normR - normR obeys regime mixture rules

Normalization and Difference Calling for Next Generation Sequencing (NGS)

Experiments via Joint Multinomial Modeling

Two NGS tracks are modeled simultaneously by fitting a binomial mixture model on mapped read counts. In the first counting process, a desired smoothing kernel (bin size) and read characteristic threshold (quality, SAMFLAG) can be specified. In a second step a binomial mixture model with a user-specified number of components is fit to the data. The fit yields different enrichment regimes in the supplied NGS tracks. Log-space computation is done in C/C++ where OpenMP enables for fast parallel computation.

Release Version

The master branch is always in sync with the normR Bioconductor release and the normR github Bioconductor mirror. A R 3.2 compliant version can be found in the normR R3.2 tree.

Installation

To install normR from the release repository, easiest way is to use Bioconductor or devtools:

#install dependencies
if (!requireNamespace("BiocManager", quietly=TRUE))
    install.packages("BiocManager")
BiocManager::install("bamsignals", suppressUpdates=T)
#fetch current normR version from github
install.packages("devtools")
require(devtools)
devtools::install_github("your-highness/normr")

Usage

See the vignette for a toy example on normR usage. The documentation of routines can be accessed from with R with ?.

Use cases

• ChIP-seq normalization / enrichment calling with an Input experiment (Whole Cell Extract, H3/IgG ChIP-seq)

• ChIP-seq differential enrichment calling for two different antigens in same sample population

• ChIP-seq identification of enrichment regimes to investigate on sample heterogeneity

• RNA-seq differential expression calling

• ChIP-seq differential enrichment calling in two different samples (be aware of CNVs!)

• CNV identification

Useful links

Be sure to check out the following amazing github projects for your upcoming NGS magic:

bamsignals - Efficient Counting in Indexed Bam Files for Single End and Paired End NGS Data

EpicSeg - Chromatin Segmentation Based on a Probabilistic Multinomial Model for Read Counts

kfoots - Fit Multivariate Discrete Probability Distributions to Count Data

deepTools - User-Friendly Tools for Normalization and Visualization of Deep-Sequencing Data