## iscream <a href="https://huishenlab.github.io/iscream/"><img src="man/figures/logo.png" align="right" height="138" style="float:right; height:138px;"/></a> <!-- badges: start --> [](https://github.com/huishenlab/iscream/actions/workflows/check-bioc.yml) [](https://app.codecov.io/gh/huishenlab/iscream) <!-- badges: end --> iscream aims to efficiently read data from any BED file into formats usable by other packages. Using [htslib](https://www.htslib.org/), iscream can query genomic regions like [tabix](https://en.wikipedia.org/wiki/Tabix), summarize the queried data and make matrices, with specific support for WGBS BED files aligned by [BISCUIT](https://huishenlab.github.io/biscuit/), [Bismark](https://www.bioinformatics.babraham.ac.uk/projects/bismark/) and [BSBolt](https://bsbolt.readthedocs.io/en/latest/). Analysis and visualization of Whole Genome Bisulfite Sequencing (WGBS)[^1] data requires reading aligned sequencing data into formats that existing packages like [BSseq](https://bioconductor.org/packages/devel/bioc/html/bsseq.html) and [scMET](https://github.com/andreaskapou/scMET) can analyze. Getting the data from on-disk BED files to a matrix of methylation values can be difficult because, with nearly 30 million CpGs, WGBS data can be quite large. iscream makes importing WGBS data for targeted exploration and analysis faster and more memory efficient. [^1]: The name iscream comes from "*Integrating Single-Cell Results for Exploring and Analyzing Methylation*" as it was originally developed to read BED files from WGBS. It was then generalized to work with any BED file. ## Dependencies ### *htslib* >= 1.17 *iscream* depends on the *htslib* header files. These may be installed with your package manager: - ubuntu/debian: `libhts-dev` - fedora/RHEL: `htslib-devel` - brew: `htslib` - nixpkgs: `htslib` - conda: `bioconda::htslib` or built manually: <https://www.htslib.org/download/>. *iscream* can use *Rhtslib* as the htslib source but we recommend installing htslib with libdeflate support for optimal performance - see [`vignette("htslib")`](https://huishenlab.github.io/iscream/articles/htslib.html) for more information. The header files may also be found among your HPC modules - make sure the `PKG_CONFIG_PATH` environment variable includes the `pkgconfig` location for your installation of *htslib*. You can verify that the *htslib* development libraries are installed with `pkg-config`: ```bash # set path if necessary export PKG_CONFIG_PATH=[path to htslib installation] # verify that htslib can be found pkg-config --cflags --libs htslib ``` #### *tabix* Some *htslib* installations do not include the *tabix* executable (on Ubuntu you need to install both *libhts-dev* and *tabix*). *iscream* will work without *tabix*, but the `tabix()` function will be faster if the executable is installed. ### GCC >= 9.4.0 GNU GCC must be installed for OpenMP support. This is usually installed by default on Linux systems, but may need to be manually installed on MacOS to use iscream with multiple threads[^2]. [^2]: Using OpenMP is also possible with Clang on MacOS (<https://mac.r-project.org/openmp/>) but installing GCC with Homebrew may be easier (<https://formulae.brew.sh/formula/gcc>). ## Installation ### Bioconductor ```r if (!requireNamespace("BiocManager", quietly = TRUE)) install.packages("BiocManager") BiocManager::install("iscream") ``` ### GitHub You can install the development version from Github by cloning the repo and running ```bash git clone https://github.com/huishenlab/iscream R CMD INSTALL iscream ``` You can also use the R [`devtools`](https://devtools.r-lib.org/) package: ```r devtools::install_github("huishenlab/iscream") ``` or [`pak`](https://pak.r-lib.org/): ```r pak::pkg_install("huishenlab/iscream") ``` ### Usage See the [quick start guide](https://huishenlab.github.io/iscream/articles/iscream.html) for an overview of iscream's functionality and the [function reference](https://huishenlab.github.io/iscream/reference/) for all available functions. Bug reports may be submitted through [GitHub issues](https://github.com/huishenlab/iscream/issues).