Chaos and Complexity in Cancer

Cancer as a dynamical system: complexity, entropy, and transcriptomic reorganization

Author

Ali M. Al-Timimi, PhD

Overview

Cancer is often described as a disease of increasing disorder.
This project tests whether that intuition is actually correct.

This site presents a reconstruction and extension of my doctoral dissertation, Chaos and Complexity in Cancer (George Mason University, 2004), using a computational framework that treats cancer as a dynamical system.

Rather than assuming uniform increases in randomness, the analysis examines how complexity, entropy, and regulatory structure are reorganized as tissue transitions from normal to malignant states.

Across cancer types, the results suggest a recurring pattern of regulatory simplification accompanied by selective consolidation of functional pathways, rather than simple increases in disorder.

Central Question

How is biological complexity reorganized during tumorigenesis?

This project evaluates whether cancer exhibits:

  • chaotic dynamics (increased heterogeneity and disorder)
  • anti-chaotic dynamics (constraint, clonality, reduced variability)
  • or a context-dependent interplay between both

Example: BLAD/TCC

A representative comparison (bladder normal vs transitional cell carcinoma) illustrates the type of structure detected.

  • Complexity: net decrease with localized functional gains
  • Entropy: heterogeneous (coexisting chaotic and anti-chaotic behavior)
  • Structure: selective organization in functional pathways

Tumorigenesis appears as a reorganization of biological complexity,
involving loss of regulatory flexibility alongside
consolidation of functional and process-specific pathways.

View full report →

Project Structure

This site is organized into:

  • Theory — conceptual framework for complexity in biological systems
  • Methods — data, preprocessing, and analytical design
  • Results and Discussion — comparison-specific reports across cancer types
  • Conclusions — synthesis of observed patterns

Conceptual Perspective

Cells are modeled as high-dimensional dynamical systems with:

  • nonlinear regulatory interactions
  • multiple attractor states
  • sensitivity to perturbation

Within this framework, cancer is interpreted as a state transition process in gene regulatory space.

Rather than uniform chaos, different cancers exhibit different balances of:

  • destabilization (chaotic behavior)
  • constraint (anti-chaotic, structured states)

Data

The analysis uses the Ramaswamy et al. (2001) dataset:

  • GEO accession: GSE68928
  • Affymetrix microarrays (~16,000 genes)

Raw CEL files are processed within this project, enabling full control over normalization, annotation, and filtering.

The dataset is reorganized into paired normal vs tumor comparisons across multiple cancer categories.

Computational Framework

The analysis is implemented as a three-stage system:

  1. Preprocessing — normalization, metadata integration, annotation
  2. Analysis — comparison mapping, filtering, complexity and entropy computation
  3. Reporting — structured summaries and Quarto-generated reports

Future Directions

This framework is being extended using:

  • latent-space modeling
  • variational autoencoders (VAE)
  • geometric representations of biological state

Acknowledgments

This work extends my doctoral dissertation, “Chaos and Complexity in Cancer” (George Mason University, 2004). I am indebted to:

  • Curtis Jamison, my dissertation director
  • Harold J. Morowitz, whose work on complexity in living systems shaped the conceptual framework of this research