README.md

In ATpoint/diffanalyseR: Wrapper functions for common tasks related to analysis of NGS counts data

diffanalyseR

Function to automate common tasks related to analysis of NGS count data in R.

Installation

#/ Check for BiocManager:
if (!requireNamespace("BiocManager", quietly = TRUE)){
  install.packages("BiocManager")
  BiocManager::install()
}

#/ Bioc packages:
pk <- c("BiocParallel", "edgeR", "GenomicRanges", "SummarizedExperiments")
BiocManager::install(pk)

install.packages("remotes")
remotes::install_github("ATpoint/diffanalyseR")

Brief overview over the functions

See the help of each functions for details and examples.

• AverageBy averages a numeric matrix by the indices of a named list

• binMatrix bins a matrix, for example a 20-col matrix with 4 bins would mean that 5 consecutive columns get averaged into one

• gosty is a wrapper around gprofiler2::gost for functional enrichment of gene lists

• makeAllContrasts takes a character of comparison groups and then forms all unique pairwise contrasts in either edgeR or DESeq2 style

• MAplotFromCts produces MAplots with base::smoothScatter based on a normalized count matrix, mainly to explore whether sample have been properly normalized

• MostVariableRows finds the most variable rows based on rowwise variance

• NormOffsetsFromTxLength calculate log-offsets that can be used in edgeR and company to correct for both sequencing depth and composition (e.g. via TMM) plus corrects for average transcript length (e.g. via tximport)

• Permut_GRanges calculate permutation-based p-values comparing an observed overlap between two GRanges objects with the expected overlap between one of the objects and a background set

• res_edgeR automates contrast testing via the edgeR QLF framework

• run_apeglm is a wrapper around the lfcShrink function with type=apeglm of DESeq2 which automates the releveling and rerunning of the Wald test in case certain contrasts are not accessable via the resultsNames function

• rowScale is a matrixStats-based wrapper to efficiently scale and center numeric matrices as efficient replacement of t(scale(t(x))) avoiding the double transposition which takes time with very large data such as single-cell stuff

• ScaleByQuantile winsorizes a numeric matrix based on user-defined quantiles. This can be useful when plotting heatmaps to avoid skewed color breaks due to outliers

ATpoint/diffanalyseR documentation built on Oct. 8, 2021, 10:11 p.m.