Intro

This document includes the analysis of 8 ensemble boolean datasets, each one corresponding to a different cell line input. In each case, we have a set of boolean models that were produced by the druglogics software pipeline module Gitsbe (\(2500\) simulations, choosing 3 best models out of each one, resulting thus in \(7500\) models in total for each cell line), using as input an automated topology (generated by the module Atopo) and a specific activity state profile for each cell line that the models were trained to. In the following chapters, we include more details for the inputs used and the analysis performed for each cell line to identify possible biomarkers. In the last chapter, we conlude the analysis by comparing the results between the 8 cell lines and the random model analysis (where the models were trained to fit a proliferation state) as well.

The main purpose of this analysis is two-fold:

  1. Show how to use the functions of the emba R package (Zobolas 2019a) on a real dataset. This means that this analysis can be used as a long-form documentation (vignette) for the R package itself. Also, functions from the usefun R package (Zobolas 2019b) are used in this analysis.
  2. Investigate if there are common biomarkers between the 8 cell lines

Note

In the beginning, this analysis included various functions that I had implemented in multiple R scripts. Only later I thought to write a more generic R package (modularizing, extending, documenting and testing the code I already had) that would help me perform this kind of analysis on new boolean model datasets, and which would be easy to use and extend with more features. So, even though all the functions now used in this book are offered/exported by the emba package, I have written additional, more generic ones that execute (almost) the full analysis for each cell line in a simple line (the graphs you can generate from the output data). To refer to these generic analysis functions, check the corresponding documentation:

  • ?emba::biomarker_tp_analysis
  • ?emba::biomarker_mcc_analysis
  • ?emba::biomarker_synergy_analysis

An example use of two of these functions is on the Random Model Analysis chapter.

The version of the emba package used in this analysis is 0.1.1. For a repository of different versions see here.