Methimpute implements a powerful HMM-based binomial test for methylation status calling. Besides improved accuracy over the classical binomial test, the HMM allows imputation of the methylation status of all cytosines in the genome. It achieves this by borrowing information from neighboring covered cytosines. The confidence in the methylation status call is reported as well. Methimpute also outputs context-specific conversion rates, which might be used to optimize the experimental procedure.
For the exact workings of methimpute we refer the interested reader to our publication TODO. A tutorial is available here.
To install the development version from Github, follow the steps given below. The installation has only been tested on Ubuntu so far, if you need to install on Windows or Mac additional steps might be necessary (e.g. installation of Rtools from https://cran.r-project.org/bin/windows/Rtools/)
-
Install a recent version of R (>=3.3.0) from https://www.r-project.org/
-
Optional: For ease of use, install Rstudio from https://www.rstudio.com/
-
Open R and install all dependencies. Please ensure that you have writing permissions to install packages. Execute the following lines one by one:
install.packages("devtools")
source("http://bioconductor.org/biocLite.R")
biocLite(c("GenomicRanges"))
library(devtools)
install_github("ataudt/methimpute")
Please refer to the vignette for a tutorial.
If you encounter errors of any kind, please file an issue here. I will try to react within one day.