Code to generate data for and to plot figures for the paper "Hierarchical optimization of biochemical networks"
by Viswan, Tribut, Gasparyan, Radulescu and Bhalla.
It is on bioRxiv at doi: https://doi.org/10.1101/2024.08.06.606818
while it goes through the review process.
The HOSS project is a framework for building data-driven pipelines for large systems optimizations, so that it is clear how the data led to the choice of parameters for the model. HOSS also introduces new optimization methodology for this class of problems. This paper describes how HOSS works. HOSS is at https://github.com/upibhalla/HOSS
This file and the files in this repository are licensed under GPL v3 or later
Current version of paper figs is 1.0 Current release of HOSS is 2.0
To install and run these scripts you will typically need to have the following packages installed:
Python3 Scipy MOOSE HillTau FindSim HOSS
To install MOOSE and HillTau you will need the C++ build environment as well as pybind11.
=============================================================================
This repository is organized by figure number. There are subdirectories named according to Fig_ where n is the figure number. The mapping to the figure numbering in the paper is as follows:
| Paper Figure number | Sub-directory | Generate figure | Plot Figure |
|---|---|---|---|
| 3 | Fig1_example_expts | None | fig1.py |
| 5 | Fig3_Expt_param_stats | None | plotFig3.py |
| 7 | Fig4_flat | run_opt.csh | plotFig4.py |
| 8 | Fig5_hoss | run_opt.csh | plotFig5.py |
| 9 | Fig6_initScram | run_opt.csh | plotFig6.py |
| 10 | Fig7_hossMC | run_opt.csh | plotFig7.py |
For Figure 7 and 8 the figure generation runs on 1 to 16 cores and can be completed in a couple of hours on a laptop. For Figure 9 and 10 the figure generation code runs on ~100 cores on a shared-memory server and may take a few hours.
The figure plotting code runs in a few seconds on a laptop.
There are also several subdirectories containing common files used by several of the figures.
All the HOSS configuration files used for different pipeline in the
paper.
Experiment files in FindSim.json format, applicable to the b2AR
optimization.
Experiment files in FindSim.json format, applicable to the EGFR
optimization.
Starting model files in HillTau (JSON) format, and .g and .sbml
formats for ODE models.
Map files used by FindSim and HOSS to map between experiment
naming of entities, and the naming used in the models.