Deep-iCE: A Deep Learning-Based Framework for Interactome Prediction from Co-elution Profile

Co-fractionation mass spectrometry (CF-MS) has emerged as a powerful approach for cell-wide identification of protein complexes that performs nearly all biological processes and cellular activities. While computational tools based on traditional machine learning is widely applied to analyze protein complexes, it remains unclear if it is the most suitable choice for analyzing CF/MS dataset. Here, we introduce Deep-iCE that employs deep learning approaches for automated scoring of CF/MS data and sophisticated clustering procedure for network inference of underlying complexes.

Installation

To install the development version in R, run:

if(!requireNamespace("devtools", quietly = TRUE)) {
  install.packages("devtools") 
}
devtools::install_github("mrbakhsh/DeepiCE")

Required Software

Although, this is a R package, users will need to install Python as well. In addition, tensorflow and keras should be installed in Python and they should be connected to R, before installing this package.