AI-first Research Institute · Budapest · Est. 2025

Not faster research.
Deeper questions.

We investigate complex systems across physics, biology, infrastructure and society — with humans and AI working as accountable thinking partners.

ISI Hypothesis · Trilith Method™ · FractAlgebra · EquoraVault
About Us

We study the connections disciplines miss

EQUORA Institute is an independent research institute based in Budapest. We pursue questions that sit between established fields, where conventional research structures often struggle to follow.

We treat the full path of inquiry as knowledge: hypotheses, source trails, methodological decisions, negative results, uncertainty, and the questions we have left open.

Our work is AI-first — because AI expands the space of hypotheses we can explore, while people remain responsible for the evidence, the interpretation, and every conclusion we publish.

We ask difficult questions, show how we reached our answers, and publish enough of the path for others to challenge them. Each thread — the ISI Hypothesis, the Trilith Method™, FractAlgebra, EquoraVault, the genomics work — has its research plan in the open.


Explore Our Research
Fractal patterns in nature
How we work

What AI-first
research means

AI participates throughout the research process: mapping literature, identifying patterns, connecting distant fields, generating competing hypotheses, and challenging early conclusions. Human researchers define the question, evaluate the sources, design the tests, examine the uncertainty, and remain accountable for the result.

See more connections

AI helps us examine more relationships across more sources than a single researcher could hold at once.

Cross disciplinary boundaries

We follow questions across physics, biology, mathematics, technology and social systems, treating the problem rather than the disciplinary border as the boundary.

Challenge the first answer

Multiple models and human reviewers test the same claim from different directions, so the reasoning that survives is stronger than the first answer.

Research Programmes

Three research programmes

Programme

Fundamental Structures

Research into the mathematical and informational structures that may underlie physical limits, self-reference, and complex systems.

ISI Hypothesis · FractAlgebra · Fractal Tokenomics →
Programme

Living Systems

Research into the interaction of biology, resources, environment, infrastructure, and local resilience.

Computational Genomics · Freshwater Systems · Unified Environmental Footprint · Living15 →
Programme

Trust & Verification

Research into how claims become evidence, how conflict shapes trust, and how communities retain the ability to measure, decide, and adapt.

Trilith Method · EquoraVault · NeverNormal · Trust, Conflict & Pride →
All research threads
Systems & Society

NeverNormal Research

The world is accelerating and "normal" is not returning. We research how individuals, teams, and organisations can build trust and adaptive capacity when change is the only constant.

See the research plan →
Regenerative Infrastructure

Living15 NanoLab

A living, scalable model of the 15-minute city in action. Researching how spatial design, resource systems, and community structure interact to produce sustainable local ecosystems.

See the research plan →
Computational Genomics

Computational Genomics

AI-augmented research into the intersection of genomic data, machine learning, and rare disease therapeutics. Calpainopathy/LGMD R1 serves as the primary disease model — from metabolic stabilisation to base editing strategies.

See the research plan →
Research Infrastructure

Trilith Method™ Development

The method claims that adversarial verification between different models catches what one model misses. Research testing that claim on the institute's own output, including the errors it failed to catch.

See the research plan →
Environment & Measurement

Unified Environmental Footprint

Carbon is one axis among six. Research into a multidimensional footprint model — carbon, water, humus, minerals, biodiversity, renewal capacity — in which the weighting is the reader's choice rather than the index author's.

See the research plan →
Water & Resources

Freshwater Systems

We know our carbon footprint. We rarely know our water footprint. Research into freshwater dynamics, consumption transparency, and planetary boundaries at the 2030 horizon.

See the research plan →
Trust & Social Systems

Trust, Conflict & Pride

Scientific consensus assumes conflict damages trust. We are testing the opposite: that conflict is a structural prerequisite of trust — and that pride is not an obstacle but a signal that something real is at stake. Live data collection open.

See the research plan →
Blockchain & Verification

EquoraVault / HFL

A self-learning oracle combining IoT sensing, machine learning, and fractal consensus to measure and verify real-world regenerative progress on a novel distributed ledger.

See the research plan →
Mathematics

FractAlgebra

Fractals are not shapes generated by algebra — they are the primitive elements of a new algebraic system. FractAlgebra proposes a formal axiomatization in which fractal combination, scale transformation, and self-reference are the foundational operations.

See the research plan →
Economics & Distributed Systems

Fractal Tokenomics

Token economic systems derived from FractAlgebra rather than designed by analogy. The UNA/NOVA/EVA mechanics of EquoraVault are the first live deployment — Fibonacci-ratio issuance, fractal governance, and scale-invariant value mechanics.

See the research plan →
Featured Research

The ISI Hypothesis

"What if every fundamental limit in physics is the same phenomenon, seen from different angles?"

The Informational Singularity Intersection hypothesis proposes a single axiomatic foundation — the number 1 as the unique Möbius-dual fixpoint — from which the measurable constants of physics emerge as necessary consequences.

Developed in collaboration with AI reasoning systems, ISI v3.5 is available as a peer-reviewable scientific preprint. The full text requires no technical background: nine chapters, from why the world exists to why π is not 1.

📄 Preprint · DOI 10.5281/zenodo.20095134
Read the accessible introduction →
Chapter overview
IWhy does something exist rather than nothing?
IIThe three singularity types: S1, S2, S3
IIIWhy the speed of light is unreachable
IVWhy absolute zero cannot be reached
VWhat it means that π is not 1
VIThe K5 mathematical proof
VIIEntropy and abstraction duality
VIIIThe Gödel-fractal construction
IXTime does not ontologically exist
Research Notes

Active observations & cross-domain connections

Working notes that emerged from active research — structural resonances, unexpected connections, and draft proposals that are not yet formal publications but are too significant to leave undocumented.

Our Mandate

The AI does not replace the thinker.
It extends what is thinkable.

For most of human history, knowledge was protected by distance — by language, by institution, by the assumption that complexity requires a credential to approach. That distance is closing. Not because the questions have become simpler. Because the tools have changed.

Today, for the first time, a person with a deep question and the willingness to pursue it can reach into the frontier of any field — and bring something back. This is not a convenience. It is a civilizational necessity.

The problems we face — water, fertility, disease, the structure of reality itself — are too large for any single discipline, any single institution, any single mind. At EQUORA Institute, we act accordingly.

Community

Where different disciplines meet

The Equora Emergence Circle brings together thinkers from different fields to share what they know and to name what they do not yet understand. The questions that follow are ones none of them would have reached alone.

Equora Emergence Circle
iterators.org
Open questions, upcoming Interference events, and a community that iterates long enough for something new to appear.
Get Involved

Bring us the question you cannot leave alone

Follow our work, join an open research conversation, or tell us about the question that keeps returning.

Write to us
Contact

Get in touch for inquiries.

Contact
Get in touch
Connect
Budapest · Barcsay utca
Computational Genomics · COMPGEN Get in Touch
EQUORA Institute › Research Domains › Computational Genomics
Active Research · Personal stake · AI-augmented

Computational Genomics

AI-augmented research at the intersection of genomic data and rare neuromuscular disease — with Calpainopathy (LGMD R1/CAPN3) as the primary focus, driven by direct personal involvement.

Domain codeCOMPGEN
Primary diseaseLGMD R1 / CAPN3
Strategy versionv3.1 (active)
Research since2024

Research disclaimer: The content on this page represents ongoing personal research and AI-augmented synthesis. It does not constitute medical advice, diagnosis, or treatment recommendation. All therapeutic strategies are experimental and in pre-clinical or monitoring stages. Consult a qualified specialist before making any health-related decisions.

Context

Research with personal stakes

This research domain exists because of a direct, personal diagnosis of Calpainopathy (LGMD R1). There is no detached academic interest here — the question of how to slow progression, preserve function, and prepare for emerging therapies is existential. AI is used as the primary synthesis partner because the patient population is too small to attract fast-moving commercial research, and because the intersection of genomics, molecular biology, and clinical trial data is too large for any individual to track without machine assistance. This is what "AI-first research" means in practice.

Calpainopathy is caused by pathogenic variants in the CAPN3 gene. Calpain-3 is a calcium-activated cysteine protease that functions not as a degrader but as a precision editor — it cuts and restructures proteins during sarcomere remodelling under mechanical stress. When absent, the structural feedback loop of muscle fibre repair breaks at a fundamental level.

The research is structured around four therapeutic pillars — from near-term metabolic support to long-horizon gene therapy monitoring — combined with a detailed personal supplement protocol (v2), biomechanical analysis of equestrian activity as a potential protective factor, and continuous AI-augmented literature monitoring across three models.

The CAPN3 file archive includes an optimised AAV-CAPN3 therapeutic vector sequence and a CureGene summary document — representing the most advanced end of the therapeutic horizon being tracked.

v3.1
Current therapeutic strategy — metabolic stabilisation, anti-fibrosis, base editing, and AAV gene therapy tracks
4 pillars
Parallel therapeutic tracks from current supplementation to gene therapy readiness
3 AI models
Claude (primary synthesis), ChatGPT Projects (monitoring), Gemini (cross-validation) — weekly sweeps
20+ years
Equestrian activity since age 4 — analysed as a potential biomechanical compensatory mechanism
Disease model

Understanding CAPN3

The key insight shaping all therapeutic strategy: CAPN3 is not a degrader — it is a precision editor. Its absence prevents dynamic remodelling, not just breakdown.

🔬
What CAPN3 does

Calcium-activated, non-lysosomal cysteine protease localised to the Z-disc. Performs limited proteolysis — precision cutting and restructuring under mechanical stress. Acts as sarcomere quality controller and mechanosensor signal generator.

What happens without it

Proteins get stuck in fixed conformations. Mechanical feedback loops break. The sarcomere cannot dynamically adapt. Result: progressive necrosis, fatty replacement, and fibrosis — particularly in proximal limb-girdle muscles.

💡
The puberty hypothesis (v3.1)

Puberty is the "critical window" where most new sarcomeres are built fastest — and without CAPN3 quality control, this is when the most defective structural elements accumulate. Functional decline at 30–40 may reflect build quality from ages 13–18.

🧬
Gene & references

CAPN3 chromosome 15q15.1 · autosomal recessive · ~550 known pathogenic variants.

OMIM #253600 · NCBI Gene: 825 · ClinicalTrials.gov

Therapeutic strategy v3.1

Four parallel tracks

Different time horizons, different mechanisms — run in parallel. Near-term metabolic support does not preclude tracking long-horizon therapies.

Pillar 01 · Current
Metabolic Stabilisation

Support mitochondrial function, reduce oxidative stress, and provide an energy buffer for the CAPN3-deficient sarcomere. The supplement stack is the active implementation.

Creatine (5g), CoQ10 ubiquinol (200mg), L-Carnitine (1g), NAC (600mg ×2), Omega-3, Curcumin, D3/K2, B-complex, C. Priority: creatine > omega-3 > D3/K2 > NAC > curcumin > L-carnitine > CoQ10.

● Active — supplement protocol v2
Pillar 02 · Current
Anti-Fibrosis & Quality Control

Slow fibrotic tissue replacement. Support autophagy as a partial compensator for missing CAPN3 function.

Intermittent fasting (16:8). Spermidine (1mg/day). Sauna 2–3×/week at 80–85°C for HSP induction. Curcumin for TGF-β pathway modulation.

● Active — partially implemented
Pillar 03 · Monitoring
Base & Prime Editing via iPSC

Track adenine base editing (ABE) and prime editing advances applicable to CAPN3 variants. Biological samples preserved as foundation for future iPSC derivation.

Addgene pEZ10 documented. FASTA maintained. CureGene AAV-CAPN3 vector archived. iPSC derivation planned 2026–2027.

◑ Monitoring — samples preserved
Pillar 04 · Horizon
AAV Gene Therapy

Track AAV-CAPN3 delivery programmes — muscle tropism (AAV8, AAVrh74, AAV9), dual-vector strategies for the ~2.4kb coding sequence, emerging clinical trials.

Optimised AAV-CAPN3 sequence archived locally. CureGene summary maintained. Weekly ClinicalTrials.gov + bioRxiv monitoring.

◑ Monitoring — trials emerging
Implementation

Supplement protocol v2

Timed around equestrian training. Fat-soluble compounds with meals; energy-support in the morning; recovery compounds post-exercise and before sleep.

TimingSupplementRationale
🌅 MorningCreatine 5g, L-Carnitine 1g, D3+K2, B-complex, C 500–1000mgEnergy buffer; fat-soluble with fat-containing breakfast
🍽 LunchCoQ10 ubiquinol 200mg, Omega-3 1000–1500mg EPA/DHA, Curcumin 500–1000mg + piperineAll fat-soluble; membrane protection + anti-inflammatory
🏇 Post-exerciseNAC 600mg, optional +Omega-3Reduce post-exercise oxidative stress; support regeneration
🌙 EveningNAC 600mg (2nd dose training days), optional 2nd Omega-3Overnight antioxidant protection; regeneration during sleep
CoQ10: ubiquinol only (not ubiquinone) — absorbs 3–8× better. Ubiquinone conversion may be impaired in CAPN3-deficient tissue. Do not combine NAC and high-dose Vitamin C within 2–3 hours.
Tracking metrics

Weekly: stair test, 30-sec squat count, 6-min walk, fatigue 1–10, soreness 1–10, HRV. Monthly: CK (creatine kinase) — primary objective biomarker. Target: downward trend or stabilisation.

Budget tiers (€/month)

Minimum (€35–50): Creatine + Omega-3 + D3/K2 + NAC

Full stack (€85–110): All above + CoQ10 ubiquinol + L-Carnitine + Curcumin + B + C

Metabolic test (planned)

4-week keto trial (<50g net carbs, 1.2–1.5g/kg protein) — test whether reducing glycaemic variability reduces CK and fatigue. Reversible if no improvement.

Biomechanical hypothesis

Equestrian activity as compensatory mechanism

Riding since age 4 while living with CAPN3-deficiency raises a significant hypothesis: decades of rhythmic, low-eccentric loading may have partially compensated for the missing calpain-3 function.

The biomechanical hypothesis

Riding at walk/trot generates ~5–10 Hz rhythmic oscillations through the rider's trunk — the frequency range that induces Ca²⁺ oscillations in muscle tissue. This may function as a natural NMES analogue. If correct, riding since age 4 may have "trained" a specific compensatory pattern that conventional gym training cannot replicate.

Why riding is protective

Riding is predominantly isometric — the rider stabilises rather than repeatedly shortening and lengthening muscles. Eccentric contractions generate the most micro-damage; CAPN3-deficient muscle cannot repair micro-damage efficiently. Walking and sitting trot minimise eccentric loading almost entirely.

The recommendation: don't stop, optimise

The working hypothesis is that riding is not just safe but actively therapeutic — a prescription, not a hobby. The protocol below reflects v3.1 risk/benefit analysis by gait and activity type.

Activity% of sessionRiskNotes
Walk40–50%LowFully safe, rhythmic Ca²⁺ oscillation, do daily
Sitting trot30–40%Low–MedGood technique essential; 3–4×/week
Canter5–10%MediumOnly when fresh; max 2×/week; monitor fatigue
Jumping0–5%HighEccentric overload on landing; avoid or minimise
Session length60–90 minOptimal2+ hours: rest day mandatory after
Supplement timing around riding

60 min before: Creatine at breakfast ✓

30 min after: NAC 600mg + electrolytes

Lunch: Omega-3 + Curcumin (anti-inflammatory recovery)

Advanced research horizon

Genomic data & therapy readiness

The CAPN3 file archive maintains molecular biology resources for long-horizon therapeutic tracks — in preparation for when gene therapy or base editing becomes clinically accessible.

Optimised AAV-CAPN3 Vector

Codon-optimised CAPN3 therapeutic vector sequence (FASTA) archived locally — from CureGene AAV programme analysis. Includes ITR sequences, muscle-specific promoter elements, and codon optimisation for human expression.

Addgene pEZ10 Reference

Documentation for pEZ10 (Addgene) — relevant plasmid for CAPN3 expression and base editing work. Maintained for future iPSC-based experiments or academic collaboration.

Addgene repository →

Biological Sample Preservation

Sample preservation protocol in place since 2024. Banked for potential future iPSC derivation, base editing studies, or biomarker research. The critical physical foundation for any patient-derived therapy.

Weekly Literature Monitoring

ChatGPT Projects: continuous monitoring of LGMD, Calpainopathy, Muscular Dystrophy. PubMed, bioRxiv, ClinicalTrials.gov. Cross-validated with Claude per AI Research Log Protocol v1.1.

Current LGMD R1 trials →

Methodology

How AI augments the research

A patient population too small for rapid commercial investment makes AI non-optional — it is the only way to maintain meaningful coverage across molecular biology, clinical trials, and genomics simultaneously.

🔬
Weekly Literature Sweep

ChatGPT Projects: continuous monitoring of LGMD, Calpainopathy, Muscular Dystrophy across PubMed, bioRxiv, ClinicalTrials.gov, OMIM — surfaced within days of publication.

🧠
Multi-model Validation

Key findings cross-checked across Claude (primary synthesis), ChatGPT Projects (monitoring), Gemini (validation). Logged with timestamps, model versions, and confidence assessments.

💊
Strategy Synthesis

Therapeutic strategy documents (v2.0 → v3.1+) produced through extended AI dialogue — hypothesis generation, mechanism analysis, supplement optimisation, biomechanical analysis of equestrian activity.

🔗
Cross-domain Integration

COMPGEN connects to EquoraVault biosensor work (IoT physiological monitoring), ISI theoretical framework (information processing in biological systems), and FRESH domain (environmental health correlates).

Connect

Researchers, clinicians, patient organisations

We are interested in connecting with anyone working in LGMD R1/CAPN3, neuromuscular disease, gene therapy, iPSC platforms, or AI-augmented rare disease research. Personal stake means we take this seriously.