Skip to content

README.md

Latest commit

 

History

History
96 lines (74 loc) · 3.9 KB

README.md

File metadata and controls

96 lines (74 loc) · 3.9 KB
Raw

Analyze the accuracy of a linear approximations for model kinetics

Objectives

  • Assess how well an SBML is approximated by one or more linear models based on
    • Consistency of Jacobian
    • Accuracy of reproducing a non-linear simulation
  • Identify "linearity bottleneck" reactions, those that must limit the viability of a linear approximation

Repository Structure

Scripts

The scripts/ directory contains executable tools for running analyses:

Script Description
analyze_biomodels.py Analyzes BioModels for linearity properties
calculate_linear_prediction_scores.py Computes linear prediction scores for models
evaluate_monomial_models.py Evaluates monomial-based kinetic models
make_biomodels_endtime.py Determines simulation end times for BioModels
make_biomodels_timecourse.py Generates timecourse data from BioModels
perturbation_study.py Runs perturbation studies to test model stability
pwla.py Piecewise Linear Approximation utilities
run_parallel_timecourse.sh Shell script for parallelized timecourse simulations

Documentation

The docs/ directory contains technical documentation and design documents:

Document Description
calculating_endtime_using_cv.md Methods for determining simulation end times using coefficient of variation
find_simulation_end_time.md Algorithms for detecting simulation end points
implementing_piecewise_segmentations.md Guide to implementing piecewise linear segmentation
linear_predictions.md Theory and implementation of linear prediction methods
model_based_design.md Model-based design principles and approaches
network_discovery.md Techniques for discovering network structures
perturbation_study.md Documentation for perturbation study methodology
piecewise_system_discovery.md Piecewise linear approximation using cluster-based Jacobian fitting
research_agenda.md Research roadmap and future directions
score.md Scoring methodology documentation
specification.md Project specifications and requirements
technical_notes.md General technical notes and implementation details

Subdirectories:

  • bugs/ — Documented bugs and issues
  • references/ — Reference materials and papers
  • superpowers/ — Advanced features and extensions

Source Code

The src/ directory contains the core Python library:

Module Description
biomodels_iterator.py Iterates over BioModels for analysis
constants.py Project-wide constants and configuration
crn_builder.py Builds Chemical Reaction Network models
dataframe_serializer.py Serialization utilities for DataFrames
linear_predictor.py Core module — Predicts model timecourses using linear ODE approximations with Jacobian-based methods
model.py Base model class for SBML models
multiple_linear_predictor.py Handles multiple simultaneous linear predictions
plot_options.py Visualization and plotting utilities
scaler.py Data scaling utilities
score.py Scoring and evaluation metrics
system_discovery.py Discovers system parameters and structures
timecourse_iterator.py Iterates over timecourse data
timecourse.py Timecourse data handling and processing
trajectory.py Trajectory class for model simulations
utils.py General utility functions

Quick Start

# Install dependencies
pip install -r requirements.txt

# Run analysis on a BioModel
python scripts/analyze_biomodels.py

# Calculate prediction scores
python scripts/calculate_linear_prediction_scores.py

Tuning SystemDiscovery

  • Need sufficient data, maybe in the 1000s.
  • Reduce threshold (sensitivity) to get more sparse coefficients
  • Set includebias = True
  • Don't include boundary species

Analyses

Versions