Clustree

Description

Deciding what resolution to use can be a difficult question when approaching a clustering analysis. One way to approach this problem is to look at how samples move as the number of clusters increases. This package allows you to produce clustering trees, a visualisation for interrogating clusterings as resolution increases.

Add node labels

Description

Add node labels to a clustering tree plot with the specified aesthetics.

Usage

add_node_labels(
 node_label,
 node_colour,
 node_label_size,
 node_label_colour,
 node_label_nudge,
 allowed
)

Arguments

Add node points

Description

Add node points to a clustering tree plot with the specified aesthetics.

Usage

add_node_points(node_colour, node_size, node_alpha, allowed)

Arguments

Aggregate metadata

Description

Aggregate a metadata column to get a summarized value for a cluster node

Usage

aggr_metadata(node_data, col_name, col_aggr, metadata, is_cluster)

Arguments

Value

data.frame with aggregated data

Assert colour node aesthetics

Description

Raise error if an incorrect set of colour node parameters has been supplied.

Usage

assert_colour_node_aes(
 node_aes_name,
 prefix,
 metadata,
 node_aes,
 node_aes_aggr,
 min,
 max
)

Arguments

Assert node aesthetics

Description

Raise error if an incorrect set of node parameters has been supplied.

Usage

assert_node_aes(node_aes_name, prefix, metadata, node_aes, node_aes_aggr)

Arguments

Assert numeric node aesthetics

Description

Raise error if an incorrect set of numeric node parameters has been supplied.

Usage

assert_numeric_node_aes(
 node_aes_name,
 prefix,
 metadata,
 node_aes,
 node_aes_aggr,
 min,
 max
)

Arguments

Build tree graph

Description

Build a tree graph from a set of clusterings, metadata and associated aesthetics

Usage

build_tree_graph(
 clusterings,
 prefix,
 count_filter,
 prop_filter,
 metadata,
 node_aes_list
)

Arguments

Value

tidygraph::tbl_graph object containing the tree graph

Calculate SC3 stability

Description

Calculate the SC3 stability index for every cluster at every resolution in a set of clusterings. The index varies from 0 to 1, where 1 suggests that a cluster is more stable across resolutions. See calc_sc3_stability_cluster() for more details.

Usage

calc_sc3_stability(clusterings)

Arguments

Value

matrix with stability score for each cluster

Calculate single SC3 stability

Description

Calculate the SC3 stability index for a single cluster in a set of clusterings. The index varies from 0 to 1, where 1 suggests that a cluster is more stable across resolutions.

Usage

calc_sc3_stability_cluster(clusterings, res, cluster)

Arguments

Details

This index was originally introduced in the SC3 package for clustering single-cell RNA-seq data. Clusters are awarded increased stability if they share the same samples as a cluster at another resolution and penalised at higher resolutions. We use a slightly different notation to describe the score but the results are the same:

s(c_{k, i}) = \frac{1}{size(L) + 1} \sum_{l \in L} \sum_{j \in N_l} \frac{size(c_{k, i} \cap c_{l, j})}{size(c_{l, j}) * size(N_l) ^ 2}

Where:

• ⁠c_{x, y}⁠ is cluster y at resolution x

• k is the resolution of the cluster we want to score

• i is the index of the cluster we want to score

• L is the set of all resolutions except k

• l is a resolution in L

• N_l is the set of clusters at resolution l that share samples with ⁠c_{k, i}⁠

• j is a cluster in N_l

Value

SC3 stability index

See Also

The documentation for the calculate_stability function in the SC3 package

Check node aes list

Description

Warn if node aesthetic names are incorrect

Usage

check_node_aes_list(node_aes_list)

Arguments

Value

Corrected node aesthetics list

Plot a clustering tree

Description

Creates a plot of a clustering tree showing the relationship between clusterings at different resolutions.

Usage

clustree(x, ...)
## S3 method for class 'matrix'
clustree(
 x,
 prefix,
 suffix = NULL,
 metadata = NULL,
 count_filter = 0,
 prop_filter = 0.1,
 layout = c("tree", "sugiyama"),
 use_core_edges = TRUE,
 highlight_core = FALSE,
 node_colour = prefix,
 node_colour_aggr = NULL,
 node_size = "size",
 node_size_aggr = NULL,
 node_size_range = c(4, 15),
 node_alpha = 1,
 node_alpha_aggr = NULL,
 node_text_size = 3,
 scale_node_text = FALSE,
 node_text_colour = "black",
 node_text_angle = 0,
 node_label = NULL,
 node_label_aggr = NULL,
 node_label_size = 3,
 node_label_nudge = -0.2,
 edge_width = 1.5,
 edge_arrow = TRUE,
 edge_arrow_ends = c("last", "first", "both"),
 show_axis = FALSE,
 return = c("plot", "graph", "layout"),
 ...
)
## S3 method for class 'data.frame'
clustree(x, prefix, ...)
## S3 method for class 'SingleCellExperiment'
clustree(x, prefix, exprs = "counts", ...)
## S3 method for class 'seurat'
clustree(x, prefix = "res.", exprs = c("data", "raw.data", "scale.data"), ...)
## S3 method for class 'Seurat'
clustree(
 x,
 prefix = paste0(assay, "_snn_res."),
 exprs = c("data", "counts", "scale.data"),
 assay = NULL,
 ...
)

Arguments

Details

Data sources

Plotting a clustering tree requires information about which cluster each sample has been assigned to at different resolutions. This information can be supplied in various forms, as a matrix, data.frame or more specialised object. In all cases the object provided must contain numeric columns with the naming structure PXS where P is a prefix indicating that the column contains clustering information, X is a numeric value indicating the clustering resolution and S is any additional suffix to be removed. For SingleCellExperiment objects this information must be in the colData slot and for Seurat objects it must be in the meta.data slot. For all objects except matrices any additional columns can be used as aesthetics, for matrices an additional metadata data.frame can be supplied if required.

Filtering

Edges in the graph can be filtered by adjusting the count_filter and prop_filter parameters. The count_filter removes any edges that represent less than that number of samples, while the prop_filter removes edges that represent less than that proportion of cells in the node it points towards.

Node aesthetics

The aesthetics of the plotted nodes can be controlled in various ways. By default the colour indicates the clustering resolution, the size indicates the number of samples in that cluster and the transparency is set to 100%. Each of these can be set to a specific value or linked to a supplied metadata column. For a SingleCellExperiment or Seurat object the names of genes can also be used. If a metadata column is used than an aggregation function must also be supplied to combine the samples in each cluster. This function must take a vector of values and return a single value.

Layout

The clustering tree can be displayed using either the Reingold-Tilford tree layout algorithm or the Sugiyama layout algorithm for layered directed acyclic graphs. These layouts were selected as the are the algorithms available in the igraph package designed for trees. The Reingold-Tilford algorithm places children below their parents while the Sugiyama places nodes in layers while trying to minimise the number of crossing edges. See igraph::layout_as_tree() and igraph::layout_with_sugiyama() for more details. When use_core_edges is TRUE (default) only the core tree of the maximum in proportion edges for each node are used for constructing the layout. This can often lead to more attractive layouts where the core tree is more visible.

Value

a ggplot object (default), a tbl_graph object or a ggraph layout object depending on the value of return

Examples

data(nba_clusts)
clustree(nba_clusts, prefix = "K")

Overlay a clustering tree

Description

Creates a plot of a clustering tree overlaid on a scatter plot of individual samples.

Usage

clustree_overlay(x, ...)
## S3 method for class 'matrix'
clustree_overlay(
 x,
 prefix,
 metadata,
 x_value,
 y_value,
 suffix = NULL,
 count_filter = 0,
 prop_filter = 0.1,
 node_colour = prefix,
 node_colour_aggr = NULL,
 node_size = "size",
 node_size_aggr = NULL,
 node_size_range = c(4, 15),
 node_alpha = 1,
 node_alpha_aggr = NULL,
 edge_width = 1,
 use_colour = c("edges", "points"),
 alt_colour = "black",
 point_size = 3,
 point_alpha = 0.2,
 point_shape = 18,
 label_nodes = FALSE,
 label_size = 3,
 plot_sides = FALSE,
 side_point_jitter = 0.45,
 side_point_offset = 1,
 ...
)
## S3 method for class 'data.frame'
clustree_overlay(x, prefix, ...)
## S3 method for class 'SingleCellExperiment'
clustree_overlay(
 x,
 prefix,
 x_value,
 y_value,
 exprs = "counts",
 red_dim = NULL,
 ...
)
## S3 method for class 'seurat'
clustree_overlay(
 x,
 x_value,
 y_value,
 prefix = "res.",
 exprs = c("data", "raw.data", "scale.data"),
 red_dim = NULL,
 ...
)
## S3 method for class 'Seurat'
clustree_overlay(
 x,
 x_value,
 y_value,
 prefix = paste0(assay, "_snn_res."),
 exprs = c("data", "counts", "scale.data"),
 red_dim = NULL,
 assay = NULL,
 ...
)

Arguments

Details

Data sources

Plotting a clustering tree requires information about which cluster each sample has been assigned to at different resolutions. This information can be supplied in various forms, as a matrix, data.frame or more specialised object. In all cases the object provided must contain numeric columns with the naming structure PXS where P is a prefix indicating that the column contains clustering information, X is a numeric value indicating the clustering resolution and S is any additional suffix to be removed. For SingleCellExperiment objects this information must be in the colData slot and for Seurat objects it must be in the meta.data slot. For all objects except matrices any additional columns can be used as aesthetics.

Filtering

Edges in the graph can be filtered by adjusting the count_filter and prop_filter parameters. The count_filter removes any edges that represent less than that number of samples, while the prop_filter removes edges that represent less than that proportion of cells in the node it points towards.

Node aesthetics

The aesthetics of the plotted nodes can be controlled in various ways. By default the colour indicates the clustering resolution, the size indicates the number of samples in that cluster and the transparency is set to 100%. Each of these can be set to a specific value or linked to a supplied metadata column. For a SingleCellExperiment or Seurat object the names of genes can also be used. If a metadata column is used than an aggregation function must also be supplied to combine the samples in each cluster. This function must take a vector of values and return a single value.

Colour aesthetic

The colour aesthetic can be applied to either edges or sample points by setting use_colour. If "edges" is selected edges will be coloured according to the clustering resolution they originate at. If "points" is selected they will be coloured according to the cluster they are assigned to at the highest resolution.

Dimensionality reductions

For SingleCellExperiment and Seurat objects precomputed dimensionality reductions can be used for x or y aesthetics. To do so red_dim must be set to the name of a dimensionality reduction in reducedDimNames(x) (for a SingleCellExperiment) or x@dr (for a Seurat object). x_value and y_value can then be set to red_dimX when red_dim matches the red_dim argument and X is the column of the dimensionality reduction to use.

Value

a ggplot object if plot_sides is FALSE or a list of ggplot objects if plot_sides is TRUE

Examples

data(nba_clusts)
clustree_overlay(nba_clusts, prefix = "K", x_value = "PC1", y_value = "PC2")

Get tree edges

Description

Extract the edges from a set of clusterings

Usage

get_tree_edges(clusterings, prefix)

Arguments

Value

data.frame containing edge information

Get tree nodes

Description

Extract the nodes from a set of clusterings and add relevant attributes

Usage

get_tree_nodes(clusterings, prefix, metadata, node_aes_list)

Arguments

Value

data.frame containing node information

Clustered NBA positions dataset

Description

NBA positions dataset clustered using k-means with a range of values of k

Usage

nba_clusts

Format

nba_clusts is a data.frame containing the NBA positions dataset with additional columns holding k-means clusterings at different values of k and the first two principal components

• Position - Player position

• TurnoverPct - Turnover percentage

• ReboundPct - Rebound percentage

• AssistPct - Assist percentage

• FieldGoalPct - Field goal percentage

• K1 - K5 - Results of k-means clustering

• PC1 - First principal component

• PC2 - Second principal component

Source

NBA positions downloaded from https://github.com/lazappi/nba_positions.

The source dataset is available from Kaggle at https://www.kaggle.com/drgilermo/nba-players-stats/data?select=Seasons_Stats.csv and was originally scraped from Basketball Reference.

See https://github.com/lazappi/clustree/blob/master/data-raw/nba_clusts.R for details of how clustering was performed.

Overlay node points

Description

Overlay clustering tree nodes on a scatter plot with the specified aesthetics.

Usage

overlay_node_points(
 nodes,
 x_value,
 y_value,
 node_colour,
 node_size,
 node_alpha
)

Arguments

Plot overlay side

Description

Plot the side view of a clustree overlay plot. If the ordinary plot shows the tree from above this plot shows it from the side, highlighting either the x or y dimension and the clustering resolution.

Usage

plot_overlay_side(
 nodes,
 edges,
 points,
 prefix,
 side_value,
 graph_attr,
 node_size_range,
 edge_width,
 use_colour,
 alt_colour,
 point_size,
 point_alpha,
 point_shape,
 label_nodes,
 label_size,
 y_jitter,
 y_offset
)

Arguments

Value

ggplot object

Simulated scRNA-seq dataset

Description

A simulated scRNA-seq dataset generated using the splatter package and clustered using the SC3 and Seurat packages.

Usage

sc_example

Format

sc_example is a list holding a simulated scRNA-seq dataset. Items in the list included the simulated counts, normalised log counts, tSNE dimensionality reduction and cell assignments from SC3 and Seurat clustering.

Source

# Simulation
library("splatter") # Version 1.2.1
sim <- splatSimulate(batchCells = 200, nGenes = 10000,
 group.prob = c(0.4, 0.2, 0.2, 0.15, 0.05),
 de.prob = c(0.1, 0.2, 0.05, 0.1, 0.05),
 method = "groups", seed = 1)
sim_counts <- counts(sim)[1:1000, ]
# SC3 Clustering
library("SC3") # Version 1.7.6
library("scater") # Version 1.6.2
sim_sc3 <- SingleCellExperiment(assays = list(counts = sim_counts))
rowData(sim_sc3)$feature_symbol <- rownames(sim_counts)
sim_sc3 <- normalise(sim_sc3)
sim_sc3 <- sc3(sim_sc3, ks = 1:8, biology = FALSE, n_cores = 1)
sim_sc3 <- runTSNE(sim_sc3)
# Seurat Clustering
library("Seurat") # Version 2.2.0
sim_seurat <- CreateSeuratObject(sim_counts)
sim_seurat <- NormalizeData(sim_seurat, display.progress = FALSE)
sim_seurat <- FindVariableGenes(sim_seurat, do.plot = FALSE,
 display.progress = FALSE)
sim_seurat <- ScaleData(sim_seurat, display.progress = FALSE)
sim_seurat <- RunPCA(sim_seurat, do.print = FALSE)
sim_seurat <- FindClusters(sim_seurat, dims.use = 1:6,
 resolution = seq(0, 1, 0.1),
 print.output = FALSE)
sc_example <- list(counts = counts(sim_sc3),
 logcounts = logcounts(sim_sc3),
 tsne = reducedDim(sim_sc3),
 sc3_clusters = as.data.frame(colData(sim_sc3)),
 seurat_clusters = sim_seurat@meta.data)

Store node aesthetics

Description

Store the names of node attributes to use as aesthetics as graph attributes

Usage

store_node_aes(graph, node_aes_list, metadata)

Arguments

Value

graph with additional attributes