| Title: | Investigating regions of interest and performing cross-regional analysis with spatial transcriptomics data |
|---|---|
| Description: | This package analyze spatial transcriptomics data through cross-regional analysis. It selects regions of interest (ROIs) and identifys cross-regional cell type-specific differential signals. The ROIs can be selected using automatic algorithm or through manual selection. It facilitates manual selection of ROIs using a shiny application. |
| Authors: | Ziyi Li [aut, cre] |
| Maintainer: | Ziyi Li <[email protected]> |
| License: | GPL-3 |
| Version: | 1.3.0 |
| Built: | 2024-09-30 05:25:06 UTC |
| Source: | https://github.com/bioc/RegionalST |
Perform GSEA analysis for cross-regional DE genes
DoGSEA(considerRes, whichDB = "hallmark", gmtdir = NULL, withProp = FALSE)DoGSEA(considerRes, whichDB = "hallmark", gmtdir = NULL, withProp = FALSE)
considerRes |
A list of cross-regional DE genes. |
whichDB |
A character string to select the database names, e.g., "hallmark", "kegg", "reactome". |
gmtdir |
Directory for external database gmt file location. |
withProp |
Whether deconvolution proportion is used in previous steps. |
A list including GSEA results for all cell types.
data(exampleRes) allCTres <- DoGSEA(exampleRes, whichDB = "hallmark", withProp = TRUE)data(exampleRes) allCTres <- DoGSEA(exampleRes, whichDB = "hallmark", withProp = TRUE)
Draw dot plot for GSEA results of cross-regional DE genes
DrawDotplot( allCTres, CT = 1, angle = 20, vjust = 0.9, hjust = 1, padj_cutoff = 1, topN = 20, chooseP = "padj", eachN = NULL )DrawDotplot( allCTres, CT = 1, angle = 20, vjust = 0.9, hjust = 1, padj_cutoff = 1, topN = 20, chooseP = "padj", eachN = NULL )
allCTres |
A list of GSEA results for all cell types. |
CT |
A number of the interested cell type, e.g., 1, 2, 3. |
angle |
A number of plotting parameter, angle of the x axis label. |
vjust |
A number of vertical adjustment in plotting. |
hjust |
A number of horizontal adjustment in plotting. |
padj_cutoff |
A cutoff number of adjusted p value. |
topN |
A number of the plotted top pathways. |
chooseP |
A character string for the p value that used in plotting, e.g., "padj" or "pval". |
eachN |
The maximum number of pathways in each cell type. |
A plot object
data(exampleRes) allCTres <- DoGSEA(exampleRes, whichDB = "hallmark", withProp = TRUE) DrawDotplot(allCTres, CT = 1, angle = 15, vjust = 1, chooseP = "padj")data(exampleRes) allCTres <- DoGSEA(exampleRes, whichDB = "hallmark", withProp = TRUE) DrawDotplot(allCTres, CT = 1, angle = 15, vjust = 1, chooseP = "padj")
Draw regional cell type distribution with cell type annotation information
DrawRegionProportion(sce, label = "celltype", selCenter = seq_len(10))DrawRegionProportion(sce, label = "celltype", selCenter = seq_len(10))
sce |
A single cell experiment object. |
label |
A string character for the cell type variable. |
selCenter |
A vector of the interested ROIs, e.g., 1:4. |
A plot object.
data("example_sce") DrawRegionProportion(example_sce, label = "celltype", selCenter = 1:3)data("example_sce") DrawRegionProportion(example_sce, label = "celltype", selCenter = 1:3)
Draw regional cell type distribution with cellular proportion information
DrawRegionProportion_withProp( sce, label = "CARD_CellType", selCenter = seq_len(10) )DrawRegionProportion_withProp( sce, label = "CARD_CellType", selCenter = seq_len(10) )
sce |
A single cell experiment object. |
label |
A string character for the cell type variable. |
selCenter |
A vector of the interested ROIs, e.g., 1:4. |
A plot object.
data("example_sce") DrawRegionProportion_withProp(example_sce, label = "Proportions", selCenter = 1:3)data("example_sce") DrawRegionProportion_withProp(example_sce, label = "Proportions", selCenter = 1:3)
A simulated example input data file
data(example_sce)data(example_sce)
A SingleCellExperiment object.
A SingleCellExperiment object.
data(example_sce)data(example_sce)
A simulated example DE output file
data(exampleRes)data(exampleRes)
A list object.
A list object.
data(exampleRes)data(exampleRes)
Identify regional cells given centers and radiuses
FindRegionalCells( sce, centerID, enhanced = FALSE, radius = 10, avern = 5, doPlot = FALSE, returnPlot = FALSE )FindRegionalCells( sce, centerID, enhanced = FALSE, radius = 10, avern = 5, doPlot = FALSE, returnPlot = FALSE )
sce |
A single cell experiment object. |
centerID |
One or a vector of spot IDs as centers of ROIs. |
enhanced |
A logical variable for plotting enhanced plot or now. Default is FALSE. |
radius |
A number of fixed ROI radius. |
avern |
A number of the average sites used to compute unit distance, default is 5. |
doPlot |
A logical variable to specify whether plot the figure or not. |
returnPlot |
a logical variable to specify whether output the plot or not. |
A list including center spot ID and regional spot IDs.
# FindRegionalCells(sce, centerID = "ACGCCTGACACGCGCT-1")# FindRegionalCells(sce, centerID = "ACGCCTGACACGCGCT-1")
Identify cross-regional differential analysis
GetCrossRegionalDE_raw( sce, twoCenter = c(3, 4), enhanced = FALSE, label = "celltype", n_markers = 10, logfc.threshold = 0.25, angle = 30, hjust = 0, size = 3, min.pct = 0.1, padj_filter = 0.05, doHeatmap = TRUE )GetCrossRegionalDE_raw( sce, twoCenter = c(3, 4), enhanced = FALSE, label = "celltype", n_markers = 10, logfc.threshold = 0.25, angle = 30, hjust = 0, size = 3, min.pct = 0.1, padj_filter = 0.05, doHeatmap = TRUE )
sce |
A single cell experiment object. |
twoCenter |
A vector of two numbers for the interested ROI numbers. |
enhanced |
A logical variable for using enhanced data or not. |
label |
A variable name that contains the cell type information. |
n_markers |
A number specifying the top DE gene number. |
logfc.threshold |
A number for the cutoff threshold of log fold change. |
angle |
A number for angle when plotting. |
hjust |
A number for horizontal justification when plotting. |
size |
A number for text font size. |
min.pct |
A number of minimum percentage specified in the Seurat DE function. |
padj_filter |
A number for filtering adjusted p values. |
doHeatmap |
Logical variable for whether drawing the heatmap. |
A list including the top DE genes (topDE), and all DE genes (allDE).
data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") # I used a very big padj filter here because this is just a toy data GetCrossRegionalDE_raw(example_sce, twoCenter = c(1,2), min.pct = 0.01, logfc.threshold = 0.01, padj_filter = 0.5)data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") # I used a very big padj filter here because this is just a toy data GetCrossRegionalDE_raw(example_sce, twoCenter = c(1,2), min.pct = 0.01, logfc.threshold = 0.01, padj_filter = 0.5)
Identify cross-regional differential analysis with proportion
GetCrossRegionalDE_withProp( sce, twoCenter = c(3, 4), label = "celltype", n_markers = 10, angle = 30, hjust = 0, size = 3, padj_filter = 0.05, doHeatmap = TRUE )GetCrossRegionalDE_withProp( sce, twoCenter = c(3, 4), label = "celltype", n_markers = 10, angle = 30, hjust = 0, size = 3, padj_filter = 0.05, doHeatmap = TRUE )
sce |
A single cell experiment object. |
twoCenter |
A vector of two numbers for the interested ROI numbers. |
label |
A variable name that contains the cell type information. |
n_markers |
A number specifying the top DE gene number. |
angle |
A number for angle when plotting. |
hjust |
A number for horizontal justification when plotting. |
size |
A number for text font size. |
padj_filter |
A number for filtering adjusted p values. |
doHeatmap |
Logical variable for whether drawing the heatmap. |
A list including the top DE genes (topDE), and all DE genes (allDE).
data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") # Since the example data is very small, I set padj filter as NULL. Default is 0.05. GetCrossRegionalDE_withProp(example_sce, twoCenter = c(1,2), padj_filter = NULL)data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") # Since the example data is very small, I set padj filter as NULL. Default is 0.05. GetCrossRegionalDE_withProp(example_sce, twoCenter = c(1,2), padj_filter = NULL)
Computer the entropy for a fixed radius
GetOneRadiusEntropy( sce, selectN, enhanced = FALSE, weight = NULL, label = "celltype", radius = 10, doPlot = FALSE, mytitle = NULL )GetOneRadiusEntropy( sce, selectN, enhanced = FALSE, weight = NULL, label = "celltype", radius = 10, doPlot = FALSE, mytitle = NULL )
sce |
A single cell experiment object. |
selectN |
A total number for selected centers. Should be smaller than the total site number. |
enhanced |
A logical variable of whether using enhanced data. |
weight |
A data frame to specify the weights of all cell types. |
label |
A variable name that contains the cell type information. |
radius |
A number for fixed radius. |
doPlot |
Logical variable about whether draw the plot. |
mytitle |
A character string for the title of the plot. |
A list including the selected centers, computed entropies, radius.
data("example_sce") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- mySpatialPreprocess(example_sce, platform="Visium") GetOneRadiusEntropy(example_sce, selectN = round(length(example_sce$spot)/2), weight = weight, radius = 5, doPlot = TRUE, mytitle = "Radius 5 weighted entropy")data("example_sce") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- mySpatialPreprocess(example_sce, platform="Visium") GetOneRadiusEntropy(example_sce, selectN = round(length(example_sce$spot)/2), weight = weight, radius = 5, doPlot = TRUE, mytitle = "Radius 5 weighted entropy")
Computer the entropy for a fixed radius with cell type proportion
GetOneRadiusEntropy_withProp( sce, selectN, weight = NULL, label = "celltype", radius = 10, doPlot = FALSE, mytitle = NULL )GetOneRadiusEntropy_withProp( sce, selectN, weight = NULL, label = "celltype", radius = 10, doPlot = FALSE, mytitle = NULL )
sce |
A single cell experiment object. |
selectN |
A total number for selected centers. Should be smaller than the total site number. |
weight |
A data frame to specify the weights of all cell types. |
label |
A variable name that contains the cell type information. |
radius |
A number for fixed radius. |
doPlot |
Logical variable about whether draw the plot. |
mytitle |
A character string for the title of the plot. |
A list including the selected centers, computed entropies, radius.
data("example_sce") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- mySpatialPreprocess(example_sce, platform="Visium") GetOneRadiusEntropy_withProp(example_sce, selectN = round(length(example_sce$spot)/10), weight = weight, radius = 5, doPlot = TRUE, mytitle = "Radius 5 weighted entropy")data("example_sce") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- mySpatialPreprocess(example_sce, platform="Visium") GetOneRadiusEntropy_withProp(example_sce, selectN = round(length(example_sce$spot)/10), weight = weight, radius = 5, doPlot = TRUE, mytitle = "Radius 5 weighted entropy")
Define an accessor method for Proportion_CARD
getProportion(card)getProportion(card)
card |
A CARD object. |
A matrix containing the spot-level cell type proportion information
# getProportion(card)# getProportion(card)
Manually select top ROIs
ManualSelectCenter(sce)ManualSelectCenter(sce)
sce |
A single cell experiment object. |
An sce object with selected centers and radiuses.
data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") # I commented this out because the shiny app will get stuck without input. # example_sce <- ManualSelectCenter(example_sce)data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") # I commented this out because the shiny app will get stuck without input. # example_sce <- ManualSelectCenter(example_sce)
Perform Preprocessing for spatial data (tailored from BayesSpace function)
mySpatialPreprocess( sce, platform = c("Visium", "ST"), n.PCs = 15, n.HVGs = 2000, skip.PCA = FALSE, assay.type = "logcounts" )mySpatialPreprocess( sce, platform = c("Visium", "ST"), n.PCs = 15, n.HVGs = 2000, skip.PCA = FALSE, assay.type = "logcounts" )
sce |
A SingleCellExperiment object. |
platform |
Which platform the data are from, Visium or ST. |
n.PCs |
Number of PCs used in the analysis. |
n.HVGs |
Number of highly variable genes used in the analysis. |
skip.PCA |
A boolean variable to choose whether skipping the PCA step or not. |
assay.type |
Which assay to use, default is logcounts. |
A processed SingleCellExperiment object.
data(example_sce) example_sce <- mySpatialPreprocess(example_sce, platform="Visium")data(example_sce) example_sce <- mySpatialPreprocess(example_sce, platform="Visium")
Hallmark database downloaded from MSigDB (Feb, 2023)
data(pathways_hallmark)data(pathways_hallmark)
A list object.
A list object.
Liberzon et al. (2015) Cell Syst. 1(6):417-425 (PubMed)
data(pathways_hallmark)data(pathways_hallmark)
KEGG database downloaded from MSigDB (Feb, 2023)
data(pathways_kegg)data(pathways_kegg)
A list object.
A list object.
Kanehisa and Goto (2000) Nucleic Acids Research 28(1):27-30 (PubMed)
data(pathways_kegg)data(pathways_kegg)
REACTOME database downloaded from MSigDB (Feb, 2023)
data(pathways_reactome)data(pathways_reactome)
A list object.
A list object.
Jassal et al. (2020) Nucleic Acids Research 28(1):27-30 (PubMed)
data(pathways_reactome)data(pathways_reactome)
Plot one selected ROI
PlotOneSelectedCenter(sce, ploti, enhanced = FALSE)PlotOneSelectedCenter(sce, ploti, enhanced = FALSE)
sce |
A single cell experiment object. |
ploti |
A number of indicate which ROI to plot. |
enhanced |
A logical variable for using enhanced data or not. |
A figure object for the selected ROI.
data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") PlotOneSelectedCenter(example_sce, ploti = 1)data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") PlotOneSelectedCenter(example_sce, ploti = 1)
Automatically rank ROI centers based on entropy
RankCenterByEntropy( sce, weight, enhanced = FALSE, selectN = round(length(sce$spot)/10), label = "celltype", topN = 10, min_radius = 10, avern = 5, radius_vec = c(10, 15, 20), doPlot = TRUE )RankCenterByEntropy( sce, weight, enhanced = FALSE, selectN = round(length(sce$spot)/10), label = "celltype", topN = 10, min_radius = 10, avern = 5, radius_vec = c(10, 15, 20), doPlot = TRUE )
sce |
A single cell experiment object. |
weight |
A data frame to specify the weights of all cell types. |
enhanced |
A logical variable of whether using enhanced data. |
selectN |
A total number for selected centers. Should be smaller than the total site number. |
label |
A variable name that contains the cell type information. |
topN |
A number to specify the total amount of top ranked ROIs. |
min_radius |
The minimum repellent radius. |
avern |
A number of the average sites used to compute unit distance, default is 5. |
radius_vec |
A vector of numbers for candidate radiuses. |
doPlot |
Logical variable about whether draw the plot. |
An sce object with selected ROI information.
data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- RankCenterByEntropy(example_sce, weight, label = "celltype", selectN = round(length(example_sce$spot)/10), topN = 3, min_radius = 10, radius_vec = c(10,15), doPlot = TRUE)data("example_sce") example_sce <- mySpatialPreprocess(example_sce, platform="Visium") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- RankCenterByEntropy(example_sce, weight, label = "celltype", selectN = round(length(example_sce$spot)/10), topN = 3, min_radius = 10, radius_vec = c(10,15), doPlot = TRUE)
Automatically rank ROI centers based on entropy with proportions
RankCenterByEntropy_withProp( sce, weight, selectN = round(length(sce$spot)/10), topN = 10, min_radius = 10, avern = 5, radius_vec = c(10, 15, 20), doPlot = TRUE )RankCenterByEntropy_withProp( sce, weight, selectN = round(length(sce$spot)/10), topN = 10, min_radius = 10, avern = 5, radius_vec = c(10, 15, 20), doPlot = TRUE )
sce |
A single cell experiment object. |
weight |
A data frame to specify the weights of all cell types. |
selectN |
A total number for selected centers. Should be smaller than the total site number. |
topN |
A number to specify the total amount of top ranked ROIs. |
min_radius |
The minimum repellent radius. |
avern |
A number of the average sites used to compute unit distance, default is 5. |
radius_vec |
A vector of numbers for candidate radiuses. |
doPlot |
Logical variable about whether draw the plot. |
An sce object with selected ROI information.
data("example_sce") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- mySpatialPreprocess(example_sce, platform="Visium") ## I set our min_raius as 10 and radius vector as 10 and 15 as the example dataset is very small example_sce <- RankCenterByEntropy_withProp(example_sce, weight, selectN = round(length(example_sce$spot)/10), topN = 3, min_radius = 10, radius_vec = c(10,15), doPlot = TRUE)data("example_sce") weight <- data.frame(celltype = c("Cancer Epithelial", "CAFs", "T-cells", "Endothelial", "PVL", "Myeloid", "B-cells", "Normal Epithelial", "Plasmablasts"), weight = c(0.25,0.05, 0.25,0.05, 0.025,0.05, 0.25,0.05,0.025)) example_sce <- mySpatialPreprocess(example_sce, platform="Visium") ## I set our min_raius as 10 and radius vector as 10 and 15 as the example dataset is very small example_sce <- RankCenterByEntropy_withProp(example_sce, weight, selectN = round(length(example_sce$spot)/10), topN = 3, min_radius = 10, radius_vec = c(10,15), doPlot = TRUE)