tests/testthat/test_aggregateNeighbors.R
In BodenmillerGroup/imcRtools: Methods for imaging mass cytometry data analysis
test_that("aggregateNeighbors function works", {
library(cytomapper)
library(dplyr)
data("pancreasSCE")
cur_sce1 <- pancreasSCE[,pancreasSCE$ImageNb == 1]
cur_sce2 <- pancreasSCE[,pancreasSCE$ImageNb == 2]
cur_sce3 <- pancreasSCE[,pancreasSCE$ImageNb == 3]
cur_sce1$Pos_X <- cur_sce1$Pos_X - min(cur_sce1$Pos_X)
cur_sce1$Pos_Y <- cur_sce1$Pos_Y - min(cur_sce1$Pos_Y)
cur_sce2$Pos_X <- cur_sce2$Pos_X - min(cur_sce2$Pos_X)
cur_sce2$Pos_Y <- cur_sce2$Pos_Y - min(cur_sce2$Pos_Y)
cur_sce3$Pos_X <- cur_sce3$Pos_X - min(cur_sce3$Pos_X)
cur_sce3$Pos_Y <- cur_sce3$Pos_Y - min(cur_sce3$Pos_Y)
pancreasSCE <- cbind(cur_sce1, cur_sce2, cur_sce3)
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10")
# Works metadata version on knn
# Directed
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
cur_dat <- cur_dat / rowSums(cur_dat)
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 10))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 215,]
cur_sce2 <- cur_sce2[,c(215, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 215])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 10,]
cur_sce2 <- cur_sce2[,c(10, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 10])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 322,]
cur_sce2 <- cur_sce2[,c(322, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 322])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
# Undirected
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
directed = FALSE,
name = "knn_10")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
cur_dat <- cur_dat / rowSums(cur_dat)
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "knn_10"))))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 215,]
cur_sce2 <- cur_sce2[,c(215, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 215])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 10,]
cur_sce2 <- cur_sce2[,c(10, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 10])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 322,]
cur_sce2 <- cur_sce2[,c(322, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 322])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
# Works expression version on knn
# Directed
# Mean
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Median
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Undirected
## Mean
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10", directed = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Median
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10",
directed = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Expansion
## metadata
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 5,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors[cur_dat[,"from"],],
DataFrame(cur_dat[,-1]))
expect_true(all(as.matrix(cur_sce$aggregatedNeighbors[-cur_dat[,"from"],]) == 0))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "exp_20"))))
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 10,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors[cur_dat[,"from"],],
DataFrame(cur_dat[,-1]))
expect_true(all(as.matrix(cur_sce$aggregatedNeighbors[-cur_dat[,"from"],]) == 0))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "exp_20"))))
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 15,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors[cur_dat[,"from"],],
DataFrame(cur_dat[,-1]))
expect_true(all(as.matrix(cur_sce$aggregatedNeighbors[-cur_dat[,"from"],]) == 0))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "exp_20"))))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "Pattern"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$Pattern <- factor(colData(cur_sce)$Pattern)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(Pattern, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "Pattern", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(as.matrix(cur_sce$aggregatedNeighbors),
as.matrix(DataFrame(cur_dat[,-1])), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "Area"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$Area <- factor(colData(cur_sce)$Area)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(Area, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "Area", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(as.matrix(cur_sce$aggregatedNeighbors),
as.matrix(DataFrame(cur_dat[,-1])), check.attributes = FALSE)
## expression
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 5,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression[unique(from(colPair(pancreasSCE,"exp_20"))),]),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$mean_aggregatedExpression[-unique(from(colPair(pancreasSCE,"exp_20"))),]))))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "counts"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression[from(colPair(pancreasSCE,"exp_20")),]),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$mean_aggregatedExpression[-from(colPair(pancreasSCE,"exp_20")),]))))
# Median
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression[unique(from(colPair(pancreasSCE,"exp_20"))),]),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$median_aggregatedExpression[-unique(from(colPair(pancreasSCE,"exp_20"))),]))))
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 15,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression[unique(from(colPair(pancreasSCE,"exp_20"))),]),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$mean_aggregatedExpression[-unique(from(colPair(pancreasSCE,"exp_20"))),]))))
# Delaunay
## metadata
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "delaunay")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat[,-1]))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "delaunay_interaction_graph"))))
cur_sce2 <- cur_sce
colPair(cur_sce2, "delaunay_interaction_graph") <- colPair(cur_sce2, "delaunay_interaction_graph")[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 215,]
cur_sce2 <- cur_sce2[,c(215, to(colPair(cur_sce2, "delaunay_interaction_graph"))[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 215])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "delaunay_interaction_graph")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "delaunay_interaction_graph") <- colPair(cur_sce2, "delaunay_interaction_graph")[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 12,]
cur_sce2 <- cur_sce2[,c(12, to(colPair(cur_sce2, "delaunay_interaction_graph"))[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 12])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "delaunay_interaction_graph")
cur_sce2$aggregatedNeighbors
## expression
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
subset_row = c("H3", "PIN")))
expect_equal(colnames(cur_sce$mean_aggregatedExpression), c("H3", "PIN"))
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs",
subset.row = c("H3", "PIN"))
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
subset_row = c(1,4)))
expect_equal(colnames(cur_sce$mean_aggregatedExpression), c("H3", "CD8a"))
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs",
subset.row = c(1,4))
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
subset_row = !(rownames(pancreasSCE) %in% c("H3", "PIN"))))
expect_equal(colnames(cur_sce$mean_aggregatedExpression), c("CD99", "CD8a", "CDH"))
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs",
subset.row = !(rownames(pancreasSCE) %in% c("H3", "PIN")))
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
# Error
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 1,
name = "exp_20")
expect_error(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"),
regexp = "No interactions found.",
fixed = TRUE)
expect_error(aggregateNeighbors("test"),
regexp = "'object' not of type 'SingleCellExperiment'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE),
regexp = "argument \"colPairName\" is missing, with no default",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = c("test", 1)),
regexp = "'colPairName' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "test"),
regexp = "'colPairName' not in 'colPairNames(object)'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata"),
regexp = "Provide a 'colData(object)' entry to aggregate by.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = 1),
regexp = "'count_by' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = c("CellType", "test")),
regexp = "'count_by' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = "test"),
regexp = "'count_by' is not a valid enty of 'colData(object)'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = "CellType",
proportions = "test"),
regexp = "'proportions' must be a single logical",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression"),
regexp = "'assay_type' not provided",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = 1),
regexp = "'assay_type' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "test"),
regexp = "'assay_type' not an assay in the 'object'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "counts",
subset_row = c(1, "test")),
regexp = "'subset_row' not in rownames(object).",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "counts",
subset_row = TRUE),
regexp = "'subset_row' logical entries must be as long as 'nrow(object)'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "counts",
name = TRUE),
regexp = "'name' must be a single string.",
fixed = TRUE)
})
BodenmillerGroup/imcRtools documentation built on July 1, 2024, 5:15 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
test_that("aggregateNeighbors function works", {
library(cytomapper)
library(dplyr)
data("pancreasSCE")
cur_sce1 <- pancreasSCE[,pancreasSCE$ImageNb == 1]
cur_sce2 <- pancreasSCE[,pancreasSCE$ImageNb == 2]
cur_sce3 <- pancreasSCE[,pancreasSCE$ImageNb == 3]
cur_sce1$Pos_X <- cur_sce1$Pos_X - min(cur_sce1$Pos_X)
cur_sce1$Pos_Y <- cur_sce1$Pos_Y - min(cur_sce1$Pos_Y)
cur_sce2$Pos_X <- cur_sce2$Pos_X - min(cur_sce2$Pos_X)
cur_sce2$Pos_Y <- cur_sce2$Pos_Y - min(cur_sce2$Pos_Y)
cur_sce3$Pos_X <- cur_sce3$Pos_X - min(cur_sce3$Pos_X)
cur_sce3$Pos_Y <- cur_sce3$Pos_Y - min(cur_sce3$Pos_Y)
pancreasSCE <- cbind(cur_sce1, cur_sce2, cur_sce3)
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10")
# Works metadata version on knn
# Directed
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
cur_dat <- cur_dat / rowSums(cur_dat)
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 10))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 215,]
cur_sce2 <- cur_sce2[,c(215, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 215])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 10,]
cur_sce2 <- cur_sce2[,c(10, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 10])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 322,]
cur_sce2 <- cur_sce2[,c(322, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 322])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
# Undirected
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
directed = FALSE,
name = "knn_10")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
cur_dat <- cur_dat / rowSums(cur_dat)
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "knn_10"))))
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"knn_10"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType[cur_dat$to])
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>%
select(-from) %>% as.matrix()
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat))
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 215,]
cur_sce2 <- cur_sce2[,c(215, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 215])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 10,]
cur_sce2 <- cur_sce2[,c(10, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 10])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "knn_10") <- colPair(cur_sce2, "knn_10")[from(colPair(cur_sce2, "knn_10")) == 322,]
cur_sce2 <- cur_sce2[,c(322, to(colPair(cur_sce2, "knn_10"))[from(colPair(cur_sce2, "knn_10")) == 322])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "knn_10")
cur_sce2$aggregatedNeighbors
# Works expression version on knn
# Directed
# Mean
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Median
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Undirected
## Mean
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10", directed = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Median
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "knn",k = 10,
name = "knn_10",
directed = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "knn_10",
aggregate_by = "expression",
assay_type = "counts",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"knn_10"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"knn_10"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
# Expansion
## metadata
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 5,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors[cur_dat[,"from"],],
DataFrame(cur_dat[,-1]))
expect_true(all(as.matrix(cur_sce$aggregatedNeighbors[-cur_dat[,"from"],]) == 0))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "exp_20"))))
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 10,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors[cur_dat[,"from"],],
DataFrame(cur_dat[,-1]))
expect_true(all(as.matrix(cur_sce$aggregatedNeighbors[-cur_dat[,"from"],]) == 0))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "exp_20"))))
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 15,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors[cur_dat[,"from"],],
DataFrame(cur_dat[,-1]))
expect_true(all(as.matrix(cur_sce$aggregatedNeighbors[-cur_dat[,"from"],]) == 0))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "exp_20"))))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "Pattern"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$Pattern <- factor(colData(cur_sce)$Pattern)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(Pattern, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "Pattern", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(as.matrix(cur_sce$aggregatedNeighbors),
as.matrix(DataFrame(cur_dat[,-1])), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "Area"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == 1))
cur_dat <- data.frame(colPair(pancreasSCE,"exp_20"))
cur_dat$Area <- factor(colData(cur_sce)$Area)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(Area, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "Area", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(as.matrix(cur_sce$aggregatedNeighbors),
as.matrix(DataFrame(cur_dat[,-1])), check.attributes = FALSE)
## expression
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 5,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression[unique(from(colPair(pancreasSCE,"exp_20"))),]),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$mean_aggregatedExpression[-unique(from(colPair(pancreasSCE,"exp_20"))),]))))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "counts"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "counts")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression[from(colPair(pancreasSCE,"exp_20")),]),
t(as.matrix(assay(cur_sce_2, "counts"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$mean_aggregatedExpression[-from(colPair(pancreasSCE,"exp_20")),]))))
# Median
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression[unique(from(colPair(pancreasSCE,"exp_20"))),]),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$median_aggregatedExpression[-unique(from(colPair(pancreasSCE,"exp_20"))),]))))
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 15,
name = "exp_20")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"exp_20"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"exp_20"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression[unique(from(colPair(pancreasSCE,"exp_20"))),]),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_true(all(is.na(as.matrix(cur_sce$mean_aggregatedExpression[-unique(from(colPair(pancreasSCE,"exp_20"))),]))))
# Delaunay
## metadata
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "delaunay")
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "metadata",
count_by = "CellType"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
# check for correct results of neighboring metadata
cur_dat <- data.frame(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_dat$celltype <- factor(colData(cur_sce)$CellType)[cur_dat$to]
cur_dat <- cur_dat %>% group_by(from) %>% count(celltype, .drop = FALSE) %>%
tidyr::pivot_wider(names_from = "celltype", values_from = "n") %>% ungroup() %>% as.matrix()
cur_dat[,-1] <- cur_dat[,-1] / rowSums(cur_dat[,-1])
expect_equal(cur_sce$aggregatedNeighbors,
DataFrame(cur_dat[,-1]))
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "metadata",
count_by = "CellType",
proportions = FALSE))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "aggregatedNeighbors"))
expect_s4_class(cur_sce$aggregatedNeighbors, "DataFrame")
expect_true(all(rowSums(as.matrix(cur_sce$aggregatedNeighbors)) == countLnodeHits(colPair(cur_sce, "delaunay_interaction_graph"))))
cur_sce2 <- cur_sce
colPair(cur_sce2, "delaunay_interaction_graph") <- colPair(cur_sce2, "delaunay_interaction_graph")[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 215,]
cur_sce2 <- cur_sce2[,c(215, to(colPair(cur_sce2, "delaunay_interaction_graph"))[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 215])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "delaunay_interaction_graph")
cur_sce2$aggregatedNeighbors
cur_sce2 <- cur_sce
colPair(cur_sce2, "delaunay_interaction_graph") <- colPair(cur_sce2, "delaunay_interaction_graph")[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 12,]
cur_sce2 <- cur_sce2[,c(12, to(colPair(cur_sce2, "delaunay_interaction_graph"))[from(colPair(cur_sce2, "delaunay_interaction_graph")) == 12])]
plotSpatial(cur_sce2, img_id = "ImageNb", node_color_by = "CellType", draw_edges = TRUE,
colPairName = "delaunay_interaction_graph")
cur_sce2$aggregatedNeighbors
## expression
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "mean_aggregatedExpression"))
expect_s4_class(cur_sce$mean_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$mean_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
statistic = "median"))
expect_s4_class(cur_sce , class = "SingleCellExperiment")
expect_equal(names(colData(cur_sce)),
c("ImageName", "Pos_X", "Pos_Y", "Area", "CellType", "ImageNb",
"CellNb", "MaskName", "Pattern", "median_aggregatedExpression"))
expect_s4_class(cur_sce$median_aggregatedExpression, "DataFrame")
expect_equal(dim(t(cur_sce$median_aggregatedExpression)), dim(cur_sce))
# check for correct results of neighboring metadata
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "median",
use.assay.type = "exprs")
expect_equal(as.matrix(cur_sce$median_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
subset_row = c("H3", "PIN")))
expect_equal(colnames(cur_sce$mean_aggregatedExpression), c("H3", "PIN"))
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs",
subset.row = c("H3", "PIN"))
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
subset_row = c(1,4)))
expect_equal(colnames(cur_sce$mean_aggregatedExpression), c("H3", "CD8a"))
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs",
subset.row = c(1,4))
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
expect_silent(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "delaunay_interaction_graph",
aggregate_by = "expression",
assay_type = "exprs",
subset_row = !(rownames(pancreasSCE) %in% c("H3", "PIN"))))
expect_equal(colnames(cur_sce$mean_aggregatedExpression), c("CD99", "CD8a", "CDH"))
cur_sce_2 <- cur_sce[,to(colPair(pancreasSCE,"delaunay_interaction_graph"))]
colData(cur_sce_2)$from <- from(colPair(pancreasSCE,"delaunay_interaction_graph"))
cur_sce_2 <- aggregateAcrossCells(cur_sce_2, ids = cur_sce_2$from, statistics = "mean",
use.assay.type = "exprs",
subset.row = !(rownames(pancreasSCE) %in% c("H3", "PIN")))
expect_equal(as.matrix(cur_sce$mean_aggregatedExpression),
t(as.matrix(assay(cur_sce_2, "exprs"))), check.attributes = FALSE)
# Error
pancreasSCE <- buildSpatialGraph(object = pancreasSCE,
img_id = "ImageNb",
type = "expansion",
threshold = 1,
name = "exp_20")
expect_error(cur_sce <- aggregateNeighbors(object = pancreasSCE,
colPairName = "exp_20",
aggregate_by = "metadata",
count_by = "CellType"),
regexp = "No interactions found.",
fixed = TRUE)
expect_error(aggregateNeighbors("test"),
regexp = "'object' not of type 'SingleCellExperiment'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE),
regexp = "argument \"colPairName\" is missing, with no default",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = c("test", 1)),
regexp = "'colPairName' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "test"),
regexp = "'colPairName' not in 'colPairNames(object)'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata"),
regexp = "Provide a 'colData(object)' entry to aggregate by.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = 1),
regexp = "'count_by' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = c("CellType", "test")),
regexp = "'count_by' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = "test"),
regexp = "'count_by' is not a valid enty of 'colData(object)'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "metadata", count_by = "CellType",
proportions = "test"),
regexp = "'proportions' must be a single logical",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression"),
regexp = "'assay_type' not provided",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = 1),
regexp = "'assay_type' must be a single string.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "test"),
regexp = "'assay_type' not an assay in the 'object'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "counts",
subset_row = c(1, "test")),
regexp = "'subset_row' not in rownames(object).",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "counts",
subset_row = TRUE),
regexp = "'subset_row' logical entries must be as long as 'nrow(object)'.",
fixed = TRUE)
expect_error(aggregateNeighbors(object = pancreasSCE, colPairName = "knn_10",
aggregate_by = "expression", assay_type = "counts",
name = TRUE),
regexp = "'name' must be a single string.",
fixed = TRUE)
})
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.