Theoretical Foundations

Research

The full papers are not published. What follows is the minimum context to understand why ZOT works. Every mechanism maps to real immunology.

McKeithan, 1995
Kinetic Proofreading
T-cell receptors require N sequential phosphorylation steps. Output scales as binding-time^N — exponential amplification discriminating nearly identical ligands.

ZOT Implementation
Each receptor runs an independent KPR cascade. N consecutive above-threshold signals required to fire.
Burnet / Nossal
Negative Selection
In the thymus, T-cells reacting to self-antigens are deleted. Only cells that ignore self survive.

ZOT Implementation
2,000 candidates tested against 60 s of quiet substrate. >10% self-fire = killed. 500 survivors.
François & Altan-Bonnet
Adaptive KPR
Adding negative feedback to KPR creates a sharp threshold where response depends on ligand quality, not quantity.

ZOT Implementation
Feedback boost parameter. After kpr_m consecutive signals, threshold rises — filtering marginal signals.
Altan-Bonnet & Germain
ERK/SHP-1 Feedback
Positive ERK + negative SHP-1 feedback create a bistable switch. The cell commits fully or not at all.

ZOT Implementation
Binary quorum decision. Regulatory suppression dampens response when threat resolves.
Voisinne et al., 2022
Multi-Step Kinase
ZAP70 activation involves multiple phosphorylation steps. Each step is an additional proofreading checkpoint.

ZOT Implementation
Variable cascade depths (kpr_n = 2–6). Fast-acting catch spikes. Slow-acting catch sustained drains.
Mayer et al.
Self/Nonself Distributions
Self and nonself distributions are nearly identical. No single threshold works — population diversity is required.

ZOT Implementation
500 receptors, different sensor pairs, weights, windows. Darwinian selection keeps what works.
Proprietary

What is not published

The theoretical architecture spans 15 internal research papers covering the existence thesis, digital organism theory, encoding problem, ZOT format specification, and the full cell biology. These remain proprietary.

What is open source is the implementation — the organism, the probes, the signal processing, the evolution engine, and the benchmark framework. The code is the proof.

Run it on your machine