Package 'scCAN'

adjustedRandIndex

Description

The function to calculate adjusted Rand index value with the inputs of true clusters and predicted clusters

Usage

adjustedRandIndex(x, y)

Arguments

x	A vector that contain predicted cluster assignment.
y	A vector that contain true cluster assignment.

Value

An value number ranging from 0 to 1 where 1 indicates a perfect clustering result and 0 means random partition.

---

calculate_celltype_prob

Description

Calculate clusters and cell types similarity based on the markers.

Usage

calculate_celltype_prob(clt_marker_list, marker_database_list, type = "jacc")

Arguments

clt_marker_list	A list of markers for all cluster.
marker_database_list	A list of markers of all reference cell types.
type	A parameter to select the method to measure cluster and cell type similarity jacc - Jaccard index. ac - Accuracy. f1 - F1 score. .

Value

A confusion matrix between clusters and cell types. Each cell represents a probabilty of a cluster belongs to a cell type.

---

curate_markers

Description

Filter genes that have low p-value and fold-change.

Usage

curate_markers(
  whole_list,
  gene_names,
  wilcox_threshold = 0.001,
  logfc_threshold = 1.5
)

Arguments

whole_list	A list of markers for all clusters.
gene_names	All the gene names of the expression matrix.
wilcox_threshold	A threshold for p-value wilcox_threshold = 0.001 by default.
logfc_threshold	A threshold for fold-change logfc_threshold = 1.5 by default.

Value

A list of markers that are strong expressed for discovered clusters.

---

find_markers

Description

Perform cluster-wise Wilcox test and fold-change for each gene.

Usage

find_markers(input_data_matrix, cluster_labels, identity = 1, threads = 8)

Arguments

input_data_matrix	An expression matrix in which rows are genes and columns are cells.
cluster_labels	A vector of cluster labels obtained from clustering methods.
identity	A parameter to select specific cluster identity = 1 by default.
threads	A parameter to control number of cores used for analysis threads = 1 by default.

Value

A list that contains p-value and fold-change ratio for all genes of each cluster.

---

find_specific_marker

Description

Calculate cluster and cell type similarity based on the markers.

Usage

find_specific_marker(gene_name, f_list, type = "jacc")

Arguments

gene_name	A list of markers belong to the cluster.
f_list	A list of markers belongs to a reference cell type.
type	A parameter to select the method to measure cluster and cell type similarity jacc - Jaccard index. ac - Accuracy. f1 - F1 score. .

Value

A vector of probabilties of a cluster belongs to cell types.

---

get_cluster_markers

Description

Find markers for each cluster

Usage

get_cluster_markers(input_data_matrix, labels_vector, threads = 1)

Arguments

input_data_matrix	An expression matrix in which rows are genes and columns are cells.
labels_vector	A vector of cluster labels ontained from clustering methods.
threads	A parameter to control number of cores used for analysis threads = 1 by default.

Value

A list that contains markers for each cluster.

---

scCAN

Description

This is the main function to perform sc-RNA seq data clustering clustering. scCAN is fully unsupervised scRNA-seq clustering framework that uses deep neural network and network fusion-based clustering algorithm. First, scCAN applies a non-negative autoencoder to filter scRNA-seq data. Second, the filtered data is passed to stacked Bayesian autoencoder to get multiple low-dimensional representations of input data. Subsequently, scCAN converts these compressed data into networks and unify those networks to a single graph. Then, scCAN uses a spectral clustering algorithm to obtain final clusters assignment.

Usage

scCAN(
  data,
  sparse = FALSE,
  n.neighbors = 30,
  alpha = 0.5,
  n.iters = 10,
  ncores = 10,
  r.seed = 1,
  subsamp = TRUE,
  k = 2:15,
  samp.size = 5000
)

Arguments

data	Gene expression matrix, with rows represent samples and columns represent genes.
sparse	Boolen variable indicating whether data is a sparse matrix. The input must be a non negative sparse matrix.
n.neighbors	Number of neighboring cells that are used to caculate the edge's weight. The number of neighbors are set n.neighbors = 30 by default.
alpha	A hyper parameter that control the weight of graph. This values is set to alpha = 0.5 by default.
n.iters	A hyper-parameter to set the number of network fusion iterations. It is set to n.iters = 10 by default.
ncores	Number of processor cores to use.
r.seed	A parameter to set a seed for reproducibility. This values is set to r.seed = 1 by default.
subsamp	Enable subsampling process for big data. This values is set to subsamp = T by default.
k	A vector to search for optimal number of cluster.
samp.size	A parameter to control number of sub-sampled cells.

Value

List with the following keys:

• cluster - A numeric vector containing cluster assignment for each sample.

• k - The optimal number of cluster.

• latent - The latent data generated from autoencoders.

References

1. Duc Tran, Hung Nguyen, Bang Tran, Carlo La Vecchia, Hung N. Luu, Tin Nguyen (2021). Fast and precise single-cell data analysis using a hierarchical autoencoder. Nature Communications, 12, 1029. doi: 10.1038/s41467-021-21312-2

Examples

## Not run: 
# Not run if scDHA has not installed yet.
# Load the package and the example data (SCE dataset)
library(scCAN)
#Load example data
data("SCE")

#Get data matrix and label
data <- t(SCE$data); label <- as.character(SCE$cell_type1)

#Generate clustering result, the input matrix has rows as samples and columns as genes
result <- scCAN(data, r.seed = 1)

#Get the clustering result
cluster <- result$cluster

#Calculate adjusted Rand Index
ari <- round(scCAN::adjustedRandIndex(cluster,label), 2)
message(paste0("ARI = ", ari))


## End(Not run)

---

SCE

Description

SCE dataset includes scRNA-seq data and cell type information.

Usage

SCE

Format

An object of class list of length 2.

---