Skip to content

Commit

Permalink
update vignette
Browse files Browse the repository at this point in the history
  • Loading branch information
@MelodyJIN-Y
MelodyJIN-Y committed Nov 9, 2024
1 parent 2cc4912 commit b501c0b
Showing 1 changed file with 153 additions and 153 deletions.
306 changes: 153 additions & 153 deletions vignettes/jazzPanda.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -169,6 +169,159 @@ The required input data structure for the main function `lasso_markers()`
is a list of matrices. We will illustrate how to simply create the required
input data from the raw output of different technologies in the following
examples:
If you have a Seurat object, you can build the required object as follows.
The defined example_vectors_cm/example_vectors_tr can be passed
to `lasso_markers` to identify marker genes.

### from a SpatialExperiment object
If you have a SpatialExperiment or SpatialFeatureExperiment object,
you can generate the required input for creating the spatial vectors by
calling `convert_data` function.
If the transcript coordinates are available, you can
use either transcript coordinates or the count matrix to define spatial vectors
for genes.
The defined example_vectors_cm/example_vectors_tr can be passed
to `lasso_markers` to identify marker genes.
```{r, eval=FALSE}
library(SpatialFeatureExperiment)
library(SingleCellExperiment)
library(TENxXeniumData)
library(ExperimentHub)
eh <- ExperimentHub()
q <- query(eh, "TENxXenium")
spe_example <- q[["EH8547"]]
# get the required input list
example_lst <- convert_data(x=spe_example, sample_names = "sample01")
# -----------------------------------------------------------------------------
# Example clustering
# get the count matrix
cm <- spe_example@assays@data$counts
all_genes = row.names(cm)
example_seu <- CreateSeuratObject(counts = cm,
min.cells = 2, min.features = 1)
example_seu <- NormalizeData(example_seu,
normalization.method = "LogNormalize")
all.genes <- rownames(example_seu)
example_seu <- ScaleData(example_seu, features = all.genes)
example_seu <- RunPCA(example_seu, features = all.genes)
ElbowPlot(example_seu)
example_seu <- FindNeighbors(example_seu, dims = 1:15)
example_seu <- FindClusters(example_seu, resolution = 0.1)
seu_clusters <- as.data.frame(example_seu$seurat_clusters,nm="cluster")
seu_clusters$cell_id <- colnames(example_seu)
# -----------------------------------------------------------------------------
# combine cluster labels and the coordiantes
# make sure the cluster information contains column names:
# cluster, x, y, sample and cell_id
clusters_info <- as.data.frame(spe_example@int_colData$spatialCoords)
colnames(clusters_info) <- c("x","y")
clusters_info$cell_id <- row.names(clusters_info)
clusters_info$sample <- spe_example$sample_id
clusters_info <- merge(clusters_info, seu_clusters, by="cell_id")
w_x <- c(floor(min(clusters_info$x,
example_lst$trans_lst$sample01$x)),
ceiling(max(clusters_info$x,
example_lst$trans_lst$sample01$x)))
w_y <- c(floor(min(clusters_info$y,
example_lst$trans_lst$sample01$y)),
ceiling(max(clusters_info$y,
example_lst$trans_lst$sample01$y)))
# -----------------------------------------------------------------------------
# build spatial vectors from count matrix and cluster coordinates
example_vectors_cm <- get_vectors(trans_lst= NULL,
cm_lst=example_lst$cm_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
# -----------------------------------------------------------------------------
# build spatial vectors from transcript coordinates and cluster coordinates
example_vectors_tr <- get_vectors(trans_lst= example_lst$trans_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
```

### from a SpatialFeatureExperiment object

```{r, eval=FALSE}
sfe_example <- toSpatialFeatureExperiment(spe_example)
example_sfe_lst <- convert_data(x=sfe_example,sample_names = "sample01")
# -----------------------------------------------------------------------------
# build spatial vectors from count matrix and cluster coordinates
# make sure the cluster information contains column names:
# cluster, x, y, sample and cell_id
example_vectors_cm <- get_vectors(trans_lst= NULL,
cm_lst=example_sfe_lst$cm_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
# -----------------------------------------------------------------------------
# build spatial vectors from transcript coordinates and cluster coordinates
example_vectors_tr <- get_vectors(trans_lst= example_sfe_lst$trans_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
```
### from a Seurat object
```{r eval=FALSE}
# cm is the count matrix
seu_obj =Seurat::CreateSeuratObject(counts = cm)
# make sure the clusters information contains column names:
# cluster, x, y and sample
clusters_info = rep1_clusters
# make sure the transcript information contains column names:
# feature_name, x, y
transcript_coords = rep1_sub$trans_info
data_example = list(cm=seu_obj@assays$RNA$counts,
trans_info =transcript_coords)
w_x = c(min(floor(min(data$x)), floor(min(clusters_info$x))),
max(ceiling(max(data$x)), ceiling(max(clusters_info$x))))
w_y = c(min(floor(min(data$y)), floor(min(clusters_info$y))),
max(ceiling(max(data$y)), ceiling(max(clusters_info$y))))
# build spatial vectors from transcript coordinates and cluster coordinates
example_vectors_tr = get_vectors(trans_lst= list("rep1"=transcript_coords),
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = row.names(cm),
w_x=w_x, w_y=w_y)
# -----------------------------------------------------------------------------
# build spatial vectors from count matrix and cluster coordinates
# make sure the cluster information contains column names:
# cluster, x, y, sample and cell_id
colnames(clusters_info)[5] = "cell_id"
example_vectors_cm = get_vectors(trans_lst= NULL,
cm_lst=list(rep1=seu_obj@assays$RNA$counts),
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = row.names(cm),
w_x=w_x, w_y=w_y)
```

### From direct platform output
#### 10x Xenium data
The constructed xenium_data can be used in function `get_vectors()`.
Expand Down Expand Up @@ -305,159 +458,6 @@ nc_lst = list("sample1"=nc_coords[,kpt_cols])
```

If you have a Seurat object, you can build the required object as follows.
The defined example_vectors_cm/example_vectors_tr can be passed
to `lasso_markers` to identify marker genes.

### from a Seurat object
```{r eval=FALSE}
# cm is the count matrix
seu_obj =Seurat::CreateSeuratObject(counts = cm)
# make sure the clusters information contains column names:
# cluster, x, y and sample
clusters_info = rep1_clusters
# make sure the transcript information contains column names:
# feature_name, x, y
transcript_coords = rep1_sub$trans_info
data_example = list(cm=seu_obj@assays$RNA$counts,
trans_info =transcript_coords)
w_x = c(min(floor(min(data$x)), floor(min(clusters_info$x))),
max(ceiling(max(data$x)), ceiling(max(clusters_info$x))))
w_y = c(min(floor(min(data$y)), floor(min(clusters_info$y))),
max(ceiling(max(data$y)), ceiling(max(clusters_info$y))))
# build spatial vectors from transcript coordinates and cluster coordinates
example_vectors_tr = get_vectors(trans_lst= list("rep1"=transcript_coords),
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = row.names(cm),
w_x=w_x, w_y=w_y)
# -----------------------------------------------------------------------------
# build spatial vectors from count matrix and cluster coordinates
# make sure the cluster information contains column names:
# cluster, x, y, sample and cell_id
colnames(clusters_info)[5] = "cell_id"
example_vectors_cm = get_vectors(trans_lst= NULL,
cm_lst=list(rep1=seu_obj@assays$RNA$counts),
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = row.names(cm),
w_x=w_x, w_y=w_y)
```

### from a SpatialExperiment object
If you have a SpatialExperiment or SpatialFeatureExperiment object,
you can generate the required input for creating the spatial vectors by
calling `convert_data` function.
If the transcript coordinates are available, you can
use either transcript coordinates or the count matrix to define spatial vectors
for genes.
The defined example_vectors_cm/example_vectors_tr can be passed
to `lasso_markers` to identify marker genes.
```{r, eval=FALSE}
library(SpatialFeatureExperiment)
library(SingleCellExperiment)
library(TENxXeniumData)
library(ExperimentHub)
eh <- ExperimentHub()
q <- query(eh, "TENxXenium")
spe_example <- q[["EH8547"]]
# get the required input list
example_lst <- convert_data(x=spe_example, sample_names = "sample01")
# -----------------------------------------------------------------------------
# Example clustering
# get the count matrix
cm <- spe_example@assays@data$counts
all_genes = row.names(cm)
example_seu <- CreateSeuratObject(counts = cm,
min.cells = 2, min.features = 1)
example_seu <- NormalizeData(example_seu,
normalization.method = "LogNormalize")
all.genes <- rownames(example_seu)
example_seu <- ScaleData(example_seu, features = all.genes)
example_seu <- RunPCA(example_seu, features = all.genes)
ElbowPlot(example_seu)
example_seu <- FindNeighbors(example_seu, dims = 1:15)
example_seu <- FindClusters(example_seu, resolution = 0.1)
seu_clusters <- as.data.frame(example_seu$seurat_clusters,nm="cluster")
seu_clusters$cell_id <- colnames(example_seu)
# -----------------------------------------------------------------------------
# combine cluster labels and the coordiantes
# make sure the cluster information contains column names:
# cluster, x, y, sample and cell_id
clusters_info <- as.data.frame(spe_example@int_colData$spatialCoords)
colnames(clusters_info) <- c("x","y")
clusters_info$cell_id <- row.names(clusters_info)
clusters_info$sample <- spe_example$sample_id
clusters_info <- merge(clusters_info, seu_clusters, by="cell_id")
w_x <- c(floor(min(clusters_info$x,
example_lst$trans_lst$sample01$x)),
ceiling(max(clusters_info$x,
example_lst$trans_lst$sample01$x)))
w_y <- c(floor(min(clusters_info$y,
example_lst$trans_lst$sample01$y)),
ceiling(max(clusters_info$y,
example_lst$trans_lst$sample01$y)))
# -----------------------------------------------------------------------------
# build spatial vectors from count matrix and cluster coordinates
example_vectors_cm <- get_vectors(trans_lst= NULL,
cm_lst=example_lst$cm_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
# -----------------------------------------------------------------------------
# build spatial vectors from transcript coordinates and cluster coordinates
example_vectors_tr <- get_vectors(trans_lst= example_lst$trans_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
```

### from a SpatialFeatureExperiment object

```{r, eval=FALSE}
sfe_example <- toSpatialFeatureExperiment(spe_example)
example_sfe_lst <- convert_data(x=sfe_example,sample_names = "sample01")
# -----------------------------------------------------------------------------
# build spatial vectors from count matrix and cluster coordinates
# make sure the cluster information contains column names:
# cluster, x, y, sample and cell_id
example_vectors_cm <- get_vectors(trans_lst= NULL,
cm_lst=example_sfe_lst$cm_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
# -----------------------------------------------------------------------------
# build spatial vectors from transcript coordinates and cluster coordinates
example_vectors_tr <- get_vectors(trans_lst= example_sfe_lst$trans_lst,
cluster_info = clusters_info,
bin_type="square",
bin_param=c(10,10),
all_genes = all_genes,
w_x=w_x, w_y=w_y)
```

## Visualise the clusters over the tissue space
We can plot the cells coordinates for each cluster of Replicate 1 subset
Expand Down

0 comments on commit b501c0b

Please sign in to comment.