Dangerous MindsDangerousMinds
Featured Case Study

OnTarget: an agentic AI co-pilot for drug discovery

Built in partnership with In Your Genes Consulting, OnTarget integrates dozens of biomedical data sources into a single LLM-orchestrated workflow — turning weeks of cross-source evidence assembly into minutes.

Agentic AIBiomedical NLPMulti-source RAGBuilt to hand over

The Client

In Your Genes Consulting

A scientific consultancy led by a PhD geneticist with 15+ years of publishing experience, specializing in early drug discovery, GWAS meta-analysis, and Mendelian randomization for biomedical clients.

inyourgenesconsulting.net

The Challenge

Drug discovery is data-rich — and bottlenecked by manual synthesis.

Scientists were spending hours querying PubMed, Open Targets, ClinVar, GTEx, ClinicalTrials and a dozen other sources for every hypothesis — then manually stitching the evidence into briefs and slides. Each new question restarted the cycle.

In Your Genes Consulting needed a system that could let scientists ask high-level questions and have an agent execute the underlying multi-source workflow, with full traceability back to every citation.

The Solution

An LLM-powered agent that talks to the entire biomedical stack.

We designed and built OnTarget around four core principles. Scientists describe the question; the agent plans, retrieves, synthesizes, and cites.

Tool-integrated agents

Secure connectors to dozens of biomedical APIs and internal data sources.

LLM planning & orchestration

Goals are decomposed into multi-step plans the agent executes and validates.

Evidence-grounded outputs

Every claim is linked back to its source — no hallucinated citations.

Governed execution

Human-in-the-loop approval gates for sensitive or irreversible actions.

Integration Coverage

One agent. Dozens of authoritative biomedical sources.

OnTarget combines public and proprietary data into repeatable workflows — not one-source-at-a-time lookups.

Literature & Evidence

PubMedPubMed Central

Targets & Disease

Open TargetsHGNCOMIM

Genetics & Variation

ClinVardbSNPgnomADCIViC

Pathways & Function

KEGGGOHPOMeSH

Expression & Structure

GTExHuman Protein AtlasAlphaFoldAlphaGenome

Clinical Landscape

ClinicalTrials.govEU Clinical TrialsIMPCMGI

Agentic Workflows

Repeatable workflows, transparent reasoning.

Four core workflows the OnTarget agent executes end-to-end.

01

Target Triage & Prioritization

Start from a gene or disease hypothesis. The agent pulls target–disease associations, expression context, human genetics support, and pathway links — then scores target suitability with transparent source traces.

02

Variant & Mechanism Deep Dive

Gather known variants and clinical significance, cross-reference structural and functional evidence, and generate a mechanism-focused summary with linked references.

03

Competitive & Clinical Landscape

Retrieve relevant trials and therapeutic programs. Summarize modality, phase, inclusion patterns, and endpoints into a concise landscape report for portfolio decisions.

04

Research Brief Generation

Compile literature, target biology, genetics, and phenotype evidence into a decision-ready report or slide narrative — with approval checkpoints before finalization.

The Outcome

From manual stitching to a governed discovery co-pilot.

OnTarget shipped as a working agentic platform that In Your Genes Consulting now operates and extends. True to our model, the codebase, runbooks, and operating knowledge live with the client.

  • Faster hypothesis-to-insight cycles for the consulting team
  • Consistent, source-traced evidence synthesis across programs
  • Reproducible, auditable decision support for client engagements
  • A scalable operating model from exploratory research to portfolio review

Have a workflow worth automating?

Let's talk about what an agentic system could look like for your team — and how to make sure you own it from day one.