README file for "Integration of Mechanistic Immunological Knowledge into a Machine Learning Pipeline Increases Clinical Predictive Power"
following is a description of the requirements, installation procedure, and some explanation for usage. All of the referenced files will be found in the included ien_materials.zip which can be downloaded in the "iEN scripts for reproduction of results: ..." section.

The ien_materials .zip contains all referenced materials as well as its own README.txt desribing all the included resources.

System Requirements:
-R 3.5.2
-R packages 'caret' ,'glmnet', and 'parallel'
-all files have been tested on ubuntu 16.04 and R 3.5.2 but are compatible with any OS
-non-standard hardware is not required for the provided software, but a high number of available cpu's wil reduce the runtime considerably 

Installation Guide:
to install the 'iEN' R package navigate to the contained 'iEN_materials' folder and run the following command in R-studio, or a terminal running r:

install.packages("iEN_0.99.0.tar.gz", repos = NULL, type="source")

This process should take no longer than 2 minutes, now the iEN package has been installed and can be accessed via the 'library(iEN)' command.


Demo:
The included reproduction/demo script 'iEN_reproduction_script.R' can be run from R with the working directory set to '.../iEN_materials/' with no additional work necessary. This script will produce an iEN vs EN boxplot from 10 repeated 10-fold Cross-Validated (CV) models for the three clinical datasets used in the paper (longitudinal term pregnancy, pregnancy validation cohort, and chronic periodontitis). These boxplots correspond to figure 4 of the manuscript with a smaller number of iterations, similarly showing iEN outperforming EN.

Runtime for this script will vary greatly depending on the hardware available. For a standard desktop/laptop expect this to run for a few hours. To speed up this process one can reduce the number of CV iterations, variable 'k', or increase the number of cores used, if additional cores are available.


Instructions For Use:
Now that the 'iEN' library is available to you cross-validated models can be generated as shown in the previous demo script; primarily through the 'cv_iEN' method. for more information see the included documentation files 'iEN.pdf' and "iEN-manual.pdf".