What Parkinson’s Cannot Hide from Geometry

Abstract

We report a 386-node geometric decomposition of the alpha-synuclein aggregation problem in Parkinson’s disease using the Omuo Genesis Engine’s Reverse Genesis capability (Mode U). Rather than proposing drug candidates, the analysis asks: what must be structurally true before any solution to this problem is possible? The engine decomposed the full paradox — intrinsic disorder, transient toxic oligomers, prion-like propagation overlapping with normal neuronal signalling — into 8 root prerequisite paths and 328 leaf atoms. Three findings diverge from the standard pharmaceutical playbook: (1) seedproof membrane handshake — intercepting prion-like templating at the vesicle fusion interface by inverting the direction of causal propagation; (2) pathological shape memory — targeting the folding trajectory and rule-propagation mechanism rather than the resulting conformation; (3) kinetic trap design with path-dependent frustration — engineering molecules that exploit the protein’s own folding history to make pathological intermediates trap themselves. The analysis independently confirmed liquid–liquid phase separation as a critical intervention layer through purely geometric reasoning, and produced the highest-confidence structural atoms (12%) in the “hidden scaffold removal” branch — suggesting the misfolded conformation is held in place by structural supports that have not yet been identified.

1. The Problem

Alpha-synuclein is a 140-amino-acid presynaptic protein that facilitates synaptic vesicle trafficking and SNARE complex assembly. In Parkinson’s disease, it misfolds into toxic oligomers and eventually insoluble fibrils that comprise Lewy bodies. Approximately 10 million people worldwide live with Parkinson’s. There is no approved disease-modifying therapy.

The therapeutic challenge is not a lack of molecular knowledge. It is a structural paradox with three interlocking components.

The protein has no stable shape. Alpha-synuclein is intrinsically disordered — it has no fixed tertiary structure to target with conventional drug design. It exists as a fluctuating ensemble of conformations. Designing a drug for a target that has no defined shape is like trying to catch a specific wave in the ocean.

The obvious target may be the wrong target. The visible Lewy body fibrils are the signature pathology, but the toxic species may be soluble oligomers — transient, heterogeneous, and difficult to isolate. The damage may be done before aggregates form, meaning therapies targeting aggregates arrive too late.

Blocking the spread risks blocking survival. Misfolded alpha-synuclein propagates between neurons in a prion-like manner through mechanisms that overlap with normal exosomal communication. Blocking propagation risks blocking the intercellular signalling that neurons depend on for survival.

Every major therapeutic modality has been attempted and failed: monoclonal antibodies against aggregated forms (prasinezumab, cinpanemab), small molecules aimed at preventing misfolding, immunotherapy targeting extracellular spread, and genetic knockdown strategies that risk eliminating the protein’s essential membrane-binding and vesicle-fusion functions.

We applied the Omuo Genesis Engine’s Reverse Genesis capability to this problem — not to propose a drug, but to decompose the structural prerequisites that must be satisfied before any solution becomes possible.

2. Method

2.1 Reverse Genesis (Mode U)

The Genesis Engine’s Mode U performs recursive unbinding: given a goal state, it identifies the structural prerequisites that must be satisfied before that goal becomes achievable. Each prerequisite is itself decomposed, producing a tree of atomic structural requirements. The engine encodes concepts as complex phasor vectors constrained to an E8 lattice (240 root vectors projected into 1,024 dimensions) and decomposes through algebraic unbinding — the inverse of the engine’s standard binding operation.

The analysis was configured as a recursive unbind tree with depth 3 and branching factor 8, producing 584 vocoder calls. The cloud vocoder (DeepSeek) was used for concept naming. The engine operated in naked unbind mode — E8 axes only, no pre-loaded manifold — ensuring all findings emerged from geometric decomposition alone rather than from prior domain knowledge.

2.2 Goal Formulation

The goal prompt encoded the full paradox structure: the intrinsic disorder of the target, the oligomer-versus-fibril ambiguity, the prion-like spreading mechanism, the overlap between pathological and physiological function, and the three decades of therapeutic failure. The prompt was designed to prevent the decomposition from collapsing into any single known approach.

3. What the Engine Found

The engine produced 386 structural nodes (328 leaf atoms) across 8 root decomposition paths.

Three of these findings diverge from the standard pharmaceutical playbook. Two confirm known approaches with added structural specificity. One independently confirms frontier research. We present each category in turn.

4. Three Novel Reframings

4.1 Seedproof Membrane Handshake

“The point where a boundary’s integrity depends not on resisting external pressure, but on selectively reversing the direction of causality across its surface.” — Causal Inversion Threshold

Nobody in the alpha-synuclein field is framing the propagation problem this way. The engine’s structural decomposition converges on a specific intervention point: not the misfolded seed in transit, and not the healthy protein waiting to be corrupted, but the exact moment of contact during vesicle fusion — the handshake itself.

The sub-branch “Causal Inversion Threshold” (9% confidence) makes a precise structural claim: the membrane’s integrity against pathological templating depends on reversing the direction of causality at the interface. In molecular terms: instead of blocking the seed from reaching healthy protein, make the healthy protein’s membrane interaction reject the pathological template from the inside out.

The deepest leaves — “Loop Collapse Point,” “Phase Lock Breakaway,” “Mirror Symmetry Shatter” (all 10%) — describe the structural prerequisites for this inversion: the feedback system must lose its ability to self-correct before cause and effect can exchange roles.

This maps to a testable molecular design principle: a molecule that modifies the vesicle fusion interface to make pathological templating thermodynamically unfavourable, without altering the fusion mechanics that alpha-synuclein normally facilitates.

4.2 Pathological Shape Memory

“The system’s memory of a shape becomes locked not by external force, but by its own internal processes continuously validating and cementing a flawed initial configuration.” — Self-reinforcing error propagation

The engine reframes the toxic oligomer problem entirely. Rather than characterising the toxic species by its static conformation, it identifies the folding trajectory as the pathological signature.

The key sub-branch — “Recursive constraint seeding” — generated the highest-confidence cluster in the entire tree (10–11%). The structural claim: prion-like propagation works not because the seed templates a shape, but because it recursively rewrites the folding rules of whatever it contacts. Sub-leaves “Self-embedding boundary” (11%), “Fractal Seed Asymmetry” (11%), and “Constraint Echo Chamber” (11%) describe a mechanism where initial misfolding conditions are copied and amplified inward at every scale, creating layers of the same restriction that become progressively harder to escape.

Each corrupted protein does not just adopt a bad conformation — it becomes a new rule-writer.

If this structural model is accurate, the therapeutic target shifts from conformational states to the rule-propagation mechanism itself — a different class of intervention than any currently pursued. This extends recent work on strain-specific propagation but goes further: the intervention point is not the strain but the rule-rewriting process that makes strains self-perpetuating.

4.3 Kinetic Trap Design

“The trap’s effectiveness relies on the target’s own momentum and previous choices making escape geometrically impossible.” — Path-dependent frustration

Rather than preventing misfolding or clearing aggregates, the engine identifies the possibility of engineering molecules that make early misfolding intermediates trap themselves.

The structural prerequisites include “Frozen choice topology” (10%), “Knots that can’t untie” (10%), and “Self-reinforcing asymmetry” (10%) — describing a mechanism where the protein’s own folding history makes the pathological pathway a geometric dead end. The sub-branch “Asymmetric escape routes” (9%) specifies that the trap must create one-way flow: easy to enter the functional state, geometrically impossible to proceed toward aggregation. “Forced sequential unlocking” (9%) adds that the pathological pathway requires states to be visited in a strict, non-optional order — suggesting that disrupting any single step in the sequence could collapse the entire aggregation cascade.

The concept of path-dependent frustration is well-studied in materials science and condensed matter physics. Its application to alpha-synuclein folding intermediates appears to be novel.

5. Frontier Confirmation

Phase boundary navigation independently arrived at liquid–liquid phase separation (LLPS) as a critical intervention layer. The engine described operating “at the precise edge where the protein’s functional liquid-like condensates transition into pathological solid aggregates.” This aligns with frontier research from the early 2020s demonstrating that alpha-synuclein undergoes LLPS prior to aggregation [4]. The engine reached this conclusion through purely geometric reasoning without prior encoding of LLPS literature.

Dynamic ensemble targeting and selective phase chaperone correspond to established research directions. The engine’s contribution here is structural specificity in the sub-decompositions: “Coherence before control” and “Field before agents” describe a prerequisite ordering — establish the conformational landscape before attempting to modulate individual conformations — that, while intuitive, is often violated in practice.

6. The Highest-Confidence Finding

The highest individual confidence scores in the entire tree appeared in the “Hidden Scaffold Removal” branch under Functional Misfold Chaperone, where “Unseen Binding Force” and “Symmetry Breaking Event” reached 12% — the maximum in the entire decomposition.

The structural claim: the misfolded conformation is held in place by invisible structural supports that have not yet been identified. “Unwritten Assembly Rules” (12%) and “Unspoken Assembly Order” (12%) describe implicit protocols that govern how the misfolded scaffold maintains itself. Removing these supports — rather than attacking the misfolded shape directly — would allow the protein to self-correct.

This is a structurally different therapeutic logic from anything currently pursued. Current approaches target the product of misfolding. The scaffold removal branch targets the infrastructure that maintains misfolding.

7. Calibration

We estimate that approximately 30–40% of the high-confidence findings are genuinely non-obvious to domain experts. Approximately 40–50% are correct but known. Approximately 10–20% are geometric artefacts. This calibration is consistent with the engine’s performance across other domains.

8. The Unified Structural Picture

The three novel branches converge on a common theme: the pathology is in the process, not the product.

Seedproof membrane handshake targets the propagation event. Pathological shape memory targets the folding trajectory. Kinetic trap design targets the intermediate states. All three shift attention from what the toxic species is to what it does — and more specifically, to the structural conditions that make its activity possible.

This process-over-product reframing may explain why three decades of structure-targeting approaches have failed: they were aiming at the output of a process rather than the process itself.

The structural prerequisites for a solution involve not the shape of the toxic species, but the geometric conditions under which shape propagation becomes possible — and the conditions under which it can be made impossible.

The connection to the engine’s cancer analysis is direct. In the cancer run, the terminal was Dissipative Symmetry Breaking — cancer as a topological phase transition that cannot be reversed from within the malignant phase. In the alpha-synuclein decomposition, the same principle appears: prion-like propagation is a symmetry-breaking cascade that cannot be reversed by targeting the broken state. Both analyses converge on the same structural insight: cure requires intervening in the breaking process, not in the broken product.

9. Limitations

The Genesis Engine does not encode biochemical knowledge. Its decompositions operate on geometric structural relationships. Concept naming is performed by a language model vocoder, which introduces interpretation bias. Confidence percentages reflect geometric binding depth, not empirical probability. The analysis was performed in naked unbind mode (no pre-loaded manifold), meaning the engine had no prior domain context.

All findings require experimental validation by domain experts before therapeutic relevance can be assessed. We report the geometry. Experimental neuroscience decides what it means.

Run Statistics

References

[1] Spillantini, M. G. et al. (1997). “α-Synuclein in Lewy bodies.” Nature, 388(6645), 839–840.
[2] Luk, K. C. et al. (2012). “Pathological α-synuclein transmission initiates Parkinson-like neurodegeneration in nontransgenic mice.” Science, 338(6109), 949–953.
[3] Pagano, G. et al. (2022). “Trial of prasinezumab in early-stage Parkinson’s disease.” New England Journal of Medicine, 387(5), 421–432.
[4] Ray, S. et al. (2020). “α-Synuclein aggregation nucleates through liquid–liquid phase separation.” Nature Chemistry, 12(8), 705–716.
[5] Fusco, G. et al. (2017). “Structural basis of membrane disruption and cellular toxicity by α-synuclein oligomers.” Science, 358(6369), 1440–1443.
[6] Brundin, P. et al. (2010). “Prion-like transmission of protein aggregates in neurodegenerative diseases.” Nature Reviews Molecular Cell Biology, 11(4), 301–307.
[7] Mekšriūnas, G. (2026). “Recursive Self-Deepening: How an E8-Constrained Geometric Lattice Discovers Its Own Structural Convergence.” Omuo Research, Zenodo.
[8] Mekšriūnas, G. (2026). “The Shape of the Hardest Problems: Geometric Unification of Structural Convergence Across the Millennium Prize Landscape.” Omuo Research, Zenodo.
[9] Mekšriūnas, G. (2026). “The Topology of Malignancy: A Geometric Theory of Cancer.” Omuo Research.

© 2026 Omou Systems, MB. All rights reserved. omuo.io — Vilnius, Lithuania