UK Series (Part 3): From Genomes to Targets: Biomarkers, Stratification, and Repurposing at Scale

Executive Summary｜UK-scale WGS and multi-ancestry GWAS enable earlier, stronger target validation, biomarker design, indication shaping, and safety anticipation. This article turns discoveries into practice—workflows, metrics, QC, and pitfalls—then closes with brief case snippets as a bridge to implementation (trials, access, manufacturing) in Part 4.

1) End-to-End Frame: Data → Hypothesis → Validation → Execution

• Data: WGS (non-coding/rare/SV) + multi-ancestry GWAS + EHR/prescriptions/labs/imaging.
• Hypothesis: Link Variant→Gene→Pathway→Phenotype; test causality (colocalization/MR) and portability across ancestries.
• Validation: Human knockouts/LoF carriers, functional annotation, fine-mapping, external replication.
• Execution: Bake stratifiers (genetic + clinical) into trial design; work backwards from access and manufacturing.

2) Target Validity: Aggregating the Right Evidence

2.1 From Association to Causation

• Step-up: GWAS locus → colocalization → fine-mapping → functional annotation → MR to raise causal confidence.
• Diversity: Exploit LD differences across ancestries to narrow candidates and detect ancestry-enriched effects.

2.2 Human Genetics as Safety “Preview”

• LoF/knockouts: Natural inactivation informs efficacy plausibility and safety bounds before pharmacology.
• Dose analogs: Genetic partial inhibition provides anchors for effect-size expectations.

3) Biomarkers & Stratification: Pairing PRS with Functional Variants

• Single-variant × PRS: Combine functional variants with PRS to control background risk and reduce misclassification.
• Multi-ancestry first: Train and validate PRS across ancestries to ensure fairness and transportability.
• Composite markers: Integrate genetics + clinical + imaging + -omics; make rules operational, not abstract.

4) Indication and Trial Design: Start with Stratification

1. Endpoint coherence: Align mechanism and clinical endpoints; minimize pathway distance.
2. Eligibility rules: Variant carriers/PRS thresholds/clinical markers as composite gates.
3. Sample size: Enrichment boosts event rates and statistical efficiency.
4. External validity: Plan bridging and generalization across ancestries upfront.

5) Safety & Pharmacology: On- and Off-Target Anticipation

• On-target: Long-run effects of natural variation around the target guide chronic-treatment limits.
• Off-target: Pathway neighbors and interaction genes inform risk; PheWAS surveys the phenotype horizon.
• Drug–gene interactions: Prescriptions × genotypes expose unexpected adverse reactions or responders.

6) Indication Expansion & Repurposing

• Pleiotropy: Multi-phenotype signals reveal candidates for new indications.
• Subtyping: Genetic clustering decomposes “one label” diseases into actionable subtypes.

7) Case Snippets (Brief)

1. Rare-disease target: LoF phenotypes indicate tolerability → enriched PoC in high-burden carriers → early efficacy signal.
2. Inflammatory disease: Ancestry-enriched effect discovered → eligibility = PRS threshold + blood marker → clearer dose–response.
3. Repurposing: PheWAS reveals metabolic benefit → pilot with surrogate endpoints → Phase II bridge.

8) Practical Workflow (Keep Handy)

1. Translate the question into a causal hypothesis (phenotype, pathway, safety priors).
2. WGS/GWAS plan (fine-mapping, MR, colocalization, functional annotation).
3. Cross-ancestry validity (portability, LD leverage, PRS performance).
4. LoF/KO analyses for pre-emptive safety bounds.
5. Define stratifiers (genetic + clinical + -omics) and replicate externally.
6. Design trials (eligibility, N, endpoints, analysis plan).
7. Backward-plan access and manufacturing assumptions.

9) QC & Pitfalls

• Multiple testing: Align thresholds with priors; separate discovery and validation.
• Population structure: Control ancestry/batch effects to avoid spurious hits.
• Phenotype drift: EHR-derived labels change; version and audit rigorously.
• External replication: Pre-commit to replication across cohorts/ancestries.

10) Short Checklist

1. Is the mechanism–endpoint chain tight?
2. Can you show cross-ancestry generalization?
3. Do LoF/KO data bracket safety?
4. Are stratifiers operational and auditable?
5. Have you reverse-planned access/manufacturing?

11) Bridge to Part 4: Execution from the UK

Next we turn designs into action—UK-led trials with parallel/expedited reviews, NHS-ready outcomes, and manufacturing/partnership structures—so teams can move from plans to timelines.

---

Up next (Part 4): “Execution from the UK: Trials, Access, and Manufacturing Playbooks.” We will pair checklists with indicative timelines.

This article was edited by the Morningglorysciences team.