Research

Correcting Protein Misfolding at the Molecular Level

A computational platform that reads the structural "grammar" of protein misfolding and identifies food-derived compounds capable of correcting it. Validated in rhodopsin, now extending to neurodegenerative disease.

Protein misfolding outpaces the cell's corrective machinery.

Proteins do not always fold correctly. Aging, genetic mutations, and environmental exposures can force a protein into an aberrant conformation. When this occurs, misfolded copies tend to aggregate, forming toxic oligomeric and fibrillar assemblies. In the brain and retina, this process underlies Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), and Retinitis Pigmentosa.

Collectively, these proteinopathies affect tens of millions of individuals worldwide. For the majority, no approved therapy exists that can halt or reverse the underlying molecular pathology.

"Disordered proteins are like sentences without punctuation." The objective is to determine the rules, the grammar, that govern how these proteins behave, and then leverage that understanding to intervene before irreversible damage occurs.

Three Persistent Gaps in the Field

1

Absence of Early Detection

No reliable method currently exists to detect misfolding at a presymptomatic stage, before neuronal or cellular damage becomes irreversible.

2

Inefficiency of Conventional Drug Discovery

Nearly all therapeutic development relies on de novo synthetic molecules, a process that averages 10 to 15 years with a failure rate exceeding 90%.

3

No Conformational Reversal

No approved therapeutic can restore misfolded proteins to their native conformation. Current interventions, at best, attenuate the rate of aggregation.

Scale of the Problem

Disease Patients (US) Disease-Modifying Tx Misfolded Protein
Alzheimer's ~6.9 million Very limited Amyloid-β, Tau
Parkinson's ~1 million None approved Alpha-synuclein
ALS ~32,000 Limited TDP-43, SOD1
Retinitis Pigmentosa ~100K (1.5M global) None approved Rhodopsin (P23H)

These diseases converge on a shared molecular mechanism: a protein loses its native conformation, and endogenous corrective pathways prove insufficient.

A Computational Platform for
Protein Correction

We have developed and validated a computational platform that decodes the structural "grammar" of protein misfolding and identifies compounds capable of correcting it. The approach has been validated in rhodopsin and is now being extended to neurodegenerative targets.

The Four-Step Pipeline

1

Decode

Compute Hamiltonian DIFF vectors from wild-type vs. mutant structures in the PDB

2

Screen

Score each compound in our library by cosine similarity to the misfolding vector

3

Rank

Identify top-scoring compounds whose signatures best match the misfolding pattern

4

Confirm

Run molecular docking, cross-check against published literature, then move to wet lab

Food-Derived Compound Library

Our compound library is deliberately curated from food-derived and naturally occurring molecules. These compounds possess established safety and toxicology profiles, which may significantly accelerate the regulatory pathway relative to novel synthetic agents. This advantage is particularly relevant for CNS targets, where safety and blood-brain barrier penetration remain the principal translational challenges.

Validated in Retinal Protein Misfolding

Our platform has been validated against a well-characterized protein misfolding target associated with hereditary retinal degeneration, a condition affecting over one million individuals worldwide with no approved disease-modifying intervention.

Validation Summary

Our platform was applied to a large compound library and screened against both wild-type and disease-associated mutant conformations. Negative controls against structurally unrelated protein targets confirmed that the platform exhibits target-specific discrimination rather than generating non-specific hits.

Key Outcomes

Target-Specific Discrimination

Statistically significant separation between target and control protein scores, confirming the platform identifies structure-specific candidates

Independent Corroboration

Multiple top-ranked compounds are independently under investigation by other research groups for related retinal conditions

Wet-Lab Advancement

Lead candidates have advanced to cell-based chaperone assays for experimental validation

Mutant and wild-type scores for the same protein clustered tightly, confirming the method captures mutation-specific structural changes

The compound library is composed of food-derived and naturally occurring molecules with established safety profiles

Beyond Stabilization

Evidence of Conformational Rescue

Preliminary results indicate that our platform can identify food-derived compounds capable of not only stabilizing misfolded proteins but also actively driving conformational rescue toward the native state. These findings extend the potential therapeutic application from prevention to reversal of protein misfolding.

Detailed compound data, scoring metrics, and experimental protocols are available upon request.

Contact Us for Additional Information

Extending to Neurodegeneration

The method validated on rhodopsin does not depend on anything specific to that protein. The same Hamiltonian DIFF vector approach works for any target with structural data on healthy and diseased conformations.

Amyloid-β

Alzheimer's Disease

Characterize the misfolding grammar of Aβ oligomerization and screen for food-derived compounds that stabilize the non-toxic monomeric form, preventing nucleation and subsequent aggregation.

τ

Tau Protein

Alzheimer's & Frontotemporal Dementia

Decode the aggregation signatures of tau hyperphosphorylation variants. Screen for chaperones that prevent formation of paired helical filaments.

αS

Alpha-synuclein

Parkinson's & Lewy Body Dementia

Map the conformational transitions from intrinsically disordered monomer to fibrillar aggregate and identify compounds capable of blocking or reversing this pathological cascade.

Get in Touch

Interested in Our
Research?

Request a briefing to learn more about our computational platform, validation results, and future directions in protein misfolding correction.