README.md

In csoneson/countsimQC: Compare Characteristic Features of Count Data Sets

countsimQC

countsimQC is an R package that provides functionality to create a comprehensive report comparing many different characteristics across multiple count data sets. One important use case is comparing one or more synthetic (e.g., RNA-seq) count matrices to a real count matrix, possibly the one based on which the synthetic data sets were generated. However, any collection of one or more count matrices can be visualized and compared.

If you use countsimQC for your work, we appreciate if you cite the accompanying paper:

• Soneson C and Robinson MD: Towards unified quality verification of synthetic count data with countsimQC. Bioinformatics 34(4):691-692 (2018).

Installation

countsimQC can be installed from Bioconductor with the following commands. Note that R version >= 3.5 and Bioconductor version >= 3.8 are required in order to use the BiocManager package. If you have an older version of R (3.4), you can still install countsimQC v0.5.4 (see the Releases tab in the GitHub repository). Please see the NEWS file for differences between versions.

## Install `BiocManager` if needed
if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

## Install countsimQC
BiocManager::install("countsimQC")

Getting started

To run countsimQC and generate a report, you simply need to call the function countsimQCReport(), with an input consisting of a named list of DESeqDataSets (see the DESeq2 package for a description of this class). Each DESeqDataSet should correspond to one data set and contain a count matrix, a data frame with sample information and a design formula, which is needed for proper dispersion calculations. To generate a DESeqDataSet from a count matrix counts, a sample information data frame sample_df and a design formula formula (of the form ~ predictors), you can do as follows:

library(DESeq2)
dds <- DESeqDataSetFromMatrix(countData = counts, 
                              colData = sample_df,
                              design = formula)

There are many other ways of generating valid DESeqDataSets, depending on in what form your counts are (e.g., reading directly from HTSeq output, or from a tximport output object (see the DESeq2 vignette).

countsimQC contains an small example list with subsets of three data sets: two synthetic ones and the real data set that was used to generate them. The following code generates a comparative report for these three data sets:

library(countsimQC)
data(countsimExample)
countsimQCReport(ddsList = countsimExample, 
                 outputFile = "countsimReport.html", 
                 outputDir = "./", 
                 description = "This is a comparison of three count data sets.")

For more detailed information about how to use the package, we refer to the vignette:

browseVignettes("countsimQC")

Example reports

• Comparison of 16S microbiome species count matrices for four body subsites from the Human Microbiome Project
• Comparison of three real bulk RNA-seq data sets
• Comparison of gene- and transcript-level count matrices for a single-cell RNA-seq data set
• Comparison of four real scRNA-seq data sets
• Comparison of two simulated scRNA-seq data sets to the underlying real data set
• Comparison of six simulated bulk RNA-seq data set with different number of genes

csoneson/countsimQC documentation built on Oct. 29, 2023, 2:14 p.m.