xiaomingfu2013 / splatter-paper

Data and analysis for the Splatter paper

Splatter paper

Data and analysis for the paper "Splatter: Simulation of Single-cell RNA sequencing data".

Directory structure

• additional - Supplementary figures and files
• analysis - Analysis files and output
• data - Input data files
• figures - Figures used in the main paper
• output - Additional intermediate files
• R - R functions used in analysis

Data

Data files used in the analysis are available in the data.tar.gz file. This file should be extracted to data before attempting to run any of the analysis.

tar -xzvf data.tar.gz

After extraction the data directory will contain the following files:

• datasets.txt - Metadata about the various datasets
• Camp.txt - The Camp dataset
• Engel.tsv - The Engel dataset
• Klein.csv - The Klein dataset
• Tung.txt - The Tung dataset
• Zeisel.txt - The Zeisel dataset

Analysis

The code for completing the analysis shown in the paper is provided as the following Rmarkdown files in the analysis directory:

• simulations.Rmd - Some examples of Splat simulations.
• performance.Rmd - Processing time benchmarking.
• datasets.Rmd - Comparison of simulations based on various datasets.
• clustering.Rmd - Example evaluation of the SC3 clustering method.
• supplementary.Rmd - Supplementary figures.

There are two additional Rmd files which are rendered via supplementary.Rmd. These are additional_figures.Rmd which combines the additional figures into a single PDF and sessionInfo.Rmd which outputs the details of all the packages used during the analysis.

Running the analysis files will produce figure files in the figures and additional directories as well as data files in the output directory.

Please be aware that some of the analysis (particularly datasets.Rmd) requires large amounts of resources (processing, memory, time) and may require slight modifications to run in your environment.

R

This directory contains the following functions used in the analysis:

• load_datasets.R
  • loadDataset - Takes a row from datasets.txt and the path to the data files then returns an expression matrix for that dataset
• simulate_datasets.R
  • simData - Takes a counts matrix, estimates parameters and simulates data using various models
• test_genes.R
  • testGenesGoF - Test the goodness of fit for each gene with regards to various distributions
• utils.R
  • chrRound - Rounds a number for presentation, converting it to a string
  • logistic - Implementation of the logistic function
  • mcri.palettes - List of colour palettes used for some plots
  • mcriPalette - Returns colour palette of particular size