p53 mutational mechanisms

Aug 2025 Labs/p53. 16 Sep 2026 README: further testing did not support the findings; lab materials withdrawn. Do not rely on them as validation. Restricted/Corrected.

Applied · Applied · Corrected · 2026-09-19 · compiled 2026-10-10 · 3 min read · 2 claims

Package: Labs/p53 · JSON · Cite

protein · p53 · Y220C · R248Q · PhiKan083

On this page

Thesis in the contract
Twelve-run scoreboard
Two failure modes, one globule

Causal stack

Labs/p53 is a 12-run coarse-grained campaign on the p53 DNA-binding domain, dated 22–23 Aug 2025 in the deliverable folders. The machine-readable contract is findings.json v2.2. The human manuscript (ReadMe.pd) is co-attributed to Gemini, ChatGPT, Grok, and Aaron Spradlin. There is no load-bearing X field note for this campaign in the current compile — it lives on GitHub. Cite the folder and the science contract, not a tweet.

What the campaign is allowed to say

Wild-type also collapses to a compact non-native globule (best RMSD 26.6 Å). The result is mechanism differentiation among misfolded globules, not native p53 folding. Contact-anneal + PhiKan083 improved Y220C from ~25 Å to 18.9 Å inside this force field. That is not a pharmacological proof and not a crystal-structure rescue.

Thesis in the contract

Across the UQCF-GEM campaign, p53-Y220C predominantly fails via a kinetic collapse trap, while p53-R248Q failure is electrostatic in nature. A protocol-level intervention—contact anneal with PhiKan083—achieves meaningful improvement (≤ 20 Å) relative to ~25 Å baselines. The campaign ultimately revealed a force-field limitation in the ligand model.

Twelve-run scoreboard

Run	Purpose	Claim type	Best RMSD (Å)	Contract verdict
1	R248Q baseline	electrostatic_failure	26.6	CONFIRMED (salt-bridge collapse; median 1 vs WT 7)
2	WT baseline	control_informative	26.6	CONFIRMED (misfolded globule)
3	Y220C coil-start	kinetic_trap	25.4	CONFIRMED
4	Y220C + L221F	negative_result	26.1	CONFIRMED (genetic rescue fails)
5	PhiKan083 native-start artifact	control_artifact	~0.0	CONFIRMED as artifact
6	PhiKan083 HERO control	ligand_partial	19.5	CONFIRMED (partial)
7	PhiKan083 HERO v1	ligand_partial	19.5	CONFIRMED (partial)
8	PhiKan083 native-start validation	native_validation_limit	4.0	CONFIRMED (drift; ligand-model limit)
9	Contact anneal + PhiKan083	rescue_protocol	18.9	INDETERMINATE in this compile (ContactAnneal folder missing)
10	PhiKan083 rescue v1 pilot	ligand_partial	25.7	CONFIRMED (no rescue)
11	PhiKan083 timing sweep	ligand_thermodynamic	18.9	CONFIRMED (stabilizer, not kinetic guide)
12	Y220F rescue attempt	negative_result	26.1	CONFIRMED (genetic rescue fails)

Manuscript numbers from Labs/p53/ReadMe.pd Table 1. This compile’s deterministic pass (7 Sep 2026) reproduced the same order from diagnostics CSVs (R248Q/WT tail-median 26.71 Å vs 26.6; Y220C coil 25.32 vs 25.4; native-start 3.82 vs 4.0; Run 11 18.86 vs 18.9): 11 CONFIRMED, 0 DEVIATION, 1 INDETERMINATE. Run 9’s ContactAnneal folder is not in the public tree — do not treat the 18.9 Å rescue as independently re-read. Start the engine smoke test on MECP2, not here.

Two failure modes, one globule

Final RMSD alone does not separate the mutants from wild-type. The mechanistic split is in the diagnostics. R248Q (DNA-contact mutant) loses the salt-bridge network (median 1.0 versus WT 7.0) while collapsing to the same 26.6 Å globule — electrostatic_failure. Y220C (pocket mutant) is diagnosed as a kinetic trap: genetic pocket patches Y220F and L221F do not improve RMSD; a gentler contact-anneal protocol plus the pharmacological chaperone PhiKan083 does, inside this model. The timing sweep says the ligand acts as a thermodynamic stabilizer of the collapsed state, not a kinetic folding guide.

Model: coarse-grained, backbone-centric UQCF-GEM. Terms named in the manuscript: steric, directional H-bond, Ramachandran, Debye–Hückel electrostatics, tunable contact springs.
Protocol: extended coil start unless marked native-start. pH 7.4, ionic strength 150 mM. k_hbond = 16.0, k_contact_springs = 0.05 as the posted sweet spot.
Negative results are first-class: L221F and Y220F fail; the native-start ligand run drifts to 4.0 Å and is filed as a force-field limit, not hidden.
PNG and PDB artifacts in Labs/ are Git LFS pointers in a plain clone. Markdown reports, raw CSVs, and findings.json are ordinary files.

Restricted

Inside this coarse-grained campaign, R248Q and Y220C reach similar misfolded globules by different routes (electrostatic network loss vs kinetic trap), and a contact-anneal + PhiKan083 protocol improved Y220C RMSD relative to ~25 Å baselines. Not native folding, not a crystal rescue, not a cancer therapy, not a TOE result.

Not claimed

This atlas does not claim that PhiKan083 rescues p53-Y220C in cells, or that UQCF-GEM replaces all-atom MD. The native-start drift is the campaign’s own limit.

Previous MECP2 Rett MBD–DNA campaign Next Frozen v9 protein bridge (1UAO / 1L2Y)

Causal stack

On this page

Thesis in the contract
Twelve-run scoreboard
Two failure modes, one globule