Identifies cell-to-cell Interaction Changed Features (ICF) for
spots, i.e. features expression residual that are different due to proximity
to other cell types. Works using results from celltype deconvolution methods
such as those from runDWLSDeconv().
findICFSpot(
gobject,
spat_unit = NULL,
feat_type = NULL,
expression_values = c("normalized", "scaled", "custom"),
ave_celltype_exp,
selected_features = NULL,
spatial_network_name = "Delaunay_network",
deconv_name = "DWLS",
minimum_unique_cells = 5,
minimum_unique_int_cells = 5,
CCI_cell_score = 0.1,
dwls_cutoff = 0.001,
diff_test = "permutation",
nr_permutations = 100,
adjust_method = "fdr",
do_parallel = TRUE,
cores = NA,
set_seed = TRUE,
seed_number = 1234,
verbose = FALSE
)A giotto object
spatial unit (e.g. 'cell')
feature type (e.g. 'rna')
expression values to use
average feature expression in each cell type
subset of selected features (optional)
name of spatial network to use
name of deconvolution/spatial enrichment values to use
minimum number of target cells required
minimum number of interacting cells required
cell proximity score to filter no interacted cell
cell type proportion cutoff to label the cell
which differential expression test
number of permutations if diff_test = permutation
which method to adjust p-values
run calculations in parallel with mclapply
number of cores to use if do_parallel = TRUE
set a seed for reproducibility
seed number
be verbose
icfObject that contains the differential feat scores
Function to calculate if features expression residual are
differentially expressed in cell types when they interact
(approximated by physical proximity) with other cell types.
Feature expression residual calculated as:
(observed expression in spot - cell_type_proportion *
average_expressed_in_cell_type)
The results data.table in the icfObject contains - at least -
the following columns:
features: All or selected list of tested features
sel: average feature expression residual in the interacting cells from the target cell type
other: average feature expression residual in the NOT-interacting cells from the target cell type
pcc_sel: correlation between cell proximity score and expression residual in the interacting cells from the target cell type
pcc_other: correlation between cell proximity score and expression residual in the NOT-interacting cells from the target cell type
pcc_diff: correlation difference between sel and other
p.value: associated p-value
p.adj: adjusted p-value
cell_type: target cell type
int_cell_type: interacting cell type
nr_select: number of cells for selected target cell type
int_nr_select: number of cells for interacting cell type
unif_int: cell-cell interaction
g <- GiottoData::loadGiottoMini("visium")
#> 1. read Giotto object
#> 2. read Giotto feature information
#> 3. read Giotto spatial information
#> 3.1 read Giotto spatial shape information
#> 3.2 read Giotto spatial centroid information
#> 3.3 read Giotto spatial overlap information
#> 4. read Giotto image information
#>
#> checking default envname 'giotto_env'
#> a system default python environment was found
#> Using python path:
#> "/usr/bin/python3"
x <- findMarkers_one_vs_all(g,
cluster_column = "leiden_clus", min_feats = 20
)
#> using 'Scran' to detect marker feats. If used in published
#> research, please cite: Lun ATL, McCarthy DJ, Marioni JC (2016).
#> 'A step-by-step workflow for low-level analysis of single-cell RNA-seq
#> data with Bioconductor.'
#> F1000Res., 5, 2122. doi: 10.12688/f1000research.9501.2.
#> start with cluster 1start with cluster 2start with cluster 3start with cluster 4start with cluster 5start with cluster 6start with cluster 7
sign_gene <- x$feats
sign_matrix <- matrix(rnorm(length(sign_gene) * 8, mean = 10),
nrow = length(sign_gene)
)
rownames(sign_matrix) <- sign_gene
colnames(sign_matrix) <- paste0("celltype_", unique(x$cluster))
#> Error in dimnames(x) <- dn: length of 'dimnames' [2] not equal to array extent
g <- runDWLSDeconv(gobject = g, sign_matrix = sign_matrix)
#> Error: package 'quadprog' is not yet installed
#>
#> To install:
#> install.packages(c("quadprog"))
ave_celltype_exp <- calculateMetaTable(g, metadata_cols = "leiden_clus")
ave_celltype_exp <- reshape2::dcast(ave_celltype_exp, variable~leiden_clus)
rownames(ave_celltype_exp) <- ave_celltype_exp$variable
ave_celltype_exp <- ave_celltype_exp[,-1]
colnames(ave_celltype_exp) <- colnames(sign_matrix)
findICFSpot(g,
spat_unit = "cell",
feat_type = "rna",
ave_celltype_exp = ave_celltype_exp,
spatial_network_name = "spatial_network"
)
#> Error in featExpDWLS(gobject = gobject, spat_unit = spat_unit, feat_type = feat_type, ave_celltype_exp = ave_celltype_exp): ncol(ave_celltype_exp) needs to be the same as
#> ncol(dwls_values) - 1