Introduction

Apart from our interactive web application (http://hrpi.ddnetbio.com/), users who are more adept with the R programming language can download our ready-to-use Seurat objects to further interrogate the rich single-cell data generated in this work. Here, we provide some examples (with code) on how to use these ready-to-use Seurat objects.

Citation: TBC

Getting ready

This ready-to-use Seurat object, containing the integrated sn/scRNA-seq of human reprogramming intermediates, can be downloaded at http://hrpi.ddnetbio.com/. Note that the Seurat object is created using Seurat v3.1.1 and may be incompatible with older or newer versions of Seurat. First, we proceed to load the required libraries and Seurat object.

rm(list=ls())
library(data.table)
library(Seurat)
library(ggplot2)
library(patchwork)
seu = readRDS("readySeu_hrpi.rds")

What is in this Seurat object?

This Seurat object, comprising 43791 cells and 11549 genes, is generated from Seurat3 integration of time-resolved snRNA-seq (14 timepoints) and media-resolved scRNA-seq (3 conditions) experiments. The integrated and original gene expression can be found in the “integrated” and “RNA” assays respectively. Five different dimensional reductions were pre-calculated to uncover different aspects of the reprogramming trajectory.

seu
## An object of class Seurat 
## 29250 features across 43791 samples within 2 assays 
## Active assay: integrated (11549 features)
##  1 other assay present: RNA
##  5 dimensional reductions calculated: pca, tsne, umap, diffmap, fdl
names(seu)
## [1] "RNA"            "integrated"     "integrated_nn"  "integrated_snn"
## [5] "pca"            "tsne"           "umap"           "diffmap"       
## [9] "fdl"

Different dimensional reductions

Here, we calculated various dimension dimensional reductions to visualise the reprogramming trajectory. These dimensional reductions include:

  • fdl: Force Directed Layout (FDL)
  • pca: Principal Component Analysis (PCA)
  • tsne: t-Distributed Stochastic Neighbor Embedding (tSNE)
  • umap: Uniform Manifold Approximation and Projection (UMAP)
  • diffmap: diffusion maps
# FDL
ggOut = DimPlot(seu, reduction = "fdl", cols = seu@misc$color$cluster)
ggOut + coord_fixed(ratio = diff(range(ggOut$data$FDL_1)) / 
                      diff(range(ggOut$data$FDL_2))) # square aspect ratio