Science & Clinical Credibility

Built on validated evidence, not novelty.

Geneprint's value is not the quantity of data it generates — it is the clinical rigor behind every prediction and recommendation. Every layer of the programme is designed to be explainable, clinically governed, and honest about what it does not yet know.

Explainable by design Clinically governed Ancestry-aware
The Geneprint AI Engine

One model. Many inputs. A single governed output.

The engine brings genomic, clinical and biological information together to estimate cancer risk and personalise prevention — but it never operates, or reports, without a defined structure and human oversight.

AI inputs

Germline variants, polygenic risk scores, age, sex and ancestry, family history, personal medical history, previous biopsies and screening results, hormonal factors, lifestyle factors, clinical biomarkers, proteomics, metabolomics, and annual follow-up data.

AI functions

Risk estimation, risk-driver analysis, screening personalisation, prevention prioritisation, action tracking, and annual comparison against a person's own prior results.

Output standard

Every prediction Geneprint issues states the cancer type, the risk category, absolute and relative risk where validated, the reference population used, the contributing factors, a confidence level, defined limitations, and a connected action. Geneprint never issues an unexplained AI score.

Risk categories

Standard population risk Moderately increased risk Significantly increased risk High inherited risk Insufficient evidence for reliable classification
Polygenic Risk

Used selectively, not by default.

Polygenic risk scores (PRS) are included in a Geneprint report only when they meet a defined evidentiary bar — not because they are available.

Polygenic scores can perform differently across ancestry groups, because the underlying genomic datasets they were built from have historically underrepresented many populations. Geneprint uses ancestry-aware methods, reports performance limitations openly, avoids overstating precision, and combines PRS with clinical and family history rather than interpreting it alone.

Analytically validated
Clinical performance is understood
Evaluated in a relevant population
Adds information beyond family and clinical history
Can influence a real screening or prevention decision
Two Layers of Evidence

A deliberate separation between what changes management, and what strengthens understanding.

Geneprint keeps clinically established findings and advanced AI/multi-omic findings distinct — in the model, in the report, and in how each is allowed to be used.

Directly influences clinical management

Clinically Established Layer

Pathogenic hereditary variants, guideline-supported risk factors, established clinical risk models, standard laboratory abnormalities, established screening recommendations, clinically actionable family-history findings, and selected validated polygenic risk scores.

Strengthens stratification & evidence

Advanced AI & Multi-Omic Layer

Proteomic and metabolomic risk patterns, integrated multi-omic predictions, biological pathway patterns, and emerging combined risk scores. This layer is never presented as a cancer diagnosis, a cancer-free certificate, definitive evidence that cancer is developing, or a reason to perform invasive testing without clinical review.

Human Clinical Review

AI supports clinical judgement. It never replaces it.

Before any report is released to a patient, a defined sequence of human checks takes place.

Laboratory quality is verified
Significant genomic findings are reviewed
AI results undergo plausibility and consistency checks
Contradictory findings are flagged
Screening recommendations are checked against applicable guidance
Significant abnormalities are reviewed by the supervising physician
The final report is approved by an authorised Geneprint or hospital physician — specialists are involved only when required
Clinical Governance

The principles that keep Geneprint accountable.

AI does not independently diagnose cancer
A physician reviews significant clinical outputs
Actionable genomic findings are clinically confirmed when required
Variants of uncertain significance are never treated as pathogenic
Standard screening is never replaced
Emerging multi-omic findings are clearly labelled as such
Every abnormal result has a predefined pathway
Model versions are recorded
Report changes are auditable
Hospitals receive updated training when the programme changes
Consent & Data Protection

Informed at every stage. Protected by design.

Part of the consent process

The purpose of testing, the types of results returned, secondary findings, family implications, storage and sharing, biobanking, future reinterpretation, research participation, and withdrawal rights are all explained and agreed before testing begins.

Data architecture

Encryption in transit and at rest, role-based access, separation of identifiable and molecular data where possible, complete audit logs, hospital-specific permissions, defined retention policies, controlled research access, and jurisdiction-specific hosting where required.

Laboratory Model & Partners

Consistent standards, wherever the sample is processed.

Partner laboratories

Geneprint partners with qualified laboratories for whole-genome sequencing, routine clinical testing, proteomics, metabolomics, and optional cell-free DNA testing.

Geneprint controls

The sample kit, quality standards, assay performance requirements, data formats, QC thresholds, sample rejection criteria, turnaround time, reanalysis rights, and reported analytes — this consistency is what makes annual biological comparison reliable over time.

Evidence Generation Network

Evidence generation is core to Geneprint — not a side activity.

Each partner hospital participates in a consented, governed evidence programme that measures the programme's real-world performance.

Actionable hereditary findings
High-risk individuals identified
Screening and referral completion
Clinically significant abnormalities
False-positive investigations
New diagnoses and stage at diagnosis
Patient and physician satisfaction
AI model performance, including performance across ancestry groups
Limitations

What Geneprint is not.

Stated plainly, because trust depends on it.

Not a direct-to-consumer DNA test
Not a diagnostic test for existing cancer
Not a replacement for mammography, colonoscopy, cervical screening or other established screening
Not a guarantee that cancer will be prevented, or that it is absent
Not an unfiltered report of hundreds of unexplained genetic variants

Ready to understand your risk?