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.


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.
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.
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.
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order of what can be acted on, medicines and laboratory results, opened from a read-only link that expires and can be withdrawn.
Each wearable account and phone is listed with its last measurement. A source that has stopped says so, with the way back in.
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.
One page for the month just gone: nights measured, stretches that stood apart, sleep, and the results and medicines added. Open, print or share.
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.
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.
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.
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.
Drug–drug interaction detection from first principles, silent-variant risk flagging, target prioritisation, and expression-yield and formulation questions in biomanufacturing.
Variant interpretation independent of conservation and machine learning, cohort validation on public datasets, and reproducible pipelines published with the results.
The operator behind HexaGene also runs chemistry and materials problems on quantum processors: battery and electrolyte design, catalyst screening and photoresist chemistry.
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.
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.
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.
Enter biomarker values to see a simulated structural stress score. This is a demonstration — production scores use the full engine.
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.
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.
Type or paste a DNA sequence. The engine walks it in real time — assigning quantum states and codeword membership to each 6-mer window.
Toggle a single nucleotide change and watch the structural disruption propagate. The score changes because the physics changed — no retraining.
Every number below is reproducible from publicly available data. Zero parameters were tuned to outcome labels.

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

Health and Retirement Study — 14,565 adults. HexaGene detects system vulnerability years before disease onset: functional decline, depression, and incident heart disease.
Same physics engine applied to protein manufacturing. Predicts expression stability across 15 organisms. Failed constructs show 14% higher structural conflict at critical junctions.
The HexaGene operator template runs identically across superconducting, neutral-atom, and photonic quantum processors. 251+ production jobs completed.
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.
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.
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.
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.
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.
The quantum chemistry workflow built on this framework was runner-up in the Global Industry Challenge 2026, set by AIST and Mitsubishi Chemical.
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.


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

Unified biological age from routine labs. 6-axis decomposition for targeted intervention. AUC 0.897 on NHANES, years of lead time on HRS.
Physics-based variant interpretation — independent of conservation and ML. ACMG-orthogonal evidence layer. 198,494 ClinVar variants scored blind.
Readmission prediction, risk stratification, clinical pathway optimization. Gain-share model: you pay nothing until savings are demonstrated.
Drug–drug interaction detection from first principles. Three known interactions detected zero-shot on quantum hardware. Silent-variant risk flagging.
Expression-yield prediction (ρ = −0.92 across 15 organisms). Aggregation flagging. Multi-excipient epistasis quantification.
Battery cathode and electrolyte design. DFT-equivalent electronic structure at hours, not weeks. Validated against published quantum chemistry methods.
251+ production jobs. Cross-domain. Per-cell reliability at zero overhead.
ACMG evidence, pharmacogenomics, polygenic risk, tissue stress maps.
Enter biomarkers, see structural state computed in real time.
Reproducible pipelines for all published validation studies.
7,939 subjects, 10-marker panels, outcome validation. Citable DOI.
Quantum-research community entries and submissions.
Active deployment across pharma, diagnostics, manufacturing, and precision health. Engagement begins with a scoping conversation.
Drug interactions ab initio · silent-variant risk · target prioritisation.
Variant interpretation · VUS rescue · ACMG-orthogonal evidence.
Biomarker-based risk · structural-decay tracking · early-warning panels.
Gain-share on readmission reduction and clinical pathway optimization.
Expression-yield · aggregation flagging · formulation epistasis.
PPA-1 through PPA-10 + non-provisional filings for in-house pipelines.
Tell us about your application. We respond within two business days.
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.