ompBAM

C++ Library for OpenMP-based multi-threaded sequential profiling of Binary Alignment Map (BAM) files

Description

Although various tools for multi-threaded BAM processing (samtools, sambamba) are available, currently there are none available to the R/Bioconductor environment. Parallelism in R is achieved using BiocParallel or related packages. Typical alternatives include processing multiple BAM files, each using a single core. Although easy to set up, memory-intensive applications limit such approaches. ompBAM is a header-only C++ library based on OpenMP, allowing developers to implement OpenMP-based parallelism for the sequential reading of BAM files. ompBAM makes it easy by handling all the parallel file-access and bgzf decompression, allowing developers to focus on multi-threaded handling of individual reads.

Installation

On current R (>= 4.0.0)

• Development version from Github:

  library("devtools")
  install_github("alexchwong/ompBAM", dependencies=TRUE, build_vignettes = TRUE)

Documentation

Access the ompBAM-API-Docs via its included vignette. This includes:

• How to set up a new package R-project, ready-to-compile with ompBAM, as well as a ‘Hello World’ equivalent example function of the ‘idxstats’ function to demonstrate ompBAM
• A step-by-step guide of how the idxstats function implemented in the example code is constructed
• Detailed documentation of the pbam_in and pbam1_t objects that comprise ompBAM.

browseVignettes("ompBAM")