Package 'BulkSignalR'

Description

We note discrepancy between format available over internet.

Usage

.formatPathwaysFromGmt(file, resourceName = NULL)

Arguments

Details

Here we consider a valid gmt file format defined on each lines as follows : First is Pathway name, Then comes the ID, Finally you will find genes symbols according to the pathway defined on the line.

You can find an example here. - For Reactome. (Directly from their website) https://reactome.org/download/current/ReactomePathways.gmt.zip Note that you need to unzip the file to read the content. The code is inspired from read.gmt function from the gsa R package.

Value

Dataframe with pathwayID, geneName and pathwayName

Description

Format dataframe according to json input

Usage

.formatPathwaysFromJson(file, resourceName = NULL)

Arguments

Value

Dataframe with pathwayID, geneName and pathwayName

Description

Read dataframe from txt file

Usage

.formatPathwaysFromTxt(file, resourceName = NULL)

Arguments

Value

Dataframe with pathwayID, geneName and pathwayName

Description

Add a comparison to a BSRDataModelComp object.

Usage

## S4 method for signature 'BSRDataModelComp'
addClusterComp(obj, cmp, cmp.name)

Arguments

Details

Add cmp to the list of comparisons contained in obj.

Value

A BSRDataModelComp object.

Examples

# prepare data
data(sdc, package = "BulkSignalR")
normal <- grep("^N", names(sdc))
bsrdm <- BSRDataModel(sdc[, -normal])

# define the comparison
bsrdm.comp <- as(bsrdm, "BSRDataModelComp")
colA <- as.integer(1:3)
colB <- as.integer(12:15)
n <- nrow(ncounts(bsrdm.comp))
stats <- data.frame(
    pval = runif(n), logFC = rnorm(n, 0, 2),
    expr = runif(n, 0, 10)
)
rownames(stats) <- rownames(ncounts(bsrdm.comp))
bsrcc <- BSRClusterComp(bsrdm.comp, colA, colB, stats)

bsrdm.comp <- addClusterComp(bsrdm.comp, bsrcc, "random.example")

Description

Representation of the links between Ligands,Receptors and Pathways.

Usage

alluvialPlot(bsrinf, keywords, type = c("L", "R", "pw.id"), qval.thres = 0.01)

Arguments

Value

NULL

This is a convenience function that relies on the ggalluvial package to propose a simple way of representing Ligands, Receptors

Examples

data(bsrinf, package = "BulkSignalR")
alluvialPlot(bsrinf,
    keywords = c("LAMC1"),
    type = "L",
    qval.thres = 0.01)

Description

Dataframe subset describing the spatial spots

Usage

data(annotation.spa)

Format

Dataframe that contains the following columns : barcode_id,sample_id, in_tissue,array_row array_col,ground_truth,reference,cell_count,idSpatial

barcode_id is the id of the spot idSpatial is the spatial id of the spot(array_rowXarray_col) ground_truth is the label (Layer1/2 were only kept)

They are the mandatory informations in order to make plots for the spatial workflow.

Source

http://spatial.libd.org/spatialLIBD/

Description

Generate a data.frame linking interactions to cell types.

Usage

assignCellTypesToInteractions(
  bsrdm,
  bsrinf,
  ct.scores,
  normalize.scores = TRUE,
  min.weight = 0.1,
  min.r2 = 0.25,
  min.r2.after = 0.35,
  lasso = TRUE,
  qval.thres = 0.001
)

Arguments

Value

A data.frame containing the cell type assignments for each L-R interaction. Unique interactions are considered only (thanks to "reduceToBestPathway" that is applied internally). An interaction can be associated with several cell types or none. In case it is associated with a single cell type, it is labelled autocrine (indicative only).

Cell type signature scores must be provided. They can be computed with BulkSignalR utility function "scoreSignatures", but also any other external tool such as CIBERSORT or BisqueRNA. In case such a tool would score cell types in a nonlinear fashion, we recommend to transform the score matrix to restore a linear relationship cell type abundance/score. By default, cell type (and L-R gene signature) scores are normalized between 0 and 1 to make the weights of each cel type in the linear models as comparable as possible.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")
data(tme.signatures, package = "BulkSignalR")

immune.signatures <- immune.signatures[immune.signatures$signature %in%
    c("T cells"), ]

signatures <- rbind(immune.signatures, tme.signatures[
    tme.signatures$signature %in% c("Fibroblasts"),
])

tme.scores <- scoreSignatures(bsrdm, signatures)

# assignment
lr2ct <- assignCellTypesToInteractions(bsrdm, bsrinf, tme.scores)

Description

A dataset containing rpkm values of brain and liver.

Usage

data(bodyMap.mouse)

Format

A data frame with 24543 rows and 8 variables.

Source

Bin Li & al., Scientific Reports, 2017;

Description

Define the columns of the expression matrix that belong to each cluster, and store the result of the cluster differences statistical analysis obtained by an external tool such as edgeR, DESeq2, etc.

Usage

BSRClusterComp(obj, col.clusterA, col.clusterB, differential.stats)

Arguments

Details

Create a BSRClusterComp object describing a comparison of two clusters of columns taken from the expression matrix in the BSRDataModelComp object obj. Such a cluster comparison description is the basis for inferring LRIs from differential expression P-values instead of correlation analysis.

The rows of differentialStats must be in the same order as those of the count matrix in obj. Alternatively, differentialStats rows can be named and a 1-1 correspondence must exist between these names and those of the count matrix.

Value

A BSRClusterComp object.

Examples

# prepare data
data(sdc, package = "BulkSignalR")
normal <- grep("^N", names(sdc))
bsrdm <- BSRDataModel(sdc[, -normal])

# define the comparison
bsrdm.comp <- as(bsrdm, "BSRDataModelComp")
colA <- as.integer(1:3)
colB <- as.integer(12:15)
n <- nrow(ncounts(bsrdm.comp))
stats <- data.frame(
    pval = runif(n), logFC = rnorm(n, 0, 2),
    expr = runif(n, 0, 10)
)
rownames(stats) <- rownames(ncounts(bsrdm.comp))
bsrcc <- BSRClusterComp(bsrdm.comp, colA, colB, stats)

Description

An S4 class to represent the comparison of two clusters of samples to infer LR interactions based on the resulting P-values, log-fold-changes (logFC), and expression values.

Slots

col.clusterA

Column indices for the samples in cluster A.

col.clusterB

Column indices for the samples in cluster B.

differential.stats

Comparison statistics A versus B as a data.frame and containing at least two columns named 'pval', 'logFC', and 'expr'.

Examples

new("BSRClusterComp")

Description

Take a matrix or data frame containing RNA sequencing, microarray, or expression proteomics data and return a BSRDataModel object ready for subsequent training. Normally, BSRDataModel objects are not instantiated directly, but through this function.

Usage

BSRDataModel(
  counts,
  normalize = TRUE,
  symbol.col = NULL,
  min.count = 10,
  prop = 0.1,
  method = c("UQ", "TC"),
  log.transformed = FALSE,
  min.LR.found = 80,
  species = "hsapiens",
  conversion.dict = NULL,
  UQ.pc = 0.75,
  x.col = NULL,
  y.col = NULL,
  barcodeID.col = NULL
)

Arguments

Details

The counts matrix or table should be provided with expression levels of protein coding genes in each samples (column) and rownames(counts) set to HUGO official gene symbols. For commodity, it is also possible to provide counts with the gene symbols stored in one of its columns. This column must be specified with symbol.col. In such a case, BSRDataModel will extract this column and use it to set the row names. Because row names must be unique, BSRDataModel will eliminate rows with duplicated gene symbols by keeping the rows with maximum average expression. Gene symbol duplication may occur in protein coding genes after genome alignment due to errors in genome feature annotation files (GTF/GFF), where a handful of deprecated gene annotations might remain, or some genes are not given their fully specific symbols. If your read count extraction pipeline does not take care of this phenomenon, the maximum mean expression selection strategy implemented here should solve this difficulty for the sake of inferring ligand-receptor interactions.

If normalize is TRUE then normalization is performed according to method. If those two simple methods are not satisfying, then it is possible to provide a pre-normalized matrix setting normalize to FALSE. In such a case, the parameter method must be used to document the name of the normalization algorithm used.

In case proteomic or microarray data are provided, min.count must be understood as its equivalent with respect to those data.

Value

A BSRModelData object with empty model parameters.

Examples

data(sdc, package = "BulkSignalR")
idx <- sample(nrow(sdc), 4000)
bsrdm <- BSRDataModel(sdc[idx, c("N22","SDC17")],
normalize = FALSE,method="UQ")

Description

An S4 class to represent the expression data used for inferring ligand-receptor interactions.

Slots

ncounts

Normalized read count matrix. Row names must be set to HUGO official gene symbols.

log.transformed

Logical indicating whether values in ncounts were log2-transformed.

normalization

Name of the normalization method.

param

List containing the statistical model parameters.

initial.organism

Organism for which the data were obtained.

initial.orthologs

List of genes for which human orthologs exist.

Examples

new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)

Description

An S4 class to represent the expression data used for inferring ligand-receptor interactions based on sample cluster comparisons.

Slots

comp

A named list of BSRClusterComp objects, one per comparison.

mu

A number representing the average value in the normalized and lop1p-transformed gene expression matrix. This value is used to compute the LR-score (cf. SingleCellSignalR paper, Cabello-Aguilar, et al., Nucleic Acids Res, 2020)

Examples

new("BSRDataModelComp")

Description

Output from the 'learnParameters' function to get BulkSignalR statistical model parameters.

Usage

data(bsrdm)

Format

An example of an object created by 'BSRDataModel' applied to an sdc subset (Patients N20,N22,SDC17,SDC25) and 10 000 genes sampled (seed set to 123) 'learnParameters' was also called to get statistical model parameters.

Description

See Vignette BulkSignalR-Differential.

Usage

data(bsrdm.comp)

Format

An example of an BSR-dataModelComp object

Description

obtained from STexampleData::Visium_humanDLPFC. A single sample (sample 151673) of human brain dorsolateral prefrontal cortex (DLPFC) in the human brain, measured using the 10x Genomics Visium platform. This is a subset of the full dataset published by Maynard and Collado-Torres et al. (2021). The subset is reproduced in the vignette. name.idx <- c("10x32","3x47","4x50", "17x111","5x59","0x20","8x100", "8x108","14x30","11x39")

Usage

data(bsrdm.spa)

Format

An example of an object created by 'BSRDataModel' applied to a subset of a spatial dataset. 'learnParameters' was also called to get statistical model parameters.

Details

Output from the 'learnParameters' function to get BulkSignalR statistical model parameters for a subset of a spatial dataset.

Source

http://spatial.libd.org/spatialLIBD/

Description

From the previous object 'bsrdm', you can generate inferences by calling its method 'BSRInference'. The resulting BSR-Inference object is 'bsrinf', It contains all the inferred L-R interactions with their associated pathways and corrected p-values.

Usage

data(bsrinf)

Format

An example of an object created by inference function

Description

See Vignette BulkSignalR-Differential.

Usage

data(bsrinf.comp)

Format

An example of an BSR-InferenceComp object

Description

see related workflow for non human organism in the vignette

Usage

data(bsrinf.mouse)

Format

An example of an object created by inference function

Description

Output from the 'learnParameters' function to get BulkSignalR statistical model parameters.

Usage

data(bsrinf.spa)

Format

From the previous object 'bsrdm.spa', you can generate inferences by calling its method 'BSRInference'. The resulting BSR-Inference object is 'bsrinf.spa', It contains all the inferred L-R interactions with their associated pathways and corrected p-values. 'learnParameters' was also called to get statistical model parameters.

Source

http://spatial.libd.org/spatialLIBD/

Description

Computes putative LR interactions along with their statistical confidence. In this initial inference, all the relevant pathways are reported, see reduction functions to reduce this list.

Usage

BSRInference(
  obj,
  rank.p = 0.55,
  min.cor = 0.25,
  restrict.genes = NULL,
  reference = c("REACTOME-GOBP", "REACTOME", "GOBP"),
  max.pw.size = NULL,
  min.pw.size = NULL,
  min.positive = NULL,
  use.full.network = FALSE,
  restrict.pw = NULL,
  with.complex = NULL,
  fdr.proc = c("BH", "Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BY", "ABH",
    "TSBH")
)

Arguments

Details

Perform the initial ligand-receptor inference. Initial means that no reduction is applied. All the (ligand, receptor, downstream pathway) triples are reported, i.e., a given LR pair may appear multiple times with different pathways downstream the receptor. Specific reduction functions are available from the package to operate subsequent simplifications based on the BSRInference object created by the initial inference.

Parameters defining minimum/maximum pathway sizes, etc. are set to NULL by default, meaning that their values will be taken from what was set during the training of the statistical model with "learnParameters"

To use different values at the time of inference sounds like a bad idea, although this could be used to explore without retraining the underlying model. Retraining of the model with adjusted parameters is advised following such an exploration.

Value

A BSRInference object with initial inferences set.

Examples

data(bsrdm, package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")

# We use a subset of the reference to speed up
# inference in the context of the example.

reactSubset <- getResource(resourceName = "Reactome",
cache = FALSE)

subset <- c("REACTOME_BASIGIN_INTERACTIONS",
"REACTOME_SYNDECAN_INTERACTIONS",
"REACTOME_ECM_PROTEOGLYCANS",
"REACTOME_CELL_JUNCTION_ORGANIZATION")

reactSubset <- reactSubset[
reactSubset$`Reactome name` %in% subset,]

resetPathways(dataframe = reactSubset,
resourceName = "Reactome")

bsrinf <- BSRInference(bsrdm,
    min.cor = 0.2,restrict.genes=immune.signatures$gene,
    reference="REACTOME")

Description

An S4 class to represent inferred ligand-receptor interactions.

Details

This class contains inferred LR interactions along with their statistical significance. Data representation supports subsequent reductions to pathways, etc. See reduction functions "reduceToBestPathway", "reduceToLigand", "reduceToReceptor" and "reduceToPathway".

Slots

LRinter

A data frame describing the (ligand,receptor,pathway) triples with P- and Q-values.

ligands

A list of ligands, one entry per LR interaction.

receptors

A list of receptors, one entry per LR interaction.

tg.genes

A list of target genes, one entry per LR interaction.

tg.corr

A list of target gene correlations to the receptor, one entry per interaction

inf.param

The parameters used for the inference.

Examples

new("BSRInference")

Description

This method supports two configurations that we refer to as paracrine and autocrine.

Usage

BSRInferenceComp(
  obj,
  cmp.name,
  src.cmp.name = NULL,
  rank.p = 0.55,
  max.pval = 0.01,
  min.logFC = 1,
  neg.receptors = FALSE,
  pos.targets = FALSE,
  neg.targets = FALSE,
  min.t.logFC = 0.5,
  restrict.genes = NULL,
  use.full.network = FALSE,
  reference = c("REACTOME-GOBP", "REACTOME", "GOBP"),
  max.pw.size = 400,
  min.pw.size = 5,
  min.positive = 2,
  restrict.pw = NULL,
  with.complex = TRUE,
  fdr.proc = c("BH", "Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BY", "ABH",
    "TSBH")
)

Arguments

Details

In the autocrine case, a single cluster comparison name is provided. In the corresponding cluster comparison, a group of samples A was compared to a group of samples B to determine fold-changes and associated P-values. The inferred ligand-receptor interactions take place in the samples of group A. They are paracrine interactions in the case of single-cell data or they take place in the same tissue represented by cluster A. A typical single-cell example would be a population of macrophages (group A) compared to all the other populations (group B) to represent specific increased or decreased expression in macrophages. The resulting ligand-receptor interactions will be autocrine interactions that are exacerbated (or reduced depending on the chosen parameters) in macrophages.

In the paracrine case, two cluster comparison names must be provided. For instance, a first comparison coul involved macrophages versus all the other cell populations as above. The second comparison could be B-cells against all the other populations. Now, calling BSRInferenceComp() with comparison macrophages vs. the rest and, as source comparison, B-cells vs. the rest, will result in inferring interactions between B-cells (ligands) and macrophages (receptors and downstream pathways). To obtain macrophages to B-cells paracrine interactions, it is necessary to call the method a second time with permuted cluster comparison names. Another example in spatial transcriptomics could be two thin bands at the boundary of two tissue regions, one emitting the ligand and the other one expressing the receptor.

In this initial inference, all the receptor-containing pathways are reported, see reduction functions to reduce this list.

Perform the initial ligand-receptor inference. Initial means that no reduction is applied. All the (ligand, receptor, downstream pathway) triples are reported, i.e., a given LR pair may appear multiple times with different pathways downstream the receptor. Specific reduction functions are available from the package to operate subsequent simplifications based on the BSRInferenceComp object created by this method.

Here, ligand-receptor interactions are inferred based on gene or protein regulation-associated P-values when comparing two clusters of samples. Since a BSRDataModelComp object can contain several such comparisons, the name of the comparison to use must be specified (parameter cmp.name).

Note that since the introduction of the use.full.network parameter (April 29, 2024), the pathway sizes are always computed before potential intersection with the observed data (use.full.network set to FALSE) for consistency. Accordingly, the minimum and maximum pathway default values have been raised from 5 & 200 to 5 & 400 respectively. By default, use.full.network is set to FALSE.

In addition to statistical significance estimated according to BulkSignalR statistical model, we compute SingleCellSignalR original LR-score, based on L and R cluster average expression. In the paracrine case, L average expression is taken from the source cluster.

Value

A BSRInferenceComp object with initial inferences set.

Examples

data(bsrdm.comp, package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")

# infer ligand-receptor interactions from the comparison
bsrinf.comp <- BSRInferenceComp(bsrdm.comp, max.pval = 1, 
reference="REACTOME",
"random.example")

Description

An S4 class to represent ligand-receptor interactions inferred from a comparison between two clusters of samples. This class inherits from BSRInference.

Details

This class is contains inferred LR interactions along with their statistical significance. Data representation supports subsequent reductions to pathways, etc. See reduction functions "reduceToBestPathway", "reduceToLigand", "reduceToReceptor" and "reduceToPathway".

Slots

cmp.name

The name of the BSRClusterComp object in a BSRDataModelComp object comp list.

src.cmp.name

The name of an optional BSRClusterComp object in a BSRDataModelComp object comp list in case paracrine inferences were performed.

tg.pval

A list of target gene P-values, one entry per interaction

tg.logFC

A list of target gene logFC, one entry per interaction

tg.expr

A list of target gene expression, one entry per interaction

Examples

new("BSRInferenceComp")

Description

Obtains gene signatures reflecting ligand-receptor as well as receptor downstream activity to score ligand-receptor pairs across samples subsequently with "scoreLRGeneSignatures"

Usage

BSRSignature(obj, pval.thres = NULL, qval.thres = NULL, with.pw.id = FALSE)

Arguments

Value

A BSRSignature object containing a gene signature for each triple ligand-receptor pair. A reduction to the best pathway for each pair is automatically performed and the gene signature is comprised of the ligand, the receptor, and all the target genes with rank equal or superior to pairs$rank.

Examples

data(bsrinf, package = "BulkSignalR")

bsrinf.redP <- reduceToPathway(bsrinf)
bsrsig.redP <- BSRSignature(bsrinf, qval.thres = 0.001)

Description

S4 class to represent gene signatures of inferred ligand-receptor interactions, including their reduced versions.

Slots

ligands

A list of ligands, one entry per LR interaction.

receptors

A list of receptors, one entry per LR interaction.

tg.genes

A list of target genes, one entry per LR interaction.

pathways

An atomic vector of pathway names, one per interaction.

tg.corr

A list of target genes correlation.

Examples

new("BSRSignature")

Description

Obtains gene signatures reflecting ligand-receptor as well as receptor downstream activity to score ligand-receptor pairs across samples subsequently with "scoreLRGeneSignatures"

Usage

BSRSignatureComp(obj, pval.thres = NULL, qval.thres = NULL, with.pw.id = FALSE)

Arguments

Value

A BSRSignatureComp object containing a gene signature for each triple ligand-receptor pair. A reduction to the best pathway for each pair is automatically performed and the gene signature is comprised of the ligand, the receptor, and all the target genes with rank equal or superior to pairs$rank.

Examples

data(bsrinf.comp, package = "BulkSignalR")

bsrinf.redP <- reduceToPathway(bsrinf.comp)
bsrsig.redP <- BSRSignatureComp(bsrinf.redP, qval.thres = 0.001)

Description

S4 class to represent gene signatures associated with ligand-receptor interactions that were inferred from the comparison of two clusters of samples. This class inherits from BSRSignature.

Slots

cmp.name

The name of the comparison.

tg.pval

A list of target genes P-values.

tg.logFC

A list of target genes logFC.

tg.expr

A list of target genes expression

Examples

new("BSRSignatureComp")

Description

Quick check to observe LR - Pathways association with their respective correlation and Q-values.

Usage

bubblePlotPathwaysLR(
  bsrinf,
  pathways,
  qval.thres = 1,
  filter.L = NULL,
  filter.R = NULL,
  color = "#16a647",
  pointsize = 6
)

Arguments

Value

A bubble plot displayed in the current viewport or in a file in case a filename was provided.

This is a convenience function to propose a simple way of representing LR - Pathways association with their respective correlation and Q-values.

Examples

data(bsrinf, package = "BulkSignalR")
pathways <- LRinter(bsrinf)[1,c("pw.name")]
bubblePlotPathwaysLR(bsrinf,
pathways = pathways,
qval.thres = 0.1,
color = "red",
pointsize = 8
)

Description

Delete the content of cache directory.

Usage

cacheClear(dir = c("both", "resources", "database"))

Arguments

Value

Returns 'NULL', invisibly.

Examples

cacheClear(dir="database")
# need to recreate database in order to run examples well
createDatabase(verbose=TRUE)

Description

Get cache content informations for specific cache dir.

Usage

cacheInfo(dir = c("both", "resources", "database"))

Arguments

Value

Returns 'NULL', invisibly.

Examples

cacheInfo()

Description

Check to see if some ressources has has been updated.

Usage

cacheVersion(dir = c("both", "resources", "database"))

Arguments

Value

Returns 'NULL', invisibly.

Examples

cacheVersion()

Description

Count how many times and with which weights cell types were involved in the (L,R,pathway) triples that targeted genes in a gene set.

Usage

cellTypeFrequency(rel, lr, min.n.genes = 1)

Arguments

Value

A list of two slots: t for counting how many times each cell type is involved; s for summing the weights of each involved cell type.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")
data(tme.signatures, package = "BulkSignalR")
data(p.EMT, package = "BulkSignalR")

immune.signatures <- immune.signatures[immune.signatures$signature %in%
    c("T cells"), ]

signatures <- rbind(immune.signatures, tme.signatures[
    tme.signatures$signature %in% c("Fibroblasts"),
])

tme.scores <- scoreSignatures(bsrdm, signatures)

# assignment
lr2ct <- assignCellTypesToInteractions(bsrdm, bsrinf, tme.scores)

# relate to p-EMT (should be done in HNSCC normally, not in SDC)
p.EMT <- p.EMT$gene
triggers <- relateToGeneSet(bsrinf, p.EMT)

# counts
cf <- cellTypeFrequency(triggers, lr2ct)

Description

Generate a igraph object including all the links between cell types.

Usage

cellularNetwork(tab)

Arguments

Value

A igraph object containing all the links in the cellular network.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")
data("tme.signatures", package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")

immune.signatures <- immune.signatures[immune.signatures$signature %in%
    c("T cells"), ]

signatures <- rbind(immune.signatures, tme.signatures[
    tme.signatures$signature %in% c("Fibroblasts"),
])

tme.scores <- scoreSignatures(bsrdm, signatures)

# assignment
lr2ct <- assignCellTypesToInteractions(bsrdm, bsrinf, tme.scores)

# cellular network
g.table <- cellularNetworkTable(lr2ct)
gCN <- cellularNetwork(g.table)

#plot(gCN, edge.width=5*E(gCN)$score)

Description

Generate a data.frame including all the links between cell types mediated by L-R interactions with their respective weights.

Usage

cellularNetworkTable(lr, autocrine = FALSE)

Arguments

Value

A data.frame containing all the links in the cellular network. A link is created between two cell types as soon as there was a L-R interaction that was associated with both cell types. The link is given a score equal to the geometric mean of each cell type assignment r2.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")
data(tme.signatures, package = "BulkSignalR")

immune.signatures <- immune.signatures[immune.signatures$signature %in%
    c("T cells"), ]

signatures <- rbind(immune.signatures, tme.signatures[
    tme.signatures$signature %in% c("Fibroblasts"),
])

tme.scores <- scoreSignatures(bsrdm, signatures)

# assignment
lr2ct <- assignCellTypesToInteractions(bsrdm, bsrinf, tme.scores)

# cellular network
g.table <- cellularNetworkTable(lr2ct)

Description

Chord diagram.

Usage

chordDiagramLR(
  bsrinf,
  pw.id.filter = NULL,
  qval.thres = 1,
  ligand = NULL,
  receptor = NULL,
  limit = 20
)

Arguments

Value

Circos Plot on the screen or a file

Examples

data(bsrinf, package = "BulkSignalR")
chordDiagramLR(bsrinf,
pw.id.filter = "R-HSA-3000178",
limit = 20,
ligand="ADAM15", 
receptor="ITGAV"
)

Description

Convert BSRDataModel to BSRDataModelComp

Arguments

Value

A BSRDataModelComp object

Examples

bsrdm <- new("BSRDataModel")
bsrdm.comp <- as(bsrdm, "BSRDataModelComp")

Description

Cluster A columns accessor

Usage

## S4 method for signature 'BSRClusterComp'
colClusterA(x)

Arguments

Value

col.clusterA

Examples

bsrcc <- new("BSRClusterComp")
colClusterA(bsrcc)

Description

Cluster B columns accessor

Usage

## S4 method for signature 'BSRClusterComp'
colClusterB(x)

Arguments

Value

col.clusterB

Examples

bsrcc <- new("BSRClusterComp")
colClusterB(bsrcc)

Description

Comparisons list accessor

Usage

## S4 method for signature 'BSRDataModelComp'
comparison(x)

Arguments

Value

comp

Examples

bsrdm.comp <- new("BSRDataModelComp")
comparison(bsrdm.comp)

Description

Comparison name accessor

Comparison name accessor

Usage

## S4 method for signature 'BSRInferenceComp'
comparisonName(x)

## S4 method for signature 'BSRSignatureComp'
comparisonName(x)

Arguments

Value

cmp.name

cmp.name

Examples

bsrinf <- new("BSRInferenceComp")
comparisonName(bsrinf)

Description

By default, BulkSignalR is designed to work with Homo sapiens. In order to work with other organisms, gene names need to be first converted to human by orthology.

Usage

convertToHuman(counts, dictionary)

Arguments

Value

Return a counts matrix transposed for Human.

Examples

data(bodyMap.mouse)

idx <- sample(nrow(bodyMap.mouse), 500)
bodyMap.mouse <- bodyMap.mouse[idx,]

ortholog.dict <- findOrthoGenes(
    from_organism = "mmusculus",
    from_values = rownames(bodyMap.mouse)
)

matrix.expression.human <- convertToHuman(
    counts = bodyMap.mouse,
    dictionary = ortholog.dict
)

Description

Fetch LR database from remote location.

Usage

createDatabase(onRequest = TRUE, verbose = FALSE)

Arguments

Value

Returns 'NULL', invisibly.

Examples

print("Function already called elsewhere by cacheClear()")
# createDatabase(onRequest = FALSE)

Description

Create cache for all resources (pathways, or PWC network) downloaded from the web when library is first loaded. This part is handled with BiocFileCache. Otherwise datatabase, is handled by another process not relying on BiocFileCache instance.

Usage

createResources(onRequest = TRUE, verbose = FALSE)

Arguments

Value

Returns 'NULL', invisibly.

Examples

createResources(onRequest=FALSE)

Description

Cluster comparison statistics accessor

Usage

## S4 method for signature 'BSRClusterComp'
differentialStats(x)

Arguments

Value

diffferential.stats

Examples

bsrcc <- new("BSRClusterComp")
differentialStats(bsrcc)

Description

By default, BulkSignalR is designed to work with Homo sapiens. In order to work with other organisms, gene names need to be first converted to human following an orthology mapping process.

Usage

findOrthoGenes(
  from_organism,
  from_values,
  method = c("gprofiler", "homologene", "babelgene")
)

Arguments

Value

Return a data frame with 2 columns containing the gene names for the two species. First column is the gene name from the source organism and the second column corresponds to the homologous gene name in Homo sapiens.

Examples

data(bodyMap.mouse)

idx <- sample(nrow(bodyMap.mouse), 20)
bodyMap.mouse <- bodyMap.mouse[idx,]

ortholog.dict <- findOrthoGenes(
    from_organism = "mmusculus",
    from_values = rownames(bodyMap.mouse)
)

Description

Generate a series of individual spatial score plots in a folder. Not limited to BulkSignalR gene signature scores.

Usage

generateSpatialPlots(
  scores,
  areas,
  plot.folder,
  width = 5,
  height = 3,
  pointsize = 8,
  rev.y = TRUE,
  ref.plot = TRUE,
  image.raster = NULL,
  x.col = "array_col",
  y.col = "array_row",
  label.col = "label",
  idSpatial.col = "idSpatial",
  cut.p = 0.01,
  low.color = "royalblue3",
  mid.color = "white",
  high.color = "orange",
  title.fs = 12,
  legend.fs = 10,
  axis.fs = 10,
  label.fs = 12,
  dot.size = 0.5,
  ref.colors = NULL
)

Arguments

Details

A set of PDF files are created in the provided folder.

Value

Create PDF file and returns 'NULL', invisibly.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(bsrinf.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

thres <- 0.01
bsrinf.red <- reduceToBestPathway(bsrinf.spa)
s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

generateSpatialPlots(scores.red[1:2,],
annotation.spa, ".", label.col = "ground_truth")

Description

Fetch LR complexes from database and and return a dataframe

Usage

getComplexes(idRelease = NULL)

Arguments

Value

Returns dataframe Complexex, invisibly.

Examples

getComplexes(idRelease=1)

Description

Fetch LR interactions from database and and return a dataframe

Usage

getInteractions(idRelease = NULL)

Arguments

Value

Returns dataframe LR interactions, invisibly.

Examples

getInteractions(idRelease=1)

Description

Generate a ligand-receptor network from a BSRInference object and add the shortest paths from the receptors to correlated target genes following Reactome and KEGG pathways.

Usage

getLRIntracellNetwork(
  bsrinf,
  pval.thres = NULL,
  qval.thres = NULL,
  min.cor = 0.25,
  max.pval = NULL,
  min.logFC = NULL,
  pos.targets = FALSE,
  neg.targets = FALSE,
  restrict.pw = NULL,
  node.size = 5
)

Arguments

Value

An igraph object featuring the ligand-receptor-downstream signaling network. Default colors and node sizes are assigned, which can be changed afterwards if necessary.

The target genes to which the min.cor correlation is imposed are those listed in tgGenes(bsrinf), correlations are in tgCorr(bsrinf). The construction of shortest paths from the receptors to those selected targets adds other genes, which were either some targets with too low correlation or genes along the shortest paths to reach the selected targets.

Examples

data(bsrinf, package = "BulkSignalR")

bsrinf.redBP <- reduceToBestPathway(bsrinf)

pairs <- LRinter(bsrinf.redBP)
top <- unique(pairs[pairs$pval < 1e-20, c("pw.id", "pw.name")])

gLRintra.res <- getLRIntracellNetwork(bsrinf.redBP,
qval.thres = 0.01,
restrict.pw = top[1,]$pw.id
)

# write.graph(gLRintra, file="SDC-LR-intracellular-network.reduced.graphml",
# format="graphml")

Description

Generate a ligand-receptor network from a ligand-receptor table.

Usage

getLRNetwork(
  bsrinf,
  pval.thres = NULL,
  qval.thres = NULL,
  node.size = 5,
  red.pairs = NULL
)

Arguments

Value

An igraph object featuring the ligand-receptor network. Default colors and node sizes are assigned, which can be changed afterwards if necessary.

Examples

data(bsrinf, package = "BulkSignalR")

gLR <- getLRNetwork(bsrinf, qval.thres = 1e-4)
# plot(gLR)
# write.graph(gLR, file="SDC-LR-network.graphml", format="graphml")

Description

Basic statistics about hit pathways

Usage

## S4 method for signature 'BSRInference'
getPathwayStats(obj, pval.thres = NULL, qval.thres = NULL)

Arguments

Value

A table with the pathways selected after the chosen threshold was applied to rows in LRinter(obj). Each pathway is reported along with various statistics: the number of selected receptors in this pathway, the total number of receptors described this pathway, the number of selected ligand-receptor pairs hitting this pathway, and the total number of ligand-receptor pairs described that could hit this pathway.

Obviously, one could imagine computing enrichment in receptors or ligand-receptor pairs based on such statistics, but the actual meaning of such an analysis would be ambiguous since the pathways were already selected as significantly regulated by the receptor. We thus did not implement this (hypergeometric test) computation.

Examples

data(bsrinf, package = "BulkSignalR")

pw.stat <- getPathwayStats(bsrinf)

Description

Get resources (pathways, or PathwayCommons network from https://www.pathwaycommons.org/) stored in the cache.

Usage

getResource(resourceName = NULL, cache = FALSE)

Arguments

Value

Returns a dataframe of the requested resource.

Examples

reactome <- getResource(resourceName = "Reactome",cache=TRUE)

Description

A dataset containing gene signatures for general immune cell populations.

Usage

data(immune.signatures)

Format

A data frame with 1541 rows and 2 variables:

gene

HUGO gene symbol

signature

cell population name

Source

PanglaoDB (Franzén et al., Database, 2019).

Description

Inference parameters accessor

Usage

## S4 method for signature 'BSRInference'
inferenceParameters(x)

Arguments

Value

inf.param

Examples

bsrinf <- new ("BSRInference")
inferenceParameters(bsrinf)

Description

organism accessor

Usage

## S4 method for signature 'BSRDataModel'
initialOrganism(x)

Arguments

Value

initialOrganism

Examples

bsrdm <- new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)
initialOrganism(bsrdm)

Description

Model parameter accessor

Usage

## S4 method for signature 'BSRDataModel'
initialOrthologs(x)

Arguments

Value

initialOrthologs

Examples

bsrdm <- new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)
initialOrthologs(bsrdm)

Description

Unique entry point for training the parameters behind BulkSignalR statistical models.

Usage

## S4 method for signature 'BSRDataModel'
learnParameters(
  obj,
  plot.folder = NULL,
  verbose = FALSE,
  n.rand.LR = 5L,
  n.rand.RT = 2L,
  with.complex = TRUE,
  max.pw.size = 400,
  min.pw.size = 5,
  min.positive = 4,
  quick = FALSE,
  null.model = c("automatic", "mixedNormal", "normal", "kernelEmpirical", "empirical",
    "stable"),
  filename = NULL,
  min.corr.LR = -1
)

Arguments

Details

Estimates the model parameters that are stored in the slot param.

In a reference pathway, i.e., a Reactome pathway or the genes of a GOBP term, the target genes are the genes coding for proteins forming a complex with the receptor and the genes in the pathway downstream the receptor, which are given as regulated by the pathway. If with.complex is set to FALSE, then only the regulated genes are considered. Participation to a complex, being regulated, and pathway directed topologies are defined by Reactome and KEGG pathways as provided by PathwayCommons.

The min.pw.size, max.pw.size, and min.positive parameters should be identical to the values intended when searching for ligand-receptor pairs with .getCorrelatedLR) and .checkReceptorSignaling) Although the statistical distributions are rather robust, it is not advisable to use different parameters that could introduce unanticipated biases, but for saving compute time and exploring.

The maximum pathway size is used to limit the redundancy inherent to GOBP and Reactome. The minimum pathway size is used to avoid overspecific, noninformative results.

BulkSignalR approach relies on modeling (Spearman) correlations and different models of null distributions are available for this purpose (parameter null.model). By default, the "automatic" option is selected meaning that censored normal and mixed normal as well as an empirical model based on Gaussian kernels (R density() function) are compared to pick the one closest to the data. Preference is given to normal and then mixture of normal over the empirical version for comparable quality of fit. It is also to bypass the automatic selection. Fitting of an alpha-stable distribution is quite time consuming as the computation of its PDF is compute-intensive. Finally, in the automaic selection mode, the choice of the actual model will be done based on the L-R null assuming a similar shape for the R-T null (with different parameters though, unless quick was set to TRUE).

Note that since the introduction of the use.full.network parameter (April 29, 2024) in the BSRInference method parameters, the pathway sizes are always computed before potential intersection with the observed data (use.full.network set to FALSE) for consistency. Accordingly, the minimum and maximum pathway default values have been raised from 5 & 200 to 5 & 400 respectively. By default, use.full.network is set to TRUE, meaning no intersection and hence larger pathways.

Value

A BSRDataModel object with trained model parameters

Examples

data(sdc, package = "BulkSignalR")
idx <- sample(nrow(sdc), 4000)
bsrdm <- BSRDataModel(sdc[idx, c("N22","SDC17")],min.LR.found = 20)
bsrdm <- learnParameters(bsrdm, n.rand.LR = 1L,
verbose=FALSE,quick=TRUE)

Description

ligands accessor

ligands accessor

Usage

## S4 method for signature 'BSRInference'
ligands(x)

## S4 method for signature 'BSRSignature'
ligands(x)

Arguments

Value

ligands

ligands

Examples

bsr.sig <- new("BSRSignature")
ligands(bsr.sig)

Description

log.transformed accessor

Usage

## S4 method for signature 'BSRDataModel'
logTransformed(x)

Arguments

Value

logTransformed

Examples

bsrdm <- new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)
logTransformed(bsrdm)

Description

LRinter accessor

Usage

## S4 method for signature 'BSRInference'
LRinter(x)

Arguments

Value

LRinter

Examples

bsrinf <- new ("BSRInference")
LRinter(bsrinf)

Description

Simplified LRinter accessor with focus on the LR-score

Usage

## S4 method for signature 'BSRInferenceComp'
LRinterScore(x)

Arguments

Value

LRinterScore

Examples

data(bsrinf.comp, package = "BulkSignalR")
LRinterScore(bsrinf.comp)[5,]

Description

Simplified LRinter accessor reporting the essential columns

Simplified LRinter accessor reporting the essential columns

Usage

## S4 method for signature 'BSRInference'
LRinterShort(x)

## S4 method for signature 'BSRInferenceComp'
LRinterShort(x)

Arguments

Value

LRinterShort

LRinterShort

Examples

data(bsrinf.comp, package = "BulkSignalR")
LRinterShort(bsrinf.comp)[5,]

Description

Get maximal ligand expression at nearby locations

Usage

maxLigandSpatialCounts(
  bsrdm,
  areas,
  nnn = 4,
  radius = NULL,
  x.col = "array_col",
  y.col = "array_row"
)

Arguments

Details

Ligand expression data contained in a BSRDataModel object are modified to consider the possibility that the ligand of a L-R interaction might be expressed at nearby locations. This is achieved replacing each ligand expression by its maximum over the central location and its neighbors. Since ligands and receptors are never used as gene targets in computing the receptor downstream signal correlations, this substitution is compatible with our statistical model. Moreover, the reciprocal configuration where the ligand is expressed at the central location and hits a receptors at a neighbor location is covered when the same ligand maximization scheme is applied to the neighbor. L-R localization and gene signature scoring is defined by the location at which the receptor is expressed after applying this function.

Two strategies are available to identify the neighbors. It is possible to simply set the number of nearest-neighbors (parameter nnn). An alternative consists in providing a distance radius (radius) along with a a maximum number of nearest-neighbors within the radius (nnn.radius). To properly define the radius, the user must know the location coordinates. The strategy with the radius enables having corner locations with two neighbors only and border locations with three neighbors only, whereas to simply set a maximum of four neighbors for instance would retrieve the four closest neighbors in every case.

Value

A BSRDataModel object containing the maximized ligand expressions.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

max.bsrdm <- maxLigandSpatialCounts(bsrdm.spa, annotation.spa,
radius = 1.2, nnn = 4)

Description

Mu accessor

Usage

## S4 method for signature 'BSRDataModelComp'
mu(x)

Arguments

Value

mu

Examples

bsrdm.comp <- new("BSRDataModelComp")
mu(bsrdm.comp)

Description

Normalized count matrix accessor

Usage

## S4 method for signature 'BSRDataModel'
ncounts(x)

Arguments

Value

ncounts

Examples

bsrdm <- new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)
ncounts(bsrdm)

Description

Normalization accessor

Usage

## S4 method for signature 'BSRDataModel'
normalization(x)

Arguments

Value

normalization

Examples

bsrdm <- new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)
normalization(bsrdm)

Description

Synthetic object used during the call to the function 'resetToInitialOrganism“

Usage

data(ortholog.dict)

Format

An example of a dataframe created by findOrthoGenes

Description

A dataset containing a partial EMT gene signature.

Usage

data(p.EMT)

Format

A data frame with 100 rows and 1 variables:

gene

HUGO gene symbol

Source

Puram, SV & al., Cell, 2017.

Description

Model parameter accessor

Usage

## S4 method for signature 'BSRDataModel'
parameters(x)

Arguments

Value

param

Examples

bsrdm <- new("BSRDataModel",
    ncounts = matrix(1.5,
        nrow = 2, ncol = 2,
        dimnames = list(c("A", "B"), c("C", "D"))
    ),
    log.transformed = TRUE,
    normalization = "TC"
)
parameters(bsrdm)

Description

pathways accessor

Usage

## S4 method for signature 'BSRSignature'
pathways(x)

Arguments

Value

pathways

Examples

bsr.sig <- new("BSRSignature")
pathways(bsr.sig)

Description

receptors accessor

receptors accessor

Usage

## S4 method for signature 'BSRInference'
receptors(x)

## S4 method for signature 'BSRSignature'
receptors(x)

Arguments

Value

receptors

receptors

Examples

bsr.sig <- new("BSRSignature")
ligands(bsr.sig)

Description

Keep one pathway per ligand-receptor pair

Keep one pathway per ligand-receptor pair

Usage

## S4 method for signature 'BSRInference'
reduceToBestPathway(obj)

## S4 method for signature 'BSRInferenceComp'
reduceToBestPathway(obj)

Arguments

Details

Ligand-receptor pairs are evaluated in relation with pathways that allow checking receptor downstream correlations. It is thus possible that several pathways are reported for a same LR pair.

Ligand-receptor pairs are evaluated in relation with pathways that allow checking receptor downstream correlations. It is thus possible that several pathways are reported for a same LR pair.

Value

A BSRInference object reduced to only report one pathway per ligand-receptor pair. The pathway with the smallest P-value is selected.

A BSRInferenceComp object reduced to only report one pathway per ligand-receptor pair. The pathway with the smallest P-value is selected.

Examples

data(bsrinf, package = "BulkSignalR")
bsrinf.redBP <- reduceToBestPathway(bsrinf)

data(bsrinf.comp, package = "BulkSignalR")


reduceToBestPathway(bsrinf.comp)

Description

Simplifies a ligand-receptor table to focus on the ligands.

Simplifies a ligand-receptor table to focus on the ligands.

Usage

## S4 method for signature 'BSRInference'
reduceToLigand(obj)

## S4 method for signature 'BSRInferenceComp'
reduceToLigand(obj)

Arguments

Value

A BSRInference object but reduced to one row per ligand. All the receptors are combined in a semi-colon-separated list surrounded by curly brackets in the tabular slot LRinter, and in vectors in the ligands (list) slot.

The reported P-value and target genes are those from the pathway with the smallest P-value.

A BSRInferenceComp object but reduced to one row per ligand. All the receptors are combined in a semi-colon-separated list surrounded by curly brackets in the tabular slot LRinter, and in vectors in the ligands (list) slot.

The reported P-value and target genes are those from the pathway with the smallest P-value. The same logic applies to the LR-score, and the receptor expression.

Examples

data(bsrinf, package = "BulkSignalR")

bsrinf.redL <- reduceToLigand(bsrinf)

data(bsrinf.comp, package = "BulkSignalR")

bsrinf.redL <- reduceToLigand(bsrinf.comp)

Description

Simplifies a ligand-receptor inference object to focus on the pathways.

Simplifies a ligand-receptor inference object to focus on the pathways.

Usage

## S4 method for signature 'BSRInference'
reduceToPathway(obj)

## S4 method for signature 'BSRInferenceComp'
reduceToPathway(obj)

Arguments

Value

A BSRInference object reduced to only report one row per pathway. The information of which ligand interacted with which receptor is lost as all the ligands and all the receptors forming pairs related to a certain pathway are combined. For a given pathway, the reported P-values and target genes are those of the best ligand-receptor pair that was in this pathway. Receptors and ligands are combined in two semi-colon-separated lists surrounded by curly brackets in the tabular slot LRinter, while the list representation slots (ligands and receptors) are update accordingly.

A BSRInferenceComp object reduced to only report one row per pathway. The information of which ligand interacted with which receptor is lost as all the ligands and all the receptors forming pairs related to a certain pathway are combined. For a given pathway, the reported P-values and target genes are those of the best ligand-receptor pair that was in this pathway. The same logic applies to the LR-score, and the ligand and receptor expression. Receptors and ligands are combined in two semi-colon-separated lists surrounded by curly brackets in the tabular slot LRinter, while the list representation slots (ligands and receptors) are update accordingly.

Examples

data(bsrinf, package = "BulkSignalR")

bsrinf.redP <- reduceToPathway(bsrinf)
data(bsrinf.comp, package = "BulkSignalR")

bsrinf.redP <- reduceToPathway(bsrinf.comp)

Description

Simplifies a ligand-receptor table to focus on the receptors.

Simplifies a ligand-receptor table to focus on the receptors.

Usage

## S4 method for signature 'BSRInference'
reduceToReceptor(obj)

## S4 method for signature 'BSRInferenceComp'
reduceToReceptor(obj)

Arguments

Value

BSRInference object reduced to one row per receptor. All the ligands are combined in a semi-colon-separated list surrounded by curly brackets in the tabular slot LRinter, and in vectors in the ligands (list) slot.

The reported P-value and target genes are those from the line with the pathway featuring the smallest P-value.

BSRInferenceComp object reduced to one row per receptor. All the ligands are combined in a semi-colon-separated list surrounded by curly brackets in the tabular slot LRinter, and in vectors in the ligands (list) slot.

The reported P-value and target genes are those from the line with the pathway featuring the smallest P-value. The same logic applies to the LR-score, and the ligand expression.

Examples

data(bsrinf, package = "BulkSignalR")

bsrinf.redR <- reduceToReceptor(bsrinf)

data(bsrinf.comp, package = "BulkSignalR")
# reduction
bsrinf.redR <- reduceToReceptor(bsrinf.comp)

Description

Finds ligands related to a gene set by following receptor, and receptor downstream pathway targets.

Usage

relateToGeneSet(bsrinf, gs, min.cor = 0.25, qval.thres = 0.001)

Arguments

Value

A data.frame listing all the (L,R,pathway) triples that lead to at least one gene in the gene set. The number of genes found by each triple is indicated in the column n.genes.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")

data(p.EMT, package = "BulkSignalR")
p.EMT <- p.EMT$gene
triggers <- relateToGeneSet(bsrinf, p.EMT)

Description

Remove a comparison from a BSRDataModelComp object.

Usage

## S4 method for signature 'BSRDataModelComp'
removeClusterComp(obj, cmp.name)

Arguments

Details

Remove the comparison with cmp.name from the list of comparisons contained in obj.

Value

A BSRDataModelComp object.

Examples

# prepare data
data(sdc, package = "BulkSignalR")
normal <- grep("^N", names(sdc))
bsrdm <- BSRDataModel(sdc[, -normal])

# define the comparison
bsrdm.comp <- as(bsrdm, "BSRDataModelComp")
colA <- as.integer(1:3)
colB <- as.integer(12:15)
n <- nrow(ncounts(bsrdm.comp))
stats <- data.frame(
    pval = runif(n), logFC = rnorm(n, 0, 2),
    expr = runif(n, 0, 10)
)
rownames(stats) <- rownames(ncounts(bsrdm.comp))
bsrcc <- BSRClusterComp(bsrdm.comp, colA, colB, stats)

bsrdm.comp <- addClusterComp(bsrdm.comp, bsrcc, "random.example")
bsrdm.comp <- removeClusterComp(bsrdm.comp, "random.example")

Description

A method to re-score an existing BSRInference object (P- and Q-value estimations).

A method to re-score an existing BSRInferenceComp object (P- and Q-value estimations).

Usage

## S4 method for signature 'BSRInference'
rescoreInference(
  obj,
  param,
  rank.p = 0.55,
  fdr.proc = c("BH", "Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BY", "ABH",
    "TSBH")
)

## S4 method for signature 'BSRInferenceComp'
rescoreInference(
  obj,
  param = NULL,
  rank.p = 0.55,
  fdr.proc = c("BH", "Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BY", "ABH",
    "TSBH")
)

Arguments

Details

A BSRInference object should be created by calling "BSRInference"

Parameters controlling the estimation of the statistical significance of the ligand/receptor/pathway triples (param) are provided at the time of calling the latter method.

Nonetheless, it might be useful to change the initially-provided parameters, in which case this method should not be called.

A BSRInferenceComp object should be created by calling "BSRInferenceComp"

Value

A BSRInference object.

A BSRInferenceComp object.

Examples

data(bsrinf, package = "BulkSignalR")
data(bsrdm, package = "BulkSignalR")

bsrinf.new <- rescoreInference(bsrinf,
param = parameters(bsrdm))
data(bsrinf.comp, package = "BulkSignalR")

bsrinf.less <- rescoreInference(bsrinf.comp, 
rank.p = 0.75)

Description

User can provide a data frame with 2 columns named ligand and receptor. This can be used to extend or replace the existing LRdb.

Usage

resetLRdb(db, switch = FALSE)

Arguments

Value

Returns 'NULL', invisibly.

Examples

resetLRdb(db = data.frame(ligand = "A2M", receptor = "LRP1"), switch = FALSE)

Description

Network is a dataframe that gives relation between genes. It's composed of 3 columns annoted as follows :

Usage

resetNetwork(network)

Arguments

Details

a.gn : Gene Symbol 1 type : controls-expression-of b.gn : Gene Symbol 2

When the user provide his own network 'type' should be set to 'controls-expression-of'.

Value

Returns 'NULL', invisibly.

Examples

BulkSignalR_Network <- getResource(resourceName = "Network",
 cache = FALSE)
resetNetwork(BulkSignalR_Network)

Description

resetPathways is a function we provide to user to refresh REACTOME and GO-BP content included in BulkSignalR.

Usage

resetPathways(
  dataframe = NULL,
  file = NULL,
  fileType = c("json", "gmt", "txt"),
  resourceName = NULL
)

Arguments

Details

Pathways are defined in Reactome and GoBP databases. Those can be updated using json files from the Human Molecular Signatures Database (MSigDB) at https://www.gsea-msigdb.org/ Gmt file format also can be imported. A dataframe can be used directly also.

Value

Returns 'NULL', invisibly.

Examples

reactSubset <- getResource(resourceName = "Reactome",
cache = TRUE)

subset <- c("REACTOME_BASIGIN_INTERACTIONS",
"REACTOME_SYNDECAN_INTERACTIONS",
"REACTOME_ECM_PROTEOGLYCANS",
"REACTOME_CELL_JUNCTION_ORGANIZATION")

reactSubset <- reactSubset[
reactSubset$`Reactome name` %in% subset,]

resetPathways(dataframe = reactSubset,
resourceName = "Reactome")

Description

Reset gene names to initial organism providen in first instance

Usage

## S4 method for signature 'BSRInference'
resetToInitialOrganism(obj, conversion.dict)

Arguments

Value

An BSRInference object updated for gene names. The gene names are replaced by the ones from the organism providen in first instance.

Examples

data(bodyMap.mouse, package = "BulkSignalR")
data(bsrinf.mouse, package = "BulkSignalR")
data(ortholog.dict, package = "BulkSignalR")

#idx <- sample(nrow(bodyMap.mouse), 7500)

#bodyMap.mouse <- bodyMap.mouse[idx,1:3]

#ortholog.dict <- findOrthoGenes(
#    from_organism = "mmusculus",
#    from_values = rownames(bodyMap.mouse)
#)

#matrix.expression.human <- convertToHuman(
#    counts = bodyMap.mouse,
#    dictionary = ortholog.dict
#)

#bsrdm <- BSRDataModel(
#    counts = matrix.expression.human,
#    species = "mmusculus",
#    conversion.dict = ortholog.dict
#)

#bsrdm <- learnParameters(bsrdm,
#    quick = TRUE  
#)

#reactSubset <- getResource(resourceName = "Reactome",
#cache = TRUE)

#subset <- c("REACTOME_BASIGIN_INTERACTIONS",
#"REACTOME_SYNDECAN_INTERACTIONS",
#"REACTOME_ECM_PROTEOGLYCANS",
#"REACTOME_CELL_JUNCTION_ORGANIZATION")

#reactSubset <- reactSubset[
#reactSubset$`Reactome name` %in% subset,]

#bsrinf.mouse <- BSRInference(bsrdm,reference="REACTOME")

bsrinf <- resetToInitialOrganism(bsrinf.mouse, 
conversion.dict = ortholog.dict)

Description

Compute ligand-receptor gene signature scores over a BSRDataModel.

Compute ligand-receptor gene signature scores over a BSRDataModelComp specific comparison.

Usage

## S4 method for signature 'BSRDataModel'
scoreLRGeneSignatures(
  obj,
  sig,
  LR.weight = 0.5,
  robust = FALSE,
  name.by.pathway = FALSE,
  abs.z.score = FALSE,
  rownames.LRP = FALSE
)

## S4 method for signature 'BSRDataModelComp'
scoreLRGeneSignatures(
  obj,
  sig,
  LR.weight = 0.5,
  robust = FALSE,
  name.by.pathway = FALSE,
  abs.z.score = FALSE,
  rownames.LRP = FALSE
)

Arguments

Value

A matrix containing the scores of each ligand-receptor gene signature in each sample.

A matrix containing the scores of each ligand-receptor gene signature in each sample.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")

bsrinf.redBP <- reduceToBestPathway(bsrinf)
bsrsig.redBP <- BSRSignature(bsrinf.redBP, qval.thres = 0.001)
res <-scoreLRGeneSignatures(bsrdm, bsrsig.redBP,
    name.by.pathway = FALSE
)
# prepare data
data(bsrdm.comp, package = "BulkSignalR")
data(bsrinf.comp, package = "BulkSignalR")

# reduction
bsrinf.red <- reduceToBestPathway(bsrinf.comp)
# signature extraction and scoring
bsrsig.red <- BSRSignatureComp(bsrinf.red, qval.thres = 1e-6)
scores.red <- scoreLRGeneSignatures(bsrdm.comp, bsrsig.red,
    name.by.pathway = TRUE, rownames.LRP = TRUE
)

Description

Scores generic gene signatures over the samples of a BSRDataModel object.

Usage

scoreSignatures(ds, ref.signatures, robust = FALSE)

Arguments

Details

This function relies on a simple average of gene z-scores over each signature. It is no replacement for mode advanced methods such as CIBERSORT or BisqueRNA. It is provided for convenience.

Value

A matrix containing the scores of each gene signature in each sample. Note that ligand-receptor gene signature scores should be computed with "scoreLRGeneSignatures" instead.

Examples

data(sdc, package = "BulkSignalR")
data(bsrdm, package = "BulkSignalR")

data(immune.signatures, package = "BulkSignalR")
imm.scores <- scoreSignatures(bsrdm, immune.signatures)

Description

A dataset containing the read counts of salivary duct carcinomas and adjacent normal tissues.

Usage

data(sdc)

Format

A data frame with 19764 rows and 26 variables.

Source

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE138581

Description

Generate a detailed view related to a chosen interaction made of series of small individual spatial plots: tissue organization (optional), gene signature score, ligand and receptor expression.

Usage

separatedLRPlot(
  v,
  L,
  R,
  ncounts,
  areas,
  inter.name = NULL,
  rev.y = TRUE,
  ref.plot = TRUE,
  image.raster = NULL,
  x.col = "array_col",
  y.col = "array_row",
  label.col = "label",
  idSpatial.col = "idSpatial",
  cut.p = 0.01,
  low.color = "royalblue3",
  mid.color = "white",
  high.color = "orange",
  title.fs = 12,
  legend.fs = 10,
  axis.fs = 10,
  label.fs = 12,
  dot.size = 0.5,
  legend.dot.factor = 10,
  ref.colors = NULL
)

Arguments

Details

A set of spatial plots are generated including an optional reference tissue plot (image or areas represented), the gene signature scores, the ligand expression values, and the receptor expression values.

Value

A set of spatial plots.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(bsrinf.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

thres <- 0.01
bsrinf.red <- reduceToBestPathway(bsrinf.spa)
s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

separatedLRPlot(scores.red, "SLIT2", "GPC1", 
ncounts(bsrdm.spa), 
annotation.spa,
label.col = "ground_truth")

Description

Generate a list of heatmaps for ligand, receptor and target genes for a specific pathway

Usage

signatureHeatmaps(
  pathway,
  bsrdm,
  bsrsig,
  h.width = 6,
  h.height = 9,
  fontsize = 6,
  show_column_names = FALSE
)

Arguments

Value

A plot is created.

This is a convenience function to propose a simple way of representing expression of genes involved in a specific pathway.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")
if(FALSE){
bsrinf.redP <- reduceToPathway(bsrinf)
bsrinf.redPBP <- reduceToBestPathway(bsrinf)
bsrsig.redPBP <- BSRSignature(bsrinf, qval.thres = 1)
pathway1 <- pathways(bsrsig.redPBP)[1]
signatureHeatmaps(
pathway = pathway1,
bsrdm = bsrdm,
bsrsig = bsrsig.redPBP,
h.width = 3,
h.height = 4,
fontsize = 1,
show_column_names = TRUE)
}

Description

Generate a heatmap representing ligand-receptor gene signature scores.

Usage

simpleHeatmap(
  mat.c,
  width = 4,
  height = 3,
  dend.row = NULL,
  dend.spl = NULL,
  cols = NULL,
  pointsize = 4,
  bottom.annotation = NULL,
  n.col.clust = 0,
  n.row.clust = 0,
  gap.size = 0.5,
  cut.p = 0.01,
  row.names = TRUE,
  column.names = TRUE,
  hcl.palette = NULL,
  reverse = FALSE
)

Arguments

Value

A heatmap. Since heatmap plotting tend to be slow on the screen, it is advisable to provide a PDF file name and plot in a file (much faster).

If hcl.palette is set, the colors parameter won't be used.

Extreme values (top and bottom) can be replaced by global quantiles at cut.p and 1-cut.p to avoid color scales shrunk by a few outliers.

This is a convenience function that relies on the ComplexHeatmap package to propose a simple way of representing signature scores. If more advanced features are needed or more graphic parameters should be controlled, users should implement their own function.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")

bsrinf.redBP <- reduceToBestPathway(bsrinf)
bsrsig.redBP <- BSRSignature(bsrinf,
    qval.thres = 0.001
)

scoresLR <- scoreLRGeneSignatures(bsrdm, bsrsig.redBP,
    name.by.pathway = FALSE
)
simpleHeatmap(scoresLR[1:3, ],
    column.names = TRUE,
    hcl.palette = "Cividis",
    width=2, 
    height=1.5)

Description

Smooth spatial expression data

Usage

smoothSpatialCounts(
  bsrdm,
  areas,
  nnn = 4,
  radius = NULL,
  weight.ratio = 0.5,
  x.col = "array_col",
  y.col = "array_row"
)

Arguments

Details

The expression data contained in a BSRDataModel object are smoothed using a weighted average of nearby locations.

Two strategies are available to identify the neighbors. It is possible to simply set the number of nearest-neighbors (parameter nnn). An alternative consists in providing a distance radius (radius) along with a a maximum number of nearest-neighbors within the radius (nnn.radius). To properly define the radius, the user must know the location coordinates. The strategy with the radius enables having corner locations with two neighbors only and border locations with three neighbors only, whereas to simply set a maximum of four neighbors for instance would retrieve the four closest neighbors in every case.

For each location, its nearest-neighbors are found and a weighted average computed with weight.ratio given to the central location itself and a total weight of 1-weight.ratio shared within the neighbors based on the inverse of their distances. In case radius is set, some locations may have less than nnn neighbors (see above). At such locations, the weight given to the central location is augmented according to 1-(1-weight.ratio)*(number of neighbors)/nnn.

Value

A BSRDataModel object containing the smoothed ncounts.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")
sm.bsrdm <- smoothSpatialCounts(bsrdm.spa, annotation.spa,
radius = 1.2, nnn = 4)

Description

Source comparison name accessor

Usage

## S4 method for signature 'BSRInferenceComp'
sourceComparisonName(x)

Arguments

Value

src.comp.name

Examples

bsrinf <- new("BSRInferenceComp")
sourceComparisonName(bsrinf)

Description

Compute the statistical association of L-R interaction score spatial distributions with tissue area labels. Not limited to BulkSignalR gene signature scores.

Usage

spatialAssociation(
  scores,
  areas,
  test = c("Kruskal-Wallis", "ANOVA", "Spearman", "r2"),
  label.col = "label",
  idSpatial.col = "idSpatial",
  fdr.proc = c("BH", "Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BY", "ABH",
    "TSBH")
)

Arguments

Details

In case the nonparametric Kruskal-Wallis test is chosen, additional columns are provided testing each label for significantly larger scores (Kruskal-Wallis is global and only says whether one or several labels show a bias). Individual labels are tested with Wilcoxon and two columns are added *per* label, one for the statistics and one for a Bonferroni-corrected P-value over all the labels.

In case an actual statistical test is chosen, a parametric test (ANOVA) and a non-parametric test (Kruskal-Wallis) are available for the global analysis. Individual labels are tested with T-tests or Wilcoxon (Bonferroni-corrected) accordingly.

In case a statistics is preferred, Spearman correlation or explained variance (r2 or coefficient of determination, through linear models) are available. They mesure the relationship between each individual area and scores. For the explained variance, a global value (R2) is also computed from a multi-linear model (the same as what is used for the ANOVA).

Value

A data.frame with the names of the interactions, the value of the chosen statistics, and the corresponding Q-value.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(bsrinf.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")
thres <- 0.01
#bsrinf.red <- reduceToBestPathway(bsrinf.spa)
#s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
#scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

# Run in other examples no need to be run again
# spatialAssociation(scores.red[c(1:2),], areas = annotation.spa,
# label.col = "ground_truth")

Description

Plot a heatmap featuring Q-values or values of statistical association between L-R interaction score spatial distributions and tissue area labels.

Usage

spatialAssociationPlot(
  associations,
  qval.thres = 0.01,
  absval.thres = 0,
  colors = NULL
)

Arguments

Details

Display a heatmap linking L-R interactions to labels.

Value

ComplexHeatmap::Heatmap object

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(bsrinf.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

thres <- 0.01
bsrinf.red <- reduceToBestPathway(bsrinf.spa)
s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

# statistical association with tissue areas based on correlations

assoc.bsr.corr <- spatialAssociation(scores.red[c(1:10), ],
areas = annotation.spa, label.col = "ground_truth",test = "Spearman")

spatialAssociationPlot(assoc.bsr.corr)

Description

Use PCA or t-SNE to obtain a 2D-projection of a set of spatial scores or associations. This plot summarizes the diversity of patterns occuring in a spatial dataset. Use the function spatialIndexPlot to create a large visual index of many spatial distributions. Not limited to BulkSignalR gene signature scores.

Usage

spatialDiversityPlot(
  scores,
  associations,
  proj = c("PCA", "tSNE"),
  score.based = FALSE,
  qval.thres = 0.01,
  val.thres = 0,
  with.names = FALSE,
  text.fs = 2.5,
  legend.fs = 10,
  axis.fs = 10,
  label.fs = 12,
  dot.size = 1,
  perplexity = 10
)

Arguments

Details

Display a 2D-projection of the score spatial distributions.

Value

Display a 2D-projection of the score spatial distributions.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(bsrinf.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

thres <- 0.01
bsrinf.red <- reduceToBestPathway(bsrinf.spa)
s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

# statistical association with tissue areas based on correlations
# For display purpose, we only use a subset here


assoc.bsr.corr <- spatialAssociation(scores.red[c(1:3), ],
annotation.spa, label.col = "ground_truth",test = "Spearman")

spatialDiversityPlot(scores.red[c(1:3),],assoc.bsr.corr)

Description

Generate an index made of series of small individual spatial score plots in a PDF. Not limited to BulkSignalR gene signature scores.

Usage

spatialIndexPlot(
  scores,
  areas,
  out.file,
  ref.plot = TRUE,
  image.raster = NULL,
  x.col = "array_col",
  y.col = "array_row",
  label.col = "label",
  idSpatial.col = "idSpatial",
  cut.p = 0.01,
  low.color = "royalblue3",
  mid.color = "white",
  high.color = "orange",
  title.fs = 12,
  legend.fs = 10,
  axis.fs = 10,
  label.fs = 12,
  dot.size = 0.25,
  ratio = 1.25,
  base.v = 2.5,
  base.h = 3,
  ref.colors = NULL
)

Arguments

Details

A PDF file is created that contains the index.

Value

Create PDF file and returns 'NULL', invisibly.

Examples

data(bsrdm.spa, package = "BulkSignalR")
data(bsrinf.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

thres <- 0.01
bsrinf.red <- reduceToBestPathway(bsrinf.spa)
s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

# generate visual index on disk in pdf file
spatialIndexPlot(scores.red[1:2,], annotation.spa,  
label.col = "ground_truth",
out.file = "spatialIndexPlot")

Description

Generate a plot with scores at the spatial coordinates of the corresponding sample locations. Not limited to BulkSignalR gene signature scores.

Usage

spatialPlot(
  v,
  areas,
  inter.name,
  rev.y = TRUE,
  ref.plot = FALSE,
  ref.plot.only = FALSE,
  image.raster = NULL,
  x.col = "array_col",
  y.col = "array_row",
  label.col = "label",
  idSpatial.col = "idSpatial",
  cut.p = 0.01,
  low.color = "royalblue3",
  mid.color = "white",
  high.color = "orange",
  title.fs = 12,
  legend.fs = 10,
  axis.fs = 10,
  label.fs = 12,
  dot.size = 0.5,
  legend.dot.factor = 10,
  ref.colors = NULL
)

Arguments

Details

A single (scores) or side-by-side (reference tissue & scores) plot is generated.

Value

A spatial plot

Examples

data(bsrinf.spa, package = "BulkSignalR")
data(bsrdm.spa, package = "BulkSignalR")
data(annotation.spa, package = "BulkSignalR")

thres <- 0.01
bsrinf.red <- reduceToBestPathway(bsrinf.spa)
s.red  <- BSRSignature(bsrinf.red, qval.thres=thres)
scores.red <- scoreLRGeneSignatures(bsrdm.spa,s.red)

inter <- "{SLIT2} / {GPC1}"

spatialPlot(scores.red[inter, ], annotation.spa, inter,
   ref.plot = TRUE, ref.plot.only = FALSE,
   image.raster = NULL, dot.size = 1,
   label.col = "ground_truth")

Description

Generate a igraph object with one link between each cell type.

Usage

summarizedCellularNetwork(tab)

Arguments

Value

A igraph object containing a summary cellular network with edge weights proportional to the sum of individual link scores. Edge weight are normalized to a total of one.

Examples

data(bsrdm, package = "BulkSignalR")
data(bsrinf, package = "BulkSignalR")
data("tme.signatures", package = "BulkSignalR")
data(immune.signatures, package = "BulkSignalR")

immune.signatures <- immune.signatures[immune.signatures$signature %in%
    c("T cells"), ]

signatures <- rbind(immune.signatures, tme.signatures[
    tme.signatures$signature %in% c("Fibroblasts"),
])

tme.scores <- scoreSignatures(bsrdm, signatures)

# assignment
lr2ct <- assignCellTypesToInteractions(bsrdm, bsrinf, tme.scores)

# cellular network
g.table <- cellularNetworkTable(lr2ct[c(1:25),])
gSummary <- summarizedCellularNetwork(g.table)
# plot(gSummary, edge.width=1+30*E(gSummary)$score)

Description

Target gene correlations accessor

Target gene correlations accessor

Usage

## S4 method for signature 'BSRInference'
tgCorr(x)

## S4 method for signature 'BSRSignature'
tgCorr(x)

Arguments

Value

tgCorr

Examples

bsr.sig <- new("BSRSignature")
tgCorr(bsr.sig)

Description

Target gene expression accessor

Target gene expression accessor

Usage

## S4 method for signature 'BSRInferenceComp'
tgExpr(x)

## S4 method for signature 'BSRSignatureComp'
tgExpr(x)

Arguments

Value

tgExpr

tg.expr

Examples

bsrinf <- new("BSRInferenceComp")
tgExpr(bsrinf)

Description

Target genes accessor

Target genes accessor

Usage

## S4 method for signature 'BSRInference'
tgGenes(x)

## S4 method for signature 'BSRSignature'
tgGenes(x)

Arguments

Value

tgGenes

tgGenes

Examples

bsr.sig <- new("BSRSignature")
tgGenes(bsr.sig)

Description

Target gene logFC accessor

Target gene logFC accessor

Usage

## S4 method for signature 'BSRInferenceComp'
tgLogFC(x)

## S4 method for signature 'BSRSignatureComp'
tgLogFC(x)

Arguments

Value

tgLogFC

tg.logFC

Examples

bsrinf <- new("BSRInferenceComp")
tgLogFC(bsrinf)

Description

Target gene P-values accessor

Target gene P-values accessor

Usage

## S4 method for signature 'BSRInferenceComp'
tgPval(x)

## S4 method for signature 'BSRSignatureComp'
tgPval(x)

Arguments

Value

tgPval

tg.pval

Examples

bsrinf <- new("BSRInferenceComp")
tgPval(bsrinf)

Description

A dataset containing gene signatures for some immune and stromal cell populations that are present in the microenvironment of a tumor.

Usage

data(tme.signatures)

Format

A data frame with 209 rows and 2 variables:

gene

HUGO gene symbol

signature

cell population name

Source

Becht & al., Genome Biol, 2016; Angelova et al., Genome Biol, 2015.

Description

A method to update the data underlying statistical significance estimations prior to rescoring for an existing BSRInferenceComp object (P- and Q-value estimations as well as LR-score).

Usage

## S4 method for signature 'BSRInferenceComp'
updateInference(
  obj,
  bsrcc,
  ncounts,
  src.bsrcc = NULL,
  rank.p = 0.55,
  max.pval = 0.01,
  min.logFC = 1,
  min.LR.score = 0,
  neg.receptors = FALSE,
  pos.targets = FALSE,
  neg.targets = FALSE,
  min.t.logFC = 0.5,
  min.positive = 2,
  fdr.proc = c("BH", "Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BY", "ABH",
    "TSBH")
)

Arguments

Details

A BSRInferenceComp object should be created by calling "BSRInferenceComp"

Value

A BSRInferenceComp object. The main application of this method is to take a "universal" inference obtained by assigning each gene to good logFC, P-values and expression levels whose role is to find all the reachable targets per receptor/pathway, and to update it by using actual logFC, P-values, and expression data. The benefit is to save time when multiple sample comparisons are performed, only one network exploration is necessary. Note that if a restrictive logic such as positive.targets=TRUE is used, the result will be correct provided all the targets were in the initial BSRInferenceComp object. If a restriction on the targets was applied, then the update is likely to miss some targets, i.e., the statistical analysis will be wrong.

Note that correlations are set to 1 to avoid lengthy computations with scRNA-seq data and multiple cell populations.

The main function of this method is to support our SingleCellSignalR v2 package.

Examples

data(bsrdm.comp, package = "BulkSignalR")
data(bsrinf.comp, package = "BulkSignalR")
colA <- as.integer(1:2)
colB <- as.integer(3:4)

#bsrdm.comp <- as(bsrdm, "BSRDataModelComp")

n <- nrow(ncounts(bsrdm.comp))
stats <- data.frame(pval = runif(n),
logFC = rnorm(n, 0, 2),
expr = runif(n, 0, 10))
rownames(stats) <- rownames(ncounts(bsrdm.comp))


# update
stats$pval <- stats$pval / 100
stats$logFC <- stats$logFC + 0.5

bsrcc.2 <- BSRClusterComp(bsrdm.comp, colA, colB, stats)
bsrinf.updated <- updateInference(bsrinf.comp, bsrcc.2,
max.pval = 1, min.logFC = 0.1)