It’s now well celebrated that genetic evidence for drug targets doubles the odds of clinical success with clear causal genes (Mendelian traits and GWAS associations linked to coding variants) being even higher.

Below are other target validation methods.

1) cis‑QTL → MR → colocalization (“genetically instrumented target modulation”)

What it is. Use variants local to the target gene (cis‑eQTL/cis‑pQTL) as instruments to estimate the causal effect of up/down‑modulating the target on disease; require colocalization with the disease GWAS to rule out LD confounding.

Why it de‑risks. Targets with MR and colocalization support are enriched among successful/approved programs; direction of effect tells you inhibitor vs. agonist.

How to use. Annotate each target–indication with (a) cis‑MR effect ± SE and (b) coloc PP4; score as high if PP4≥0.7 and MR FDR passes a prespecified threshold, then compute RS as in Minikel.

https://pmc.ncbi.nlm.nih.gov/articles/PMC7610464/

2) Human “natural experiments” of loss‑ or gain‑of‑function (LoF/GoF)

What it is. Individuals with naturally occurring pLoF/GoF alleles in the drug gene provide in vivo perturbations of the mechanism; LoF tolerance (LOEUF) quantifies population‑level constraint.

Why it de‑risks. Human LoF/GoF has repeatedly anticipated both efficacy and on‑target safety (e.g., PCSK9 and HMGCR for LDL‑C, diabetes risk with statin mechanism). Constraint (LOEUF) helps flag potential on‑target toxicity of inhibition. Nature+2Nature+2

How to use. Add binary/graded features: (i) presence of benign heterozygous LoF carriers, (ii) phenotype of LoF/GoF carriers aligns with desired effect, (iii) LOEUF decile. Use these as efficacy/safety priors in RS.

https://www.nature.com/articles/s41586-020-2308-7

https://www.nature.com/articles/s41586-020-2267-z

3) Phenome‑wide genetics for safety (PheWAS of the target gene)

What it is. Map the gene (or cis‑instruments) across thousands of EHR phenotypes to anticipate on‑target side‑effects and organ‑system liabilities.

Why it de‑risks. There’s a documented correlation between organ systems flagged by target genetics and the organ systems where drug side‑effects occur; tissue‑specific eQTL/PheWAS further improves side‑effect prediction. Nature+1

How to use. Build a side‑effect risk score per organ system from gene‑level PheWAS; include as a negative prior (penalty) in target triage.