Relatively Human | Season 1, Episode 6: The Cell That Remembers
You started as a single fertilized egg holding roughly 750 megabytes of genetic data. Today, you are a staggering constellation of 37 trillion cells. How does a biological system process information and actually gain complexity, seemingly violating the fundamental Data Processing Inequality?
The answer lies in a paradigm-shifting realization sweeping across modern science: biological systems are not "memoryless" or Markovian. They remember.
In this episode of Relatively Human, we explore the awe-inspiring convergence of four unrelated biomedical fields—cancer research, developmental biology, neuroscience, and protein folding. We discover how researchers, hitting the limits of traditional biology, are independently borrowing mathematical frameworks from 1960s gas physics (the Mori-Zwanzig projection) and 1990s communication theory (Directed Information) to decode the hidden histories of cells.
In this episode, we unravel:
• The Cancer Paradox: How cancer cell populations use non-Markovian memory to resist chemotherapy without acquiring new genetic mutations. We also dive into how memory kernels govern the mechanics and collective migration of tumor cells.
• The Embryo's Journey: Why the famous Waddington landscape of embryonic development is getting a mathematical update to account for lineage memory and the lasting impact of morphogen signals.
• Untangling the Brain and Genes: How Directed Information and Transfer Entropy are replacing old correlation models to map the true causal highways inside dense neural circuits and intricate gene regulatory networks.
• Compressing the Impossible: How AI and Memory Kernel Minimization Neural Networks (MEMnets) are making impossible all-atom protein folding simulations a reality by using deep learning to discover exactly which molecular details can be safely "forgotten".
Every cell in your body carries a history-dependent story that no current molecular measurement can fully read. But for the first time, we have the mathematics to ask the right questions—and the CRISPR-tracing technologies to listen for the answers.
--------------------------------------------------------------------------------
Episode Reference List:
1. Information Theory & Genetics: Tkačik, G., & Gregor, T. "The many bits of positional information." | Tkačik, G., & Walczak, A. M. "Information transmission in genetic regulatory networks: a review."
2. Cancer & Cell Migration: Stichel, D., et al. "An individual-based model for collective cancer cell migration explains speed dynamics and phenotype variability in response to growth factors." | Lin, S.-Z., et al. "Dynamic Migration Modes of Collective Cells."
3. Protein Folding & MEMnets: Liu, B., et al. "Memory Kernel Minimization Based Neural Networks for Discovering Slow Collective Variables of Biomolecular Dynamics."
4. Directed Information & Causal Inference: Tsur, D., et al. "Directed Information: Estimation, Optimization and Applications in Communications and Causality." | Kornai, D., et al. "AGM-TE: Approximate Generative Model Estimator of Transfer Entropy for Causal Discovery." | Rahimzamani, A., et al. "Restricted Directed Information for Gene Regulatory Interactions Inference."
5. Cellular Sensing Dynamics: Nandan, A. P. "Dynamical basis of cellular sensing and responsiveness to spatial-temporal signals."
