments

Details

NOTE: spaces are treated as regular characters, no trimming is applied before or after escaping.

Value

vector of annotation strings with elements escaped

See Also

unescapeAnnotation

Examples

annVec<-apply(matrix(letters, ncol=13), 2, paste, collapse=';')
cbind(annVec, escapeAnnotation(annVec, ';', '|'))

Description

Function to compare two distance distributions using the Kolmogorov-Smirnov test. Where the first distance distribution is generated internally and calculates the distance between random graph centralities. The second distance distribution is generated externally, and measures the distance between random and the original graph centralities.

Usage

evalCentralitySignificance(dmi, dme)

Arguments

Value

list of lists for each centrality value in the input matrix three element list is created where ks contains Kolmogorov-Smirnov test result from class ks.test; pval contains Kolmogorov-Smirnov test pvalue; and dt contains input distribution.

See Also

ks.test

Examples

data(karate,package='igraphdata')
m<-getCentralityMatrix(karate)
gnp<-list()
for(i in 1:10){
    gnp[[i]]<-getRandomGraphCentrality(karate,type = 'gnp')
}
gnpIDist<-calcCentralityInternalDistances(gnp)
gnpEDist<-calcCentralityExternalDistances(m,gnp)

simSig<-evalCentralitySignificance(gnpIDist,gnpEDist)
sapply(simSig,function(.x).x$ks$p.value)

Description

Find Largest Connected Component of the graph

Usage

findLCC(GG)

Arguments

Value

igraph representation LCC

Examples

g1 <- make_star(10, mode="undirected") %du% make_ring(7) %du% make_ring(5)
lcc<-findLCC(g1)
summary(lcc)

Description

Fit a Powerlaw distribution to graph's degree distribution using the R “PoweRlaw” package (version 0.50.0) (Gillespie, 2015)

Usage

fitDegree(
  DEG,
  Nsim = 100,
  plot = FALSE,
  DATAleg = "Fit power-law",
  threads = 4,
  WIDTH = 480,
  HEIGHT = 480,
  legpos = "bottomleft",
  showErr = TRUE
)

Arguments

Value

an object of class law-class with results of fitting

Examples

##No: of bootstrap iterations use nsim > 100 for reliable result
nsim <- 10

##Legend Titles
Legend <- "Presynaptic PPI"

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
pFit <- fitDegree( as.vector(igraph::degree(graph=gg)),
DATAleg=Legend,threads=1, Nsim=nsim)

Description

This function calculates fit of the Fold-Enrichment distribution to the sigmoid function with the levels of noise specidied in SDV for all clustering algorithms, which have non-zero SUM3$`Psig&ORsig` in the enrichment table summary results. The function returns the list in which each element contains result for one of the noise level.

Usage

fitSigmoid(stat, SDv = c(0, 0.05, 0.1, 0.5))

Arguments

Details

Results are repersented as a list with five elements:

• gridplot that allow comparison of fitting for different clustering algorithms;

• plots the list of individual plots from gridplot;

• fitInfo the data.frame that contains results of fitting, such as message, number of iterations and exit code;

• parInfo values and standard deviations for all sigmoid parameters;

• ks table of Kolmogorov-Smirnov test p-values.

Grid plot is designed in a way to be viewed in the device at least 12 inches in width and 12 inches in height.

Value

list of fitted functions tables and plots

See Also

summaryStats()

Description

Annotation derived from Human Disease Ontology database (HDO). The table contains three columns; the first containing gene Entrez IDs, the second gene Human Disease Ontology (HDO) ID terms, the third gene HDO description terms; in csv format

See Also

annotateTopOntoOVG

Description

Annotation, downloaded from Gene Ontology for Biological Proceess domain. The table has columns: the first containing gene gene functional group ID terms, the second gene functional group description terms, the third - Human gene Entrez IDs; in csv format

See Also

annotateGoBP

Description

Annotation, downloaded from Gene Ontology for Cellular Compartment domain. The table has columns: the first containing gene gene functional group ID terms, the second gene functional group description terms, the third - Human gene Entrez IDs; in csv format

See Also

annotateGoCC

Description

Annotation, downloaded from Gene Ontology for Molecular Function domain. The table has columns: the first containing gene gene functional group ID terms, the second gene functional group description terms, the third - Human gene Entrez IDs; in csv format

See Also

annotateGoMF

Description

It is not uncommon to find both duplicated vertex annotation terms, and vertices annotated with multiple terms, in a given annotation list. This function creates a vector of unique annotation terms for each vertex given an input annotation list.

Usage

getAnnotationList(
  annVec,
  col = COLLAPSE,
  sort = c("none", "string", "frequency")
)

Arguments

Value

vector of unique annotation terms

See Also

getAnnotationVertexList

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
annVec<-V(gg)$TopOntoOVG
al<-getAnnotationList(annVec)
al

Description

For different purposes annotation of graph vertices could be represented in three forms:

Pairs

dataframe with vertex ID and annotation terms

Vertex Annotation

list named with vertex ID and containing terms annotating each vertex

Annotation Vertices

list named with term and containing vertex IDs

Usage

getAnnotationVertexList(g, name, vid = "name", col = COLLAPSE)

Arguments

Details

This function takes Vertex Annotation from vertex attribute and convert it to Annotation Vertices form.

Value

named list with annotation in Annotation Vertices form

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
avl<-getAnnotationVertexList(gg, 'TopOntoOVGHDOID')
head(avl)

Description

Bridginess takes into account a vertices shared community membership together with its local neighbourhood. It was proposed in Nepusz et al., 2008 doi:10.1103/PhysRevE.77.016107.

Usage

getBridgeness(gg, alg, conmat)

Arguments

Details

Function assumes clustering already been performed by the clustering algorithm, and its membership values stored in vertex attributes. If clustering algorithm vertex alg attribute is not found an error will be issued.

Value

data.frame with first column contains vertex ID, if GeneName attribute assigned to the vertices its value will be stored as a second column, the last column contains bridginess values for the

Examples

library(BioNAR)
karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
gg <- calcClustering(karate, 'louvain')
cnmat <- makeConsensusMatrix(gg, N=10, alg = 'louvain', type = 2, mask = 10)
br<-getBridgeness(gg, alg = 'louvain', cnmat)

Description

Calculate centrality measures for graph nodes.

Usage

getCentralityMatrix(gg, weights = NULL)

Arguments

Details

The edge attribute weights treated differently by different functions calculating centrality measures. For example, betweenness use weights as an edge length, while in page_rank "an edge with a larger weight is more likely to be selected by the surfer", which infer the opposite meaning. Taking into account that all methods in getClustering treat edge weights in the same way as page_rank, we calculate the distance=1/weights as edge weights for BET, dBET, mnSP, and sdSP values. So we treat weights in the package consistently as the strength and closiness of vertices, rather the distance between them.

Value

data.frame with following columns:

• ID - vertex ID

• DEG - degree

• iDEG - in-degree (directed graph only)

• oDEG - out-degree (directed graph only)

• BET - betweenness for undirected graph

• dBET - betweenness when directionality is taken into account (directed graph only)

• CC - clustering coefficient

• SL - semilocal centrality

• mnSP - mean shortest path

• PR - page rank for undirected graph

• dPR - page rank when directionality is taken into account (directed graph only)

• sdSP - standard deviation of the shortest path

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
m<-getCentralityMatrix(gg)

Description

Wrapper function for calculation of clustering for predefined set of ten algorithms:

• lec – leading eigenvector community (version of cluster_leading_eigen), directed graph will be converted to undirected by as_undirected with mode collapse;

• wt – walktrap community cluster_walktrap;

• fc – fastgreedy community cluster_fast_greedy, directed graph will be converted to undirected by as_undirected with mode collapse;

• infomap – infomap community cluster_infomap;

• louvain – cluster_louvain cluster_louvain, directed graph will be converted to undirected by as_undirected with mode collapse;

• sgG1 – spin-glass model and simulated annealing clustering (version of cluster_spinglass with spins=500 and gamma=1);

• sgG2 – spin-glass model and simulated annealing clustering (version of cluster_spinglass with spins=500 and gamma=2);

• sgG5 – spin-glass model and simulated annealing clustering (version of cluster_spinglass with spins=500 and gamma=7);

• spectral – spectral modularity clustering spectral_igraph_communities;

Usage

getClustering(
  gg,
  alg = c("lec", "wt", "fc", "infomap", "louvain", "sgG1", "sgG2", "sgG5", "spectral"),
  weights = NULL
)

Arguments

Details

graph suppose to be undirected. If algorithm failed warning will be issued and function returned NULL.

Algorithm names are verified with match.arg.

Value

communities object or NULL if algorithm failed.

Examples

data(karate,package='igraphdata')
c<-getClustering(karate,'lec')
c$modularity

Description

Function reads in a graph gg, vertex cluster membership vector mem, and returns an induced subgraph given a cluster membership number 'clID'.

Usage

getClusterSubgraphByID(clID, gg, mem)

Arguments

Value

induced subgraph as igraph object

Examples

data(karate,package='igraphdata')
alg<-'louvain'
c<-getClustering(karate,alg = alg)
gc3<-getClusterSubgraphByID(3,karate,membership(c))
#plot(gc3,vertex.label=V(gc3)$name)

Description

The idea based upon this StackOverflow answer

Usage

getCommunityGraph(gg, membership)

Arguments

Value

community graph

Examples

data(karate,package='igraphdata')
alg<-'louvain'
mem<-calcMembership(karate,alg = alg)
cg<-getCommunityGraph(karate,mem$membership)

Description

Return vector of HDO disease IDs for synaptic PPI analysis.

Usage

getDiseases()

Value

vector of disease IDs of interest

See Also

getDType

Examples

getDiseases()

Description

Return vector of disease abbreviations for synaptic PPI analysis.

Usage

getDType()

Value

vector of disease abbreviations for synaptic PPI analysis.

See Also

getDiseases

Examples

getDType()

Description

This function calculates sensitivity matrix that represents perturbation patterns defined by topology and edge weights of the network. If weights are signed value sensitivity matrix is able to reproduce not only activation but inhibition relationships in the network.

Usage

getDYNAMO(g, attr = NULL, vid = "name", alpha = 0.9)

Arguments

Details

Algorithm proposed in:

Santolini,M. and Barabasi,A.-L. (2018) Predicting perturbation patterns from the topology of biological networks. Proc Natl Acad Sci USA, 169, 201720589.

Value

sparce sensitivity matrix defined by the network topology and edge values

Examples

data(karate, package='igraphdata')
upgrade_graph(karate)
d<-getDYNAMO(karate,attr='weight')
df<-metlMatrix(d)
head(df)

Description

According to Teschendorf et al., 2010, network entropy measure quantifies the degree of randomness in the local pattern information flux around single genes. For instance, in metastatic cancer this measure was found significantly higher than in non-metastatic and helped to identify genes and entire pathways involved on metastasis. However, for the assessment of scale-free structure we do not actually require gene expression data as it based solely on the network topology.

Usage

getEntropy(gg, maxSr = NULL, exVal = NULL)

Arguments

Details

In this function, following procedure described in (Teschendorff et al., 2015), all vertexes are artificially assigned a uniform weight then sequentially perturbed with the global entropy rate (SR) after each protein’s perturbation being calculated and plotted against the log of the protein’s degree. In case of scale-free or approximate scale-free topologies, we see a clear bi-modal response between over-weighted vertices and their degree and an opposing bi-phasic response in under-weighted vertices and their degrees.

Value

matrix containing for each Gene:

• Entrez ID,

• Name,

• Degree,

• UP – Graph Entropy values when gene is expressed up,

• DOWN – Graph Entropy values when gene is expressed down.

Note

Entropy is calculated with respect to GeneName property, if there is no such vertex attribute in the graph vertex name will be copied to the GeneName attribute. If any NA is found in GeneNames error will be thrown.

See Also

Other Entropy Functions: calcEntropy(), getEntropyRate(), plotEntropy()

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
gg<-annotateGeneNames(gg)
any(is.na(V(gg)$GeneName))
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(gg)$name == '80273')
paste(V(gg)$GeneName[idx], 'GRPEL1')
e<- getEntropy(gg)

Description

This function calculates the maximum entropy rate $maxSR$ (maxSr) and initial entropy rate $SR_0$ (SRo) given a connected network.

Usage

getEntropyRate(gg)

Arguments

Details

The maximum entropy rate being calculated from the network’s adjacency matrix:

$maxSR = \sum_{i,j} p_{ij} = \frac{A_{ij}\nu_j}{\lambda\nu_i}$

where $\nu$ and $\lambda$ are the leading eigenvector and eigenvalue of the network adjacency matrix $A$ respectively.

The initial configuration occurs when the entropy for each node is maximal. This can be calculated by setting the expression value for each gene/node in the network to be the same, and thus the maximal node entropy is dependent only on the node’s degree k:

$SR_0 = \frac{1}{N\bar{k}} \sum_j k_j \log k_i$

where N here is the number of nodes and $\bar{k}$ the average node degree found in the network.

Value

list with values of maxSr and SRo

See Also

Other Entropy Functions: calcEntropy(), getEntropy(), plotEntropy()

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
ent <- getEntropyRate(karate)

Description

Function generates random G(n,p) Erdos-Renyi graph (sample_gnp) with the same number of vertices and edges as in in the reference graph gg.

Usage

getGNP(gg, ...)

Arguments

Value

new instance of the random graph.

Examples

data(karate,package='igraphdata')
vcount(karate)
ecount(karate)
rg<- getGNP(karate)
vcount(rg)
ecount(rg)

Description

Convert centrality matrix into ECDF

Usage

getGraphCentralityECDF(m)

Arguments

Value

list of several ecdf objects, corresponding to values in centrality matrix from getCentralityMatrix invocation.

See Also

getCentralityMatrix

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
m<-getCentralityMatrix(gg)
ecdfL<-getGraphCentralityECDF(m)

Description

Utility function to get vertex ids from vertex attributes The function obtain attribute values and check duplicates in it. It fails if any duplicate found.

Usage

getIDs(gg, idatt)

Arguments

Value

idatt attribute values

Description

The function generates random Barabasi-Albert graph (sample_pa) with the same vertex number as in the reference graph gg and the power specified by parameter pwr. If pwr is missing, we are trying to estimate pwr from the reference graph gg.

Usage

getPA(gg, pwr, ...)

Arguments

Value

new instance of the random graph.

Examples

data(karate,package='igraphdata')
vcount(karate)
ecount(karate)
rg<- getPA(karate,pwr=1.25)
vcount(rg)
ecount(rg)

Description

Generate a random graph that mimics the properties of the input graph and calls getCentralityMatrix to calculate all available vertex centrality measures. There are four different types of random graph to generate

Usage

getRandomGraphCentrality(
  gg,
  type = c("gnp", "pa", "cgnp", "rw"),
  power = NULL,
  weights = NULL,
  ...
)

Arguments

Value

matrix of random graph vertices centrality measure.

See Also

getCentralityMatrix() for explanation of the use of weights.

Examples

data(karate,package='igraphdata')
m<-getRandomGraphCentrality(karate,'pa',threads=1)
# to avoid repetitive costy computation of PowerLaw fit
# power parameter could be send explicitly:
pFit <- fitDegree( as.vector(igraph::degree(graph=karate)),
Nsim=10, plot=FALSE,threads=1)
pwr <- slot(pFit,'alpha')
m<-getRandomGraphCentrality(karate,'pa',power=pwr)
lpa<-lapply(1:5,getRandomGraphCentrality,gg=karate,type='pa',
            power=pwr,weights = NULL)

Description

This function takes as argument a network (gg), the name of a clustering algorithm (alg) which can be found in the network, and a consensus matrix (conmat) generated from the clustering network. The function uses the consensus matrix to generate a measure of cluster robustness $C_{rob}$ (Crob) for each cluster (C) using the R function clrob. Briefly, this is done by summing elements of the consensus matrix that are found in the same cluster, and dividing this by the total number of entries in the matrix:

$C_{rob}=\frac{2}{C_n(C_n-1)} \sum_{i,j \in I_C \atop i \le j} conmat_{i,j}$

where $I_C$ – indices of vertices of the cluster $C$, $C_n$ is the number of nodes found inside the cluster $C$.

Usage

getRobustness(gg, alg, conmat)

Arguments

Value

data.frame that for each cluster C shows

• its size $C_n$ (Cn),

• robustness $C_{rob}$ (Crob) and

• robustness scaled to range between 0 and 1 (CrobScaled).

See Also

Other Robustness functions: makeConsensusMatrix()

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
alg<-'louvain'
gg<-calcClustering(karate, alg = alg)
conmat<-makeConsensusMatrix(gg, N=100, mask = 10, alg = alg, type = 2)
clrob<-getRobustness(gg, alg = alg, conmat)
clrob

Description

This is internal function and do not suppose to be called by user.

Usage

gofs(x, rate, model, sigma2 = NULL, countDATA = TRUE)

Arguments

Value

list of ks.test values for each value in rate

Description

Result of PawerLaw fit

Slots

fit

displ-class result of power law fit.

p

numeric.

alpha

numeric degree of power-law.

SDxmin

numeric bootstrap sd of Xmin.

SDalpha

numeric bootstrap sd of alpha.

Description

Function to split graph into clusters and layout each cluster independently..

Usage

layoutByCluster(gg, mem, layout = layout_with_kk)

Arguments

Value

Layout in a form of 2D matrix.

See Also

igraph::layout_

Examples

data(karate,package='igraphdata')
alg<-'louvain'
mem<-calcMembership(karate,alg = alg)
lay<-layoutByCluster(karate,mem)
#plot(karate,layout=lay)

Description

Takes results of recluster and apply layoutByCluster to each

Usage

layoutByRecluster(gg, remem, layout = layout_with_kk)

Arguments

Value

Layout in a form of 2D matrx.

Examples

data(karate,package='igraphdata')
alg<-'louvain'
mem<-calcMembership(karate,alg = alg)
remem<-calcReclusterMatrix(karate,mem,alg,10)
lay<-layoutByRecluster(karate,remem)
#plot(karate,layout=lay)

Description

Function to make random resampling consensus matrix in memory

Usage

makeConsensusMatrix(
  gg,
  N = 500,
  mask = 20,
  alg,
  type,
  weights = NULL,
  reclust = FALSE,
  Cnmax = 10
)

Arguments

Details

Function to assess the robustness of network clustering. A randomisation study is performed apply the same clustering algorithm to N perturbed networks, and which returns the consensus matrix where each vertex pair is assigned the probability of belong to the same cluster. The inputted network is perturbed by randomly removing a mask percentage of edges (type=1) or vertices (type=2) from the network before clustering.

Value

consensus matrix of Nvert X Nvert

See Also

Other Robustness functions: getRobustness()

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
alg<-'louvain'
gg<-calcClustering(karate, alg = alg)
conmat<-makeConsensusMatrix(gg, N=100, mask = 10, alg = alg, type = 2)
dim(conmat)

Description

Create membership data.frame from graph vertex attribute or vector of cluster names, IDs or indices. This function is simular to calcMembership but do not linked to clustering algorithm.

Usage

makeMembership(gg, membership)

Arguments

Details

Any annotation coercible to factor could be converted to the membership data.frame. This function is useful, for example, to make layout with layoutByCluster.

Value

data.frame with two columns names and membership

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
m<-makeMembership(karate,rep(c(1,2),length.out=vcount(karate)))
head(m)

Description

Algorithm proposed in:

Usage

markBowTie(g)

Arguments

Details

"Bow-tie Decomposition in Directed Graphs" - Yang et al. IEEE (2011)

Value

graph with BowTie vertex attribute

Description

For very large graphs handling adjacency-like matrices is difficult due to its sparse nature. This function convert sparse matrix into triplet data.frame with row and column indices and names, and cell value.

Usage

metlMatrix(sparceM)

Arguments

Value

data.frame with three colums:

• i – row index;

• j – column index;

• x – cell value;

• Rname – i-th row name;

• Cname – j-th column name.

Examples

data(karate, package='igraphdata')
upgrade_graph(karate)
Ws <- as_adjacency_matrix(karate,type='both',attr='weight',sparse = TRUE)
mdf<-metlMatrix(Ws)
head(mdf)

Description

Function to compare network Modularity of input network with networks of different size and connectivity.

Usage

normModularity(
  gg,
  alg = c("lec", "wt", "fc", "infomap", "louvain", "sgG1", "sgG2", "sgG5"),
  Nint = 1000,
  weights = NULL
)

Arguments

Details

Used the normalised network modularity value Qm based on the previous studies by Parter et al., 2007, Takemoto, 2012, Takemoto, 2013, Takemoto and Borjigin, 2011, which was defined as:

$Q_m = \frac{Q_{real}-Q_{rand}}{Q_{max}-Q_{rand}}$

Where $Q_{real}$ is the network modularity of a real-world signalling network and, $Q_{rand}$ is the average network modularity value obtained from 10,000 randomised networks constructed from its real-world network. $Q_{max}$ was estimated as: 1 - 1/M, where M is the number of modules in the real network.

Randomised networks were generated from a real-world network using the edge-rewiring algorithm (Maslov and Sneppen, 2002).

Value

normalized modularity value

References

Takemoto, K. & Kihara, K. Modular organization of cancer signaling networks is associated with patient survivability. Biosystems 113, 149–154 (2013).

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)

nm<-normModularity(gg, alg='louvain',Nint=10)

Description

This function is a convinience wrapper to sample with replace= FALSE

Usage

permute(GNS, N)

Arguments

Value

random list of GNS values

Examples

permute(LETTERS, 15)

Description

Semi-local centrality measure (Chen et al., 2011) lies between 0 and 1 indicating whether protein is important globally or locally. By plotting Bridgeness against semi-local centrality we can categorises the influence each protein found in our network has on the overall network structure:

• Region 1, proteins having a 'global' rather than 'local' influence in the network (also been called bottle-neck bridges, connector or kinless hubs (0<Sl<0.5; 0.5<Br<1).

• Region 2, proteins having 'global' and 'local' influence (0.5<Sl<1, 0.5<Br<1).

• Region 3, proteins centred within the community they belong to, but also communicating with a few other specific communities (0<Sl<0.5; 0.1<Br<0.5).

• Region 4, proteins with 'local' impact , primarily within one or two communities (local or party hubs, 0.5<Sl<1, 0<Br<0.5).

Usage

plotBridgeness(
  gg,
  alg,
  VIPs,
  Xatt = "SL",
  Xlab = "Semilocal Centrality (SL)",
  Ylab = "Bridgeness (B)",
  bsize = 3,
  spsize = 7,
  MainDivSize = 0.8,
  xmin = 0,
  xmax = 1,
  ymin = 0,
  ymax = 1,
  baseColor = "royalblue2",
  SPColor = "royalblue2"
)

Arguments

Value

ggplot object with plot

Examples

karate <- make_graph("Zachary")
# We need vertex ID in the 'name' attribute of the vertex
V(karate)$name<-c(LETTERS,letters)[1:vcount(karate)]
set.seed(100)
gg <- calcClustering(karate, 'louvain')
gg <- calcCentrality(gg)
cnmat <- makeConsensusMatrix(gg, N=10, alg = 'louvain', type = 2, mask = 10)
gg<-calcBridgeness(gg, alg = 'louvain', cnmat)
plotBridgeness(gg,alg = 'louvain',VIPs=c("Mr Hi","John A"))

Description

Following procedure described in (Teschendorff et al., 2015), all vertexes are artificially assigned a uniform weight then sequentially perturbed with the global entropy rate (SRprime) after each protein’s perturbation being calculated by getEntropy function.

Usage

plotEntropy(SRprime, subTIT = "Entropy", SRo = NULL, maxSr = NULL)

Arguments

Details

This function plot SRprime against the log of the protein’s degree. In case of scale-free or approximate scale-free topologies, we see a clear bi-modal response between over-weighted vertices and their degree and an opposing bi-phasic response in under-weighted vertices and their degrees.

If maxSr or SRo is set to their default value NULL getEntropyRate will be called and returned values will be used in the following calculations. As maxSr is required for SRprime calculation by getEntropy using explicit values could save some time in the case of large network.

Value

ggplot2 object with diagram

See Also

getEntropy()

Other Entropy Functions: calcEntropy(), getEntropy(), getEntropyRate()

Examples

file <- system.file("extdata", "PPI_Presynaptic.csv", package = "BioNAR")
tbl <- read.csv(file, sep="\t")
gg <- buildNetwork(tbl)
gg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(gg)$name == '80273')
paste(V(gg)$GeneName[idx], 'GRPEL1')
ent <- getEntropyRate(gg)
SRprime <- getEntropy(gg, maxSr = NULL)
plotEntropy(SRprime, subTIT = "Entropy", SRo = ent$SRo, maxSr = ent$maxSr)

Description

Plot fraction of enriched communities

Usage

plotRatio(
  x,
  desc = "",
  anno = "",
  LEGtextSize = 1.5,
  LEGlineSize = 4,
  type = NULL
)

Arguments

Value

ggplot object

Description

Plot results of the sigmoid fit

Usage

plotSigmoid(x, rates, model, alg = "", pv = 0)

Arguments

Value

ggplot object with sigmoid fit plot

Description

Protein-protein interactions (PPIS) for presynaptic compartment, extracted from Synaptome.db, in a csv form. Columns A and B correspond to Entrez IDs for interacting proteins A and B (node names); column We contains the edge weights, if available.

See Also

buildNetwork

Description

Protein-protein interactions (PPIS) for presynaptic compartment, extracted from Synaptome.db, and saved in a graph format. Graph contains node attributes, such as names (Entrez IDs), Gene Names, disease association (TopOntoOVG, TopOntoOVGHDOID), annotation with schizophrenia-related genes (Schanno (v/c), function annotation from GO (GOBPID, GOBP, GOMFID, GOMF, GOCCID, GOCC), centrality measures (DEG - degree, BET - betweenness, CC - clustering coefficient, SL - semilocal centrality, mnSP - mean shortest path, PR - page rank, sdSP - standard deviation of the shortest path), and clustering memberships for 8 clustering algorithms (lec, wt, fc, infomap, louvain, sgG1, sgG2, sgG5)

Description

Function to return vertex annotation from a graph in the Vertex Annotation form and format it for further analysis.

Usage

prepareGDA(gg, name)

Arguments

Value

escaped annotation in Vertex Annotation form

See Also

getAnnotationVertexList

escapeAnnotation

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
agg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(agg)$name == '80273')
paste(V(agg)$GeneName[idx], 'GRPEL1')
gda<-prepareGDA(agg, 'TopOntoOVGHDOID')
gda<-prepareGDA(agg, 'TopOntoOVGHDOID')
head(gda)

Description

Presynaptic genes functional annotation derived from Boyken at al. (2013) doi:10.1016/j.neuron.2013.02.027. The table has columns: the first containing functional group ID terms, the second - gene functional group description terms, third - gene Human Entrez Ids; in csv format

See Also

annotatePresynaptic

Description

Function reads in a graph GG with cluster membership stored in vertex attribute ALGN, and reapplies the clustering algorithm ALGN to all clusters larger than CnMAX

Usage

recluster(GG, ALGN, CnMAX, weights = NULL)

Arguments

Value

remembership matrix, that contains vertex ID membership and result of reclustering

Examples

data(karate,package='igraphdata')
alg<-'louvain'
mem<-calcMembership(karate,alg = alg)
remem<-calcReclusterMatrix(karate,mem,alg,10)

Description

Remove vertex property.

Usage

removeVertexTerm(GG, NAME)

Arguments

Value

igraph object with attribute removed

Examples

data(karate, package='igraphdata')
upgrade_graph(karate)
vertex_attr_names(karate)
m<-removeVertexTerm(karate, 'color')
vertex_attr_names(m)

Description

Function to calculate the disease-pair overlap characteristics of an inputted network, before applying Nperm permutations on the disease annotations of #' type "random" or "binned" permute. From the permuted networks the function estimates the significance of disease overlap: p-value, Bonferoni-adjusted p-value, and q-value in the Disease_overlap_sig. The function also compares the average disease separation between inputted and permuted networks, and calculates its significance using the Wilcox test and store. Significance of disease-pair overlap and disease separation results are stored in the matrix Disease_location_sig.

Usage

runPermDisease(
  gg,
  name,
  diseases = NULL,
  Nperm = 100,
  permute = c("random", "binned"),
  alpha = c(0.05, 0.01, 0.001)
)

Arguments

Details

Run with care, as large number of permutations could require a lot of memory and be timeconsuming.

Value

list of two matrices: Disease_overlap_sig gives s tatistics for each pair of disease, and Disease_location_sig gives intra-disease statistics

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
agg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(agg)$name == '80273')
paste(V(agg)$GeneName[idx], 'GRPEL1')
r <- runPermDisease(
agg,
name = "TopOntoOVGHDOID",
diseases = c("DOID:10652", "DOID:3312", "DOID:12849", "DOID:1826"),
Nperm = 10,
alpha = c(0.05, 0.01, 0.001))
r$Disease_location_sig

Description

Function to randomly shuffle vertex annotation terms, whilst preserving the vertex degree originally found with that annotation term..

Usage

sampleDegBinnedGDA(org.map, term)

Arguments

Value

vertex IDs to assign term in shuffled annotation

See Also

degreeBinnedGDAs

Examples

file <- system.file("extdata", "PPI_Presynaptic.gml", package = "BioNAR")
gg <- igraph::read_graph(file, format="gml")
agg<-annotateGeneNames(gg)
# due to error in org.Hs.eg.db we have to manually check annotation of one node
idx <- which(V(agg)$name == '80273')
paste(V(agg)$GeneName[idx], 'GRPEL1')
gda<-prepareGDA(agg, 'TopOntoOVGHDOID')
diseases<-getAnnotationList(gda)
m<-degreeBinnedGDAs(agg, gda, diseases)
sampleDegBinnedGDA(m, diseases[1])

Description

Function will mask mask a percentage of edges (type=1) or vertices (type=2) from the network, find the largest connected component of the masked network and cluster it. The clustering results are stored in a three column matrix: the first column contains the vertex IDs of input network; the second column the vertex IDs of the subsampled network, or -1 if the vertex has been masked; the third column the cluster membership of subsampled network, or -1 if vertex has been masked.

Usage

sampleGraphClust(
  gg,
  mask = 20,
  alg,
  type,
  weights = NULL,
  reclust = FALSE,
  Cnmax = 10
)

Arguments

Details

This is internal function and not supposed to be calle by end user.

Value

list of Nx3 matrices

Examples

data(karate,package='igraphdata')
alg<-'louvain'
mem<-calcMembership(karate,alg = alg)
smpl<-BioNAR:::sampleGraphClust(karate,mask=10,alg,type=2)

Description

Annotation, manually curated from an external file: Lips et al., (2012) doi:10.1038/mp.2011.117.The table has columns: the first containing gene Human Entrez IDs, the second gene functional group ID terms, the third gene functional group description terms; in csv format

See Also

annotateSCHanno

Description

Calculate summary statistics from enrichment table

Usage

summaryStats(RES, ALPHA, usePadj = FALSE, FeMAX = 0, FcMAX = 0)

Arguments

Value

list of data.frame

Description

Function to remove all escape characters from annotation strings (opposite to escapeAnnotation).

Usage

unescapeAnnotation(annVec, col = COLLAPSE, esc = ESC)

Arguments

Details

NOTE: spaces are treated as regular characters, no trimming is applied before or after escaping.

Value

vector of annotation strings with removed escape characters

See Also

escapeAnnotation

Examples

annVec<-apply(matrix(letters, ncol=13), 2, paste, collapse=';')
escVec<-escapeAnnotation(annVec, ';', '|')
cbind(annVec, escVec, unescapeAnnotation(escVec, ';', '|'))

Description

Auxiliary function to replace NAs with zeros.

Usage

zeroNA(x)

Arguments

Value

matrix or vector with NAs replaced by zero.

Examples

x<-matrix(NA,nrow = 3,ncol = 3)
zeroNA(x)