Glossary

Mendelian Randomization

An epidemiological method using genetic variants as instrumental variables to assess the causal effect of a modifiable drug target on a disease outcome, mimicking a randomized controlled trial.

Get in touch Learn more

Control room desk with laptops and a large orchestration network display.

CAUSAL INFERENCE IN DRUG TARGET VALIDATION

What is Mendelian Randomization?

An epidemiological method leveraging genetic variants as instrumental variables to infer the causal effect of a modifiable drug target on a disease outcome, effectively mimicking a randomized controlled trial using observational data.

Mendelian Randomization (MR) is a statistical framework that uses germline genetic variants—typically single nucleotide polymorphisms (SNPs)—as instrumental variables to assess whether a modifiable exposure, such as a drug target biomarker, exerts a causal effect on a health outcome. By exploiting the random assortment of alleles at conception, MR mitigates confounding and reverse causation biases that plague conventional observational studies, providing genetically validated evidence for target-disease associations.

In computational drug repurposing, MR integrates with genome-wide association studies (GWAS) and protein quantitative trait loci (pQTL) data to prioritize therapeutic targets. A genetic variant reliably associated with altered levels of a circulating protein serves as a proxy for a pharmacological intervention. If that variant also associates with reduced disease risk, it provides causal evidence that modulating the target will yield clinical benefit, thereby de-risking pipeline decisions.

CAUSAL INFERENCE FRAMEWORK

Core Characteristics of Mendelian Randomization

Mendelian randomization leverages germline genetic variants as instrumental variables to strengthen causal inference in drug target validation, effectively mimicking a randomized controlled trial using observational data.

Genetic Instrumental Variables

Uses single nucleotide polymorphisms (SNPs) as proxies for drug target modulation. A valid instrument must satisfy three core assumptions: relevance (SNP robustly associates with the exposure), independence (SNP is not associated with confounders), and exclusion restriction (SNP affects the outcome only through the exposure). Common instruments include cis-eQTLs and pQTLs that directly influence gene expression or protein levels.

Mimicking Randomized Trials

Because alleles are randomly allocated at conception according to Mendel's laws, genetic variants are largely independent of the behavioral and environmental confounders that plague observational epidemiology. This natural randomization process breaks the confounding by indication that biases conventional drug-outcome association studies, providing unconfounded estimates of lifelong exposure effects.

Two-Sample MR Design

A powerful framework where genetic variant-exposure associations are estimated in one genome-wide association study (GWAS) cohort, and genetic variant-outcome associations are estimated in a separate, independent GWAS cohort. This design dramatically increases statistical power by leveraging large, publicly available summary statistics without requiring individual-level data from a single cohort.

Inverse-Variance Weighted Analysis

The primary meta-analysis method in MR that combines Wald ratio estimates from multiple independent genetic instruments. Each instrument's causal estimate is weighted by the inverse of its variance, giving more influence to precise instruments. The IVW method assumes all instruments are valid and provides the most statistically efficient estimate under the no horizontal pleiotropy assumption.

Pleiotropy Robust Methods

Sensitivity analyses that relax the exclusion restriction assumption to account for horizontal pleiotropy, where a genetic variant affects the outcome through pathways independent of the exposure. Key methods include:

MR-Egger regression: detects and corrects for directional pleiotropy
Weighted median estimator: consistent when at least 50% of instruments are valid
MR-PRESSO: identifies and removes outlier instruments

Colocalization Analysis

A Bayesian statistical approach that distinguishes causal variant sharing from mere genomic proximity. Colocalization tests whether the same causal variant drives both the exposure and outcome GWAS signals, ruling out linkage disequilibrium confounding where distinct variants in the same genomic region independently influence each trait. Essential for confirming target-disease causality.

MENDELIAN RANDOMIZATION EXPLAINED

Frequently Asked Questions

Clear, technically precise answers to the most common questions about using genetic variants as instrumental variables to infer causal relationships between modifiable risk factors and disease outcomes.

Mendelian randomization (MR) is an epidemiological method that uses genetic variants—typically single nucleotide polymorphisms (SNPs)—as instrumental variables (IVs) to assess the causal effect of a modifiable exposure (e.g., a drug target biomarker) on a health outcome. The method leverages the random assortment of alleles during gamete formation, which mimics the randomization process of a controlled trial. For an SNP to serve as a valid instrument, it must satisfy three core assumptions: (1) the relevance assumption—the variant is robustly associated with the exposure; (2) the independence assumption—the variant is not associated with confounders of the exposure-outcome relationship; and (3) the exclusion restriction—the variant affects the outcome only through the exposure, not via alternative pathways. In practice, MR is implemented using summary-level data from genome-wide association studies (GWAS) through methods like inverse-variance weighted (IVW) regression, MR-Egger, or weighted median estimators.

CAUSAL INFERENCE COMPARISON

Mendelian Randomization vs. Other Causal Inference Methods

A comparison of Mendelian randomization with alternative causal inference frameworks used in drug repurposing and epidemiological research.

Feature	Mendelian Randomization	Randomized Controlled Trial	Propensity Score Matching	Difference-in-Differences
Core Principle	Genetic variants as instrumental variables to mimic randomization	Physical randomization of treatment assignment	Matching treated and control units on estimated propensity scores	Comparing pre-post treatment changes between groups
Confounding Control	Strong; genetic variants are fixed at conception and not influenced by confounders	Strongest; randomization balances both measured and unmeasured confounders	Moderate; only controls for measured confounders included in the propensity model	Moderate; controls for time-invariant unmeasured confounders
Reverse Causation Protection
Data Requirements	Genome-wide association study summary statistics and outcome data from independent cohorts	Prospective enrollment with randomized treatment arms	Observational data with comprehensive covariate measurements	Longitudinal data with pre- and post-intervention measurements
Key Assumptions	Relevance, independence, and exclusion restriction (no horizontal pleiotropy)	Successful randomization, no attrition bias, no crossover	No unmeasured confounding, positivity, correct model specification	Parallel trends assumption, no simultaneous interventions
Typical Cost	$50K-200K per analysis using existing GWAS data	$5M-50M+ for a Phase III trial	$10K-50K using existing observational datasets	$50K-200K depending on data access and complexity
Time to Completion	2-6 months with available summary statistics	3-7 years from enrollment to readout	1-3 months with clean observational data	3-12 months depending on data availability
Drug Repurposing Applicability	High; can assess causal effect of drug target modulation on disease using eQTL or pQTL data	Gold standard but prohibitively expensive and slow for repurposing screening	Useful for generating hypotheses from electronic health records	Applicable when policy changes create natural experiments

Enabling Efficiency, Speed & Accuracy

Intelligent Analysis, Decision & Execution

We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.

Talk to Us

Search across company data

Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.

Useful when people spend too long searching or get different answers from different systems.

Enterprise searchRAGPermissions

Automate internal workflows

Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.

Useful when repetitive work moves across multiple tools and teams.

AI agentsWorkflow automationGovernance

Add AI to products and internal tools

Build assistants, guided actions, or decision support into the software your team or customers already use.

Useful when AI needs to be part of the product, not a separate tool.

AI integrationDecision supportModel routing

CAUSAL INFERENCE

Mendelian Randomization in Drug Repurposing: Key Examples

Mendelian randomization (MR) uses germline genetic variants as instrumental variables to estimate the causal effect of a drug target on disease outcomes, mimicking a randomized controlled trial to prioritize repurposing candidates.

PCSK9 Inhibition & Cardiovascular Disease

Genetic variants in PCSK9 that lower LDL cholesterol are associated with reduced coronary heart disease risk. This MR evidence validated PCSK9 as a causal target, leading to the development of monoclonal antibodies (evolocumab, alirocumab) that were later repurposed from hyperlipidemia to broader atherosclerotic cardiovascular disease prevention.

Instrument: PCSK9 loss-of-function variants
Exposure: Lower lifelong LDL cholesterol
Outcome: Reduced myocardial infarction and stroke
Mechanism: Upregulation of hepatic LDL receptor recycling

15-20%

Relative risk reduction in MACE

IL-6 Receptor Blockade & Rheumatoid Arthritis to COVID-19

MR studies using IL6R variants (e.g., rs2228145) demonstrated that genetically proxied IL-6 receptor inhibition reduced rheumatoid arthritis risk. During the COVID-19 pandemic, this causal evidence supported the rapid repurposing of tocilizumab from RA to severe COVID-19 pneumonia.

Instrument: IL6R missense variant rs2228145
Exposure: Reduced IL-6 signaling
Outcome: Lower RA activity; reduced COVID-19 mortality in RECOVERY trial
Validation: Concordance between genetic and pharmacologic effects

14%

Mortality reduction in severe COVID-19

HMG-CoA Reductase & Statin Repurposing

Variants in HMGCR that mimic statin-mediated HMG-CoA reductase inhibition show causal protection against coronary artery disease. MR analyses further revealed that genetically proxied statin use increases type 2 diabetes risk, a finding later confirmed in clinical trials, demonstrating MR's ability to predict both efficacy and adverse event profiles for repurposed drugs.

Instrument: HMGCR expression quantitative trait loci
Exposure: Lower LDL cholesterol via HMGCR inhibition
Outcome: Reduced CAD; increased T2D risk
Application: Preclinical safety screening for repurposing candidates

~10%

Increased T2D risk per mmol/L LDL reduction

SGLT2 Inhibition & Heart Failure

Genetic variants in SLC5A2 that reduce SGLT2 expression are associated with lower risk of heart failure hospitalization, independent of glycemic control. This MR finding supported the repurposing of SGLT2 inhibitors (empagliflozin, dapagliflozin) from type 2 diabetes to heart failure with reduced and preserved ejection fraction.

Instrument: SLC5A2 loss-of-function variants
Exposure: Reduced renal glucose reabsorption
Outcome: Reduced heart failure hospitalization
Key insight: Benefit independent of diabetes status

30%

Reduction in HF hospitalization

TYK2 Inhibition & Psoriasis to Autoimmune Spectrum

A TYK2 loss-of-function variant (rs34536443) protects against psoriasis and other autoimmune conditions. This human genetic evidence directly informed the development of deucravacitinib, an allosteric TYK2 inhibitor, and supports its repurposing from psoriasis to psoriatic arthritis, lupus, and inflammatory bowel disease.

Instrument: TYK2 P1104A missense variant
Exposure: Reduced TYK2 kinase activity
Outcome: Protection across multiple autoimmune diseases
Advantage: Selective targeting avoids JAK1/JAK3-related toxicities

Autoimmune indications under investigation

CETP Inhibition & Alzheimer's Disease

MR analyses using CETP variants that raise HDL cholesterol suggest a protective effect against Alzheimer's disease. This finding has prompted investigation into repurposing CETP inhibitors, originally developed for cardiovascular disease, toward dementia prevention.

Instrument: CETP loss-of-function variants
Exposure: Elevated HDL cholesterol levels
Outcome: Reduced Alzheimer's disease risk
Controversy: Discordance with negative cardiovascular trials highlights target-specific pleiotropy

20-30%

Estimated AD risk reduction per SD HDL increase

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.

Limited slotsGet a Free AI Consultation

How We Work

Custom AI workflows for your Business

One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.