A quick tour of RCSL

In RCSL: Rank Constrained Similarity Learning for Single Cell RNA Sequencing Data

library(knitr)
opts_chunk$set(
  collapse = TRUE,
  fig.align = 'center', fig.width = 6, fig.height = 5, dev = 'png',
  comment = "#>"
)

Introduction

RCSL is an R toolkit for single-cell clustering and trajectory analysis using single-cell RNA-seq data.

Installation

Install RCSL package and other requirements

RCSL can be installed directly from GitHub with 'devtools'.

library(devtools)
devtools::install_github("QinglinMei/RCSL")

Now we can load RCSL. We also load the SingleCellExperiment, ggplot2 and igraph package.

library(RCSL)
library(SingleCellExperiment)
library(ggplot2)
library(igraph)
library(umap)

Run RCSL

Load dataset (yan)

We illustrate the usage of RCSL on a human preimplantation embryos and embryonic stem cells(Yan et al., (2013)). The 'yan' data is distributed together with the RCSL package, with 90 cells and 20,214 genes:

head(ann)
yan[1:3, 1:3]
origData <- yan
label <- ann$cell_type1

1. Pre-processing

In practice, we find it always beneficial to pre-process single-cell RNA-seq datasets, including: 1. Log transformation. 2. Gene filter

data <- log2(as.matrix(origData) + 1)
gfData <- GenesFilter(data)

2. Calculate the initial similarity matrix S

resSimS <- SimS(gfData)

3. Estimate the number of clusters C

Estimated_C <- EstClusters(resSimS$drData,resSimS$S)

4. Calculate the block diagonal matrix B

resBDSM <- BDSM(resSimS$S, Estimated_C)

Calculate accuracy of the clustering

ARI_RCSL <- igraph::compare(resBDSM$y, label, method = "adjusted.rand")

Trajectory analysis to time-series datasets

DataName <- "Yan"
res_TrajecAnalysis <- TrajectoryAnalysis(gfData, resSimS$drData, resSimS$S,
                                         clustRes = resBDSM$y, TrueLabel = label, 
                                         startPoint = 1, dataName = DataName)

Display the constructed MST

res_TrajecAnalysis$MSTPlot

Display the plot of the pseudo-temporal ordering

res_TrajecAnalysis$PseudoTimePlot

Display the plot of the inferred developmental trajectory

res_TrajecAnalysis$TrajectoryPlot

RCSL documentation built on April 19, 2021, 9:06 a.m.