DIMMSC

An R package for clustering droplet-based single cell transcriptomic data.

Version: 0.2.1 (Date: 2017-08-17)

See Homepage @ Wei Chen's Lab

Introduction

DIMMSC is an R package for clustering droplet-based single cell transcriptomic data. It uses Dirichlet mixture prior to characterize variations across different clusters. An expectation-maximization algorithm is used for parameter inference. This package can provide clustering uncertainty.

plot_tsne_clusters performs log2 normalization, PCA and t-SNE on the data and plot clusters on the t-SNE projection.

Installation

Install third-party R package cellrangerRkit from 10X Genomics.

Install DIMMSC with R command

install.packages(pkgs = "DIMMSC.tar.gz", repos = NULL, type = "source")

# Or install from GitHub
library(devtools)
install_github("wt2015-github/DIMMSC")

Or terminal command

R CMD INSTALL DIMMSC.tar.gz

Usage

DIMMSC(data, K = 2, method_cluster_initial = "kmeans", method_alpha_initial = "Ronning", maxiter = 200, tol = 1e-04, lik.tol = 0.01, customized_initial = NULL)

plot_tsne_clusters(data, cluster)

Arguments

• data : a G*C matrix with G genes and C cells
• K : number of desired clusters
• method_cluster_initial : method for initializing clusters, "kmeans", "random" or "customized"
• method_alpha_initial : method for initializing the alpha matrix for EM algorithm, "Ronning" (Ronning's method, 1989) or "Weir" (Weir and Hill's method 2002)
• maxiter : maximum number of iterations
• tol : a convergence tolerance for the difference of vector pie between iterations
• lik.tol : a convergence tolerance for the difference of log-likelihoods between iterations
• customized_initial : a vector of positive integers {1,2,...,K} for the customized initial clusters, the length should be equal to the number of cells
• cluster : a vector of clustering member ship, e.g. mem of DIMMSC output

Values

DIMMSC returns a list object containing:

• pie : a vector of pie estimates
• delta : a C*K matrix with probability that each cell belongs to each cluster
• alpha : a K*G matrix of alpha estimates
• mem : a vector of clustering label
• loglik : the final log likelihood after iterations
• AIC : Akaike information criterion (AIC)
• BIC : Bayesian information criterion (BIC)

plot_tsne_clusters outputs a figure of t-SNE and clusters

Example:

# Load the example data data_DIMMSC
data("data_DIMMSC")

# Run DIMMSC
result <- DIMMSC(data=data_DIMMSC, K=3, method_cluster_initial="kmeans", method_alpha_initial="Ronning", maxiter=200, tol=1e-4, lik.tol=1e-2)

# Plot t-SNE and clusters
plot_tsne_clusters(data=data_DIMMSC, cluster=result$mem)

Publications

• Zhe Sun, Ting Wang, Ke Deng, Xiao-Feng Wang, Robert Lafyatis, Ying Ding, Ming Hu, Wei Chen. DIMM-SC: A Dirichlet mixture model for clustering droplet-based single cell transcriptomic data. Bioinformatics 2017.

Contact

Ting Wang ([email protected]), Wei Chen.