Zero trained parameters · Quantum-validated · Runner-up, Global Industry Challenge 2026

The single measurable state variable for human longevity.

One unified health score from routine bloodwork — decomposed into six actionable biological axes. No black boxes. No training data. Fully interpretable, clinically validated on 29,400 subjects.

A double helix of amber and pale-blue light. Illustration.
A glass double helix in morning light. Illustration.
0
Quantum hardware jobs
0
QPU platforms validated
0
Genetic variants scored
Trained parameters
The App · hexagene.app

Your genome, bloods, wearables and medicines, read together.

HexaGene is now a product people use. One health profile from four sources, with a report for the person and a report for their clinician, in English and Arabic.

76 seconds, with sound. A produced film with actors; the screens in it are illustrative. Decision support to be read with a qualified clinician, not a diagnosis.

Read against your own normal

Each night from a wearable is read against the person's own usual range, beside their bloods and genes. When nights stand apart, the app asks what happened and keeps the answer with those nights.

The genome in plain words

89 topics answered from a genome file: medicines, family, nutrition and risks, each with its evidence. The same genome is told as a story in fourteen chapters.

One report a clinician can act on

A Clinician Summary with findings in order of what can be acted on, medicines and laboratory results, opened from a read-only link that expires and can be withdrawn.

Every source, and whether it is still sending

Each wearable account and phone is listed with its last measurement. A source that has stopped says so, with the way back in.

An estimate is shown as an estimate

A marker worked out from other results is drawn apart from what was measured, and replaced by a laboratory result when one is on file.

A month in review

One page for the month just gone: nights measured, stretches that stood apart, sleep, and the results and medicines added. Open, print or share.

Open hexagene.app →Ask for an evaluation →
Use Cases

One engine. Who uses it, and for what.

The same engine serves a person reading their own health, a clinic, a pharmaceutical team and a research group. Each starts from the data it already has.

People

Start with a wearable, a blood panel or a genome file. Each is read on its own, and together once a second arrives. Nine areas of health, searchable, with the evidence two taps away.

Clinicians

A clinical view of the same profile: a summary to read as a referral, a care plan, a screening calendar to review together, and pharmacogenomic findings beside the medicines list.

Clinics, employers and health programmes

An evaluation on a demonstration profile under an evaluation agreement, then a programme set up with you. Risk stratification and pathway work on a gain-share basis for health systems.

Pharmaceutical R&D

Drug–drug interaction detection from first principles, silent-variant risk flagging, target prioritisation, and expression-yield and formulation questions in biomanufacturing.

Research groups

Variant interpretation independent of conservation and machine learning, cohort validation on public datasets, and reproducible pipelines published with the results.

Frontier: materials, energy and quantum

The operator behind HexaGene also runs chemistry and materials problems on quantum processors: battery and electrolyte design, catalyst screening and photoresist chemistry.

01 How It Works

Six axes of health. One unified score.

Existing risk tools are optimized for single diseases. AI models offer predictions but no explanation of why. HexaGene provides multi-system biological insight that clinicians and patients can act on immediately.

Many strands of light meet at one point and leave as six beams. Illustration.

From routine labs to structural biology

Standard bloodwork — albumin, hemoglobin, hs-CRP, HbA1c, eGFR, RDW, uric acid — is mathematically mapped into a single structural stress variable. That variable then decomposes into six independent homeostatic axes, each pointing to a specific clinical action.

No genetic sequencing required. No wearable integration required. The engine runs on data every hospital already has.

+
Structural
Tissue integrity and anabolic reserve.
Albumin · Hemoglobin

Captures frailty and malnutrition risk before clinical thresholds are crossed. Action: nutritional optimization, iron/B12 evaluation.

+
Inflammatory
Chronic immune activation.
hs-CRP · NLR

Captures "inflammaging" — chronic low-grade inflammation accelerating biological aging. Action: anti-inflammatory dietary patterns, weight management.

+
Metabolic
Energy regulation and lipid metabolism.
HbA1c · Triglycerides

Captures insulin resistance trajectory years before diabetes diagnosis. Action: pharmacologic therapy, lifestyle modification.

+
Redox
Cellular stress and oxidative burden.
RDW

Red cell distribution width predicts all-cause mortality independent of anemia. Action: correct nutritional deficiencies, reduce oxidative stressors.

+
Kinetic
Systemic circulation and clearance.
eGFR · Lipid transport

Captures cardio-renal stress — the coupling between cardiac output and kidney filtration. Action: blood pressure optimization, nephroprotective strategies.

+
System Balance
Global load distribution.
Uric Acid

Uric acid is a coupling marker — elevated levels signal cross-system stress that single-disease scores miss entirely. Action: dietary purine moderation.

The Core Principle

Biology as a physical system

A biomarker spike is not an isolated bad number. Biological anomalies redistribute structural load across the entire physiological system — like a suspension bridge with a broken cable. HexaGene measures the resulting stress landscape across the entire structure, not the broken cables.

A suspension bridge at dawn, cables fanning from a tower. Illustration.

Try it — quick risk preview

Enter biomarker values to see a simulated structural stress score. This is a demonstration — production scores use the full engine.

0.22
Low Risk
Demonstration only. Production uses the full S21 engine.
The Engine

See the genome as physics.

Every 6-nucleotide window maps to one of 64 quantum states, partitioned into 21 codewords. Same input, same output — always. Hover the radar, type a sequence, toggle a mutation.

Interactive axis profile

Hover each axis point to see the biomarker values driving the score. The shape of the radar tells the clinical story — which systems are under stress, how severe, and what to do about it.

This patient shows elevated inflammatory and metabolic stress with adequate kinetic clearance — a pattern traditional single-disease scores would miss entirely.

Structural Inflammatory Metabolic Redox Kinetic Balance
0.68
Unified Risk Score

Live sequence reader

Type or paste a DNA sequence. The engine walks it in real time — assigning quantum states and codeword membership to each 6-mer window.

Mutation impact — live

Toggle a single nucleotide change and watch the structural disruption propagate. The score changes because the physics changed — no retraining.

...ATGGCTAGAT...
Stiffness
Harmony
Ti/Tv
Codeword
Stacking
0.08
Pathogenicity score
Reference — baseline
02 Clinical Validation

Tested on real patients. Published datasets.

Every number below is reproducible from publicly available data. Zero parameters were tuned to outcome labels.

A field of small beads grading from green to coral. Illustration.
Population · Longevity
AUC 0.897

All-cause mortality prediction

NHANES national cohort — 29,400 adults. 4.62σ separation between survivors and non-survivors. Absolute risk estimates align perfectly with observed event rates across all deciles.

4.62σ
Separation
98.4%
Classification
N=29,400
Subjects
10
Marker panel
Diagnostic · Genomic
198,494

Genetic variant interpretation

Blind pathogenicity scoring on ClinVar variants with zero tuned parameters. Provides an independent, physics-based evidence layer for clinical variant classification (ACMG-orthogonal).

T = 16.39
Effect size
AUC 0.67
Zero-param
p < 10⁻⁶⁰
Significance
0
Tuned params
A stone path with a line of light running to the horizon. Illustration.
Longitudinal · Early Warning
AUC 0.79

Years of lead time

Health and Retirement Study — 14,565 adults. HexaGene detects system vulnerability years before disease onset: functional decline, depression, and incident heart disease.

0.79
Functional decline
0.77
Depression onset
0.71
Heart disease
14,565
Subjects
Manufacturing
ρ = −0.92

Biologic stability prediction

Same physics engine applied to protein manufacturing. Predicts expression stability across 15 organisms. Failed constructs show 14% higher structural conflict at critical junctions.

15
Organisms
p = 10⁻⁷
Significance
14%
Junction conflict
ρ = 0.67
GLP-1 valid.
03 Quantum Hardware

One architecture. Three quantum platforms. Same operator.

The HexaGene operator template runs identically across superconducting, neutral-atom, and photonic quantum processors. 251+ production jobs completed.

A superconducting quantum computer's gold-plated refrigerator stages. Illustration.

IBM ibm_fez · 156 qubits

Heron r2 superconducting gate processor. 238+ production jobs across four application domains — pharmacogenomics, catalyst design, cardiac stratification, and protein formulation — using the same circuit architecture.

ρ = 0.9954 · two-run reproducibility
4 domains · zero-shot, no retraining between domains
18/18 gauge · built-in error detection at zero qubit overhead

Gauge-symmetry error detection on NISQ hardware

S. H. Bachani · Merlin Digital · April 2026
Read the manuscript →
Quantum Framework

How the quantum framework builds and improves HexaGene.

HexaGene is one application of a single physics operator. The same operator is run on quantum processors, and what those runs show is what the health engine is built from.

01 · ONE OPERATOR

The engine and the hardware share one description

The operator that scores a variant or a blood panel is the one programmed onto the processor. Nothing is retrained between a genome, a drug pair and a material.

02 · THE HARDWARE IS THE TEST

Predictions are checked on machines, with nothing tuned to fit

251+ production jobs across superconducting, neutral-atom and photonic machines. A result that holds on three kinds of hardware is kept; one that fails there is withdrawn.

03 · MEDICINES

Drug interaction work feeds the medicines layer

Pharmacogenomic and drug–drug interaction runs on hardware are the research behind the product's medicines area, where the app reports from graded published evidence.

04 · JUDGED OUTSIDE

The methods have been judged by others

The quantum chemistry workflow built on this framework was runner-up in the Global Industry Challenge 2026, set by AIST and Mitsubishi Chemical.

Global Industry Challenge 2026

Runner-up, AIST & Mitsubishi Chemical challenge.

Merlin Digital, the company behind HexaGene, was named runner-up in the Advanced Materials challenge of the Global Industry Challenge 2026, presented by Connected DMV. The submission modelled a tin-oxo compound used in EUV photoresists, scaling a quantum chemistry workflow from 12 to 40 qubits.

Four people on a stage in front of a screen reading Global Industry Challenge 2026, Advanced Materials, Runner-up, Merlin Digital
The Advanced Materials award, Global Industry Challenge 2026.
A glass trophy engraved Connected DMV 2026 Global Industry Challenge: AIST and Mitsubishi Chemical, Runner-up, Merlin Digital
The trophy: Connected DMV, 2026 Global Industry Challenge, AIST & Mitsubishi Chemical.
04 Applications

One engine. Six commercial verticals.

The same physics runs across longevity, diagnostics, pharma, manufacturing, health systems, and materials. No retraining between domains.

Glass models of a protein, a DNA helix and a crystal lattice. Illustration.

Precision health & longevity

Unified biological age from routine labs. 6-axis decomposition for targeted intervention. AUC 0.897 on NHANES, years of lead time on HRS.

Genomic diagnostics

Physics-based variant interpretation — independent of conservation and ML. ACMG-orthogonal evidence layer. 198,494 ClinVar variants scored blind.

Health systems

Readmission prediction, risk stratification, clinical pathway optimization. Gain-share model: you pay nothing until savings are demonstrated.

Pharmaceutical R&D

Drug–drug interaction detection from first principles. Three known interactions detected zero-shot on quantum hardware. Silent-variant risk flagging.

Biomanufacturing

Expression-yield prediction (ρ = −0.92 across 15 organisms). Aggregation flagging. Multi-excipient epistasis quantification.

Materials & energy

Battery cathode and electrolyte design. DFT-equivalent electronic structure at hours, not weeks. Validated against published quantum chemistry methods.

05 Science & Demos

Open validation. Protected method.

Manuscript

Gauge-symmetry error detection on NISQ hardware

251+ production jobs. Cross-domain. Per-cell reliability at zero overhead.

academia.edu →
Clinical demo

20-section genome report

ACMG evidence, pharmacogenomics, polygenic risk, tissue stress maps.

View demo →
Interactive demo

Metabolic-risk engine

Enter biomarkers, see structural state computed in real time.

Try it →
GitHub

Validation scripts

Reproducible pipelines for all published validation studies.

github.com →
Zenodo

NHANES dataset

7,939 subjects, 10-marker panels, outcome validation. Citable DOI.

zenodo.org →
aqora

Research profile

Quantum-research community entries and submissions.

aqora.io →
06 Engage

Production-ready. Commercially deployable.

Active deployment across pharma, diagnostics, manufacturing, and precision health. Engagement begins with a scoping conversation.

Pharmaceutical R&D

Drug interactions ab initio · silent-variant risk · target prioritisation.

Genomic diagnostics

Variant interpretation · VUS rescue · ACMG-orthogonal evidence.

Precision health

Biomarker-based risk · structural-decay tracking · early-warning panels.

Health systems

Gain-share on readmission reduction and clinical pathway optimization.

Biomanufacturing

Expression-yield · aggregation flagging · formulation epistasis.

Patent licensing

PPA-1 through PPA-10 + non-provisional filings for in-house pipelines.

Get in touch.

Tell us about your application. We respond within two business days.

Intellectual Property

The internal structure of the cell Hamiltonian, bond-term construction, dose-encoding maps, gate decomposition, energy terms, and application-specific input mappings are proprietary subject matter covered by U.S. provisional patent applications PPA-1 through PPA-10 (incl. US 64/027,290) and non-provisional filings, owned by Merlin Digital.