Roughly 10⁷ developing T cells enter thymic selection each day; on the order of 2% leave the thymus as mature, self-tolerant T cells. The other ∼98% die in place. This is not wastage — it is proofreading, structurally the same discrimination-by-irreversible-step logic as Hopfield's 1974 kinetic proofreading and this corpus's own K-operator threshold, run here across three sequential filters instead of one. This chapter lays out the three stages, states plainly what is an established immunological fact and what is this corpus's own operator-chain framing, and does not blur the two.
Double-positive (CD4⁺CD8⁺) thymocytes are produced in the thymic cortex after successful V(D)J recombination of the T-cell receptor (TCR) locus — the combinatorial step that generates an estimated 10⁷ distinct receptor specificities from roughly 50 V, 25 D, and 6 J gene segments (plus junctional diversity). Each of these receptors is then tested, in sequence, against self-peptide–MHC complexes displayed on thymic epithelial cells. The outcome of that test is not binary tolerance or rejection; it is an affinity measurement with three possible results, and getting the biology of that measurement right is the entire content of this section (Klein et al. 2014; Hogquist, Baldwin & Jameson 2005).
DP thymocyte TCR engages self-peptide–MHC on cortical thymic epithelial cells. No selection yet — this stage only establishes whether contact of sufficient quality occurs at all.
Thymocytes whose TCR fails to engage self-MHC with at least a minimum affinity receive no survival signal and die by neglect — apoptosis by default, the majority outcome at this stage.
Of the survivors, thymocytes whose TCR binds self-peptide–MHC too strongly are deleted (clonal deletion) or diverted to the regulatory T-cell lineage, mainly in the medulla, via AIRE-dependent presentation of tissue-restricted antigens on mTECs.
This corpus's own convention, used throughout Book VI's bio-domain chapters, reads C (Contact), K (threshold/Kommit), F (Fold), U (Unfold/express) as a generic sequence: establish geometric contact, cross an irreversible threshold, compress onto an attractor, express the output. Thymic selection instantiates the first three operators directly, and the window structure in §1's callout is exactly why the third stage is F and not another K: a second monotonic threshold would just be another commitment step, but a window with two boundaries — die below K*, die above K** — is a fold in the technical sense used elsewhere in this corpus (a stability region bounded on two sides, collapsing to a point if the window closes), not a second one-sided cutoff.
| Operator | dm³ role | Thymic selection | Key molecules |
|---|---|---|---|
| C — Contact | Establish geometric/recognition contact | TCR–self-peptide–MHC engagement on cortical epithelium | TCR, MHC-I/II, CD4, CD8 |
| K — Threshold | Irreversible commitment given adequate signal | Positive selection. Affinity ≥ K* required to survive; below it, death by neglect (the default fate). | ZAP-70, ERK signaling intensity/duration |
| F — Fold | Bounded stability region, collapsing at its edges | Negative selection. The surviving affinity window K* ≤ affinity ≤ K** is the fold; crossing the upper edge triggers clonal deletion or Treg diversion. | AIRE, mTECs, dendritic cells, Bim-dependent apoptosis |
A fourth, U (Unfold/express), is the mature single-positive T cell's export to the periphery and its functional repertoire there — outside the thymus proper, and not this chapter's subject.
The headline figures in this chapter's own hero line — roughly 10⁷ thymocytes produced daily, roughly 2% surviving to export — are the widely cited order-of-magnitude estimates in the immunology literature (Klein et al. 2014; Palmer 2003). What is not as cleanly settled in the literature is the precise per-stage split: what fraction dies specifically at positive selection versus specifically at negative selection varies by study, repertoire, and measurement method, and no single agreed number exists to two significant figures for each individual stage. This chapter states the two well-established facts — the large-order input, the small-order output — and does not manufacture false precision for the intermediate split.
The corpus's n-bonacci recurrence ladder — φ (Fibonacci, n=2), η (Tribonacci, n=3), Δ (Tetranacci, n=4), each the unique real root >1 of xⁿ = xⁿ⁻¹+…+x+1, converging monotonically toward the embodiment threshold τ=2 as n grows — is real and independently checkable arithmetic, not tied to any biological measurement:
This chapter's three biological stages (Contact, positive selection, negative selection) sit in the same three-deep sequence as φ, η, Δ sit in the recurrence ladder — three successive steps in a chain that this corpus elsewhere associates with a convergence toward τ=2. That is a structural, naming-level correspondence: three stages, three constants, same chain convention used across this book's chapters. It is not a numerical fit. No published thymic-selection survival ratio is claimed here to equal, approximate, or derive φ, η, or Δ — §3 was explicit that the per-stage percentages aren't even settled precisely enough in the literature to compare against anything. The n-bonacci arithmetic above is checked and real; the biological numbers in §1 and §3 are checked and real; the bridge between them offered in this section is a naming convention, stated as exactly that.
Unlike the differential-equations chapters elsewhere in Book VI, this chapter has no Lean scaffold file yet. The natural formal object — a thymocyte's affinity as a real number required to lie in an open window (K*, K**) — is straightforward to state and is sketched here for the record, not claimed as done:
-- sketch only, not yet a file in this repo
structure Thymocyte where
affinity : ℝ
-- survives iff affinity lies strictly inside the selection window
def survives (Kstar Kstarstar : ℝ) (t : Thymocyte) : Prop :=
Kstar < t.affinity ∧ t.affinity < Kstarstar
-- not proved: existence/nonemptiness of the window for a given TCR repertoire,
-- which is an empirical immunological question, not a theorem
No claim is made that this window's existence is provable from first principles — whether a given repertoire has a nonempty (K*,K**) window for a given self-antigen set is an empirical question about real receptor sequences and real MHC molecules, not a mathematical one. The sketch above states the shape of the object honestly; it formalizes nothing yet.
Klein, L., Kyewski, B., Allen, P.M., Hogquist, K.A. (2014). Positive and negative selection of the T cell repertoire: what thymocytes see (and don't see). Nat. Rev. Immunol. 14, 377–391. — The standard modern review of the three-stage process described in §1–2.
Hogquist, K.A., Baldwin, T.A., Jameson, S.C. (2005). Central tolerance: learning self-control in the thymus. Nat. Rev. Immunol. 5, 772–782. — Affinity-threshold model of positive/negative selection.
Palmer, E. (2003). Negative selection — clearing out the bad apples from the T-cell repertoire. Nat. Rev. Immunol. 3, 383–391. — Negative selection mechanism and the medullary/AIRE component.
Hopfield, J.J. (1974). Kinetic proofreading: a new mechanism for reducing errors in biosynthetic processes requiring high specificity. PNAS 71, 4135–4139. — The general irreversible-threshold discrimination mechanism this chapter's K operator borrows its logic from.
Nogueira Grossi, P. (2026). Principia Orthogona Vol I. G6 LLC. doi:10.5281/zenodo.19117400 — dm³ framework foundations: C, K, F, U operators and the n-bonacci recurrence ladder.
See also: Waddington's Landscape as Lyapunov Potential · Immune Adaptation (Book 3, Ch. 5) · Book VI Index