Graduation Year
2024
Date of Submission
12-2024
Document Type
Campus Only Senior Thesis
Degree Name
Bachelor of Arts
Department
Physics
Reader 1
Sarah Marzen
Reader 2
Adam Landsberg
Terms of Use & License Information
Abstract
We propose an information-theoretic framework to analyze the relationship between memory and prediction in gene regulatory networks (GRNs). Further, by reformulating total predictable information (TPI) using entropy-based metrics, we enable practical estimation in high-dimensional modeling environments. This reformulation supports a Python-based simulation pipeline for exploring GRN dynamics, integrating SBMLtoODEjax for efficient numerical methods and autodiscjax for perturbation-driven analysis.
Utilizing Markov and semi-Markov models to calculate forward and reverse time causal states through their dwell time distributions, the pipeline leverages these features of ϵ-machines to quantify how GRNs memorize and predict. This approach may identify pathways for shifting GRNs between attractor states, with implications for experimental biophysics, developmental biology, and potentially therapeutics.
Recommended Citation
Hutton, Wolfgang, "Memory and the Quantification of Prediction in Transcriptional Network Models" (2024). CMC Senior Theses. 3848.
https://scholarship.claremont.edu/cmc_theses/3848
This thesis is restricted to the Claremont Colleges current faculty, students, and staff.