Polygenic score services

An end-to-end PRS portfolio spanning programme design, genotype data, quality control, imputation, scoring, interpretation, reporting, and integration.

Platform Contact

One focused service

Polygenic scores, delivered end to end

The public portfolio is focused on PRS and PGS. Each module supports a different part of the same controlled path from programme design to reporting and integration.

Intended use

Programme & pathway design

Define population, intended use, evidence boundaries, consent, reporting roles, and implementation requirements before scoring begins.

Applications: Prevention · stratification · cohorts
Methods: Use-case design · governance · validation plan

Sample or data

Genotyping & data intake

Work from a qualified genotyping route or controlled customer genotype data, with identity, format, coverage, and sample-quality checks.

Applications: New samples · existing cohorts · partner data
Methods: Genome-wide arrays · harmonisation · QC

Comparable input

Imputation & harmonisation

Versioned reference panels and controlled harmonisation create the variant coverage required by the selected score models.

Applications: Multi-array cohorts · longitudinal programmes · validation
Methods: Phasing · imputation · build alignment

PRS / PGS

Versioned polygenic scoring

Selected models run through transparent, versioned pipelines with allele alignment, model checks, score calculation, and traceable provenance.

Applications: Disease risk · quantitative traits · cohort stratification
Methods: Model registry · scoring · technical QC

Evidence in context

Interpretation & validation

Scores are interpreted against calibration, ancestry, reference populations, published evidence, and the clinical or research context.

Applications: Local validation · professional review · evidence updates
Methods: Calibration · benchmarking · evidence review

White-label delivery

Reporting & integration

Recipient, professional, technical, and structured outputs can be delivered under your brand and returned to the systems your organisation uses.

Applications: Clinics · laboratories · research platforms
Methods: Reports · APIs · structured data

Fit-for-purpose PRS technology

The right input and pipeline for the score

PRS delivery depends on reliable genotype data, appropriate reference panels, reproducible model execution, and validation that matches the intended population and use.

Scalable genotype input

Genotyping arrays

Genome-wide array data for polygenic scoring, with platform-specific sample and variant QC.

Reuse controlled data

Existing genotype data

Customer or cohort data can enter through a defined format, provenance, identity, and quality review.

Coverage and comparability

Reference panels

Versioned phasing and imputation resources support the variants required by selected score models.

Reproducible calculation

Scoring pipelines

Allele alignment, model versioning, score calculation, and QC remain traceable for every result.

Population fit

Validation datasets

Calibration, discrimination, ancestry context, and local performance determine responsible interpretation.

PRS / PGS

Available PRS indications

The current portfolio now covers a broader set of clinically relevant PRS indications across cardiovascular disease, metabolic risk, inflammatory disease, reproductive timing, bone health, and selected screening-oriented applications. The aim is still the same: to make score outputs interpretable, governable, and usable inside real healthcare and diagnostics workflows.

Availability does not mean every indication should be deployed in the same way everywhere. Some indications fit near-term laboratory or pathway design work more directly, while others require tighter local validation, governance review, or screening-policy alignment before use.

Cardiovascular

Vascular, rhythm, and blood-pressure indications for prevention and pathway design.

Coronary artery disease (CAD)

Cardiovascular risk stratification in prevention and follow-up pathways.

Atrial fibrillation

Inherited rhythm-related risk in cardiovascular settings.

Hypertension

Blood-pressure-related inherited risk alongside standard predictors.

Stroke

Vascular risk context interpreted alongside standard care variables.

Metabolic & lipid

Metabolic burden, lipid profiles, and inherited cardiovascular signals interpreted together.

Type 2 diabetes

Metabolic risk stratification interpreted with routine clinical context.

High body mass index (BMI)

Trait-linked metabolic burden with pathway relevance.

High triglycerides

Lipid-associated metabolic interpretation with practical relevance.

Low HDL cholesterol (hypo-HDL cholesterolemia)

HDL-related inherited burden in PRS context.

Polygenic hypercholesterolemia

LDL-oriented inherited burden and stratification.

Elevated lipoprotein(a)

Lipoprotein(a)-related inherited cardiovascular signal.

Immune & inflammatory

Immune-mediated and inflammatory susceptibility for diagnostic and cohort settings.

Inflammatory bowel disease (IBD)

Inflammatory risk framing across gastroenterology pathways.

Celiac disease

Immune-mediated susceptibility in diagnostic and cohort settings.

Reproductive & bone

Reproductive timing and bone-health susceptibility for prevention and counseling contexts.

Early menopause

Reproductive timing risk in preventive and counseling contexts.

Osteoporosis

Bone-health susceptibility for screening and prevention pathways.

Configurable reporting

From analysis to reporting in your care model

See how Starling translates polygenic scores into a visual recipient layer, professional interpretation, model-level QC, and structured output for your organisation.

Starling Genomics · Fictional demonstration data · configurable B2B output

Heart & circulation · at a glance

Inherited predisposition

2 elevated of 4

Hypertension Position 94 out of 100

High

Your position

Lower inherited predispositionHigher inherited predisposition

Of 100 people, about 94 have a lower or equal inherited predisposition and 6 have a higher predisposition.

Model comparison: 1.72x
Coverage: 100%

Coronary artery disease Position 93 out of 100

High

Your position

Lower inherited predispositionHigher inherited predisposition

Of 100 people, about 93 have a lower or equal inherited predisposition and 7 have a higher predisposition.

Model comparison: 2.12x
Coverage: 100%

Stroke Position 39 out of 100

Not elevated

Your position

Lower inherited predispositionHigher inherited predisposition

Of 100 people, about 39 have a lower or equal inherited predisposition and 61 have a higher predisposition.

Model comparison: 0.96x
Coverage: 100%

Atrial fibrillation Position 17 out of 100

Not elevated

Your position

Lower inherited predispositionHigher inherited predisposition

Of 100 people, about 17 have a lower or equal inherited predisposition and 83 have a higher predisposition.

Model comparison: 0.63x
Coverage: 100%

Detailed results

Heart & circulation

4 scores

Coronary artery disease (CAD)

High

93Position 93 out of 100

Your position

Lower inherited predispositionHigher inherited predisposition

Of 100 people, about 93 have a lower or equal inherited predisposition and 7 have a higher predisposition.

What this means

This fictional result places the individual near the 93rd percentile of the stated reference population. It indicates relatively higher inherited predisposition and should be interpreted alongside age, blood pressure, cholesterol, family history, and the intended care pathway.

PRS_CAD · vJ.01Coverage 100%Z-score +1.48
ADMIX reference contextModel comparison 2.12x
AUC 0.735 · QC Pass

Restricted research pathway for certain territories

Cancer-oriented scores and Alzheimer's disease are handled separately in certain territories. In those settings, use may need to remain inside a restricted research pathway because Medical Research Involving Human Subjects Act requirements and national screening-permit frameworks can apply.

That distinction is not a disclaimer for its own sake. It reflects the fact that population screening, inherited cancer susceptibility, and neurodegenerative risk communication can trigger a different governance threshold than routine pathway support in other PRS indications.