All case studies
AstraZeneca

Drug discovery · Case study

AstraZeneca: choose the right targets

Give every candidate gene a sourced dossier, a ranked call and an owning scientist’s decision, and reopen it when new trials or variants appear.

By OneShot · Updated

Read the solution ↓

What the workflow produces

Target evidence dossier
Target
Gene, protein, summary and pathways, with the date of each lookup.
Genetic and structural evidence
Variants with clinical significance, protein domains, function and structure coverage.
Chemistry and clinic
Compounds with properties and hazards, related trials and adverse event reports.
Ranking and decision
Proposed rank with its drivers, the owner’s decision and what reopened the file.

The documented problem

AstraZeneca says that after a major review of its research and development strategy in 2010 it created a 5R framework to guide how it discovers and develops drug candidates: the right target, right tissue, right safety, right patient and right commercial potential. The company describes the right target as uncovering, selecting and validating new targets with a strong link to disease, and says a stronger focus on biological rationale and understanding of the target markedly reduced the number of projects in its discovery portfolio.

AstraZeneca reports that its success rate from pre-clinical development to Phase III completion rose from 4% in 2005-2010 to 19% in 2012-2016. Its genomics page says drug targets guided by human genetic evidence are up to seven times more likely to achieve FDA approval, and that its human genetics research has supported more than 80 pipeline decisions since 2017. These are AstraZeneca’s own figures from its own pipeline. The OneShot workflow below assembles the evidence behind each candidate target and keeps it current as new trials and variants appear.

The OneShot solution

Candidate targets weighed on current, sourced evidence

Give every candidate target a current, sourced dossier and a recorded pursue, hold or drop decision by the scientist who owns the program.

Work enters the queue
A new candidate list, a new or changed clinical trial on the target, a reclassified variant or a new adverse event signal for a related drug.
Decision owner
Program scientist who owns the target
  1. 01

    Resolve the target

    The OneShot agent resolves each candidate gene symbol to one gene record and its protein, so aliases and retired symbols do not split the evidence. It pulls the gene summary and the pathways the gene sits in, and records the date of each lookup.

  2. 02

    Weigh the human genetic evidence

    List the gene’s known variants, resolve each to a standard identifier and read its clinical significance and the conditions it is reported in. Keep pathogenic, benign and uncertain findings apart, and record a missing classification as missing, not as benign. Research the published human genetics where the record is thin.

  3. 03

    Check the protein and the chemistry

    Read the protein’s function, its domains and whether a solved structure covers the region a drug would need to reach. Resolve the compounds known to act on the target, then pull their physical and chemical properties and their hazard classifications.

  4. 04

    Read the clinical record

    Find running and completed trials against the target or its pathway, and pull adverse event reports for approved drugs that act on it. Adverse event reports are recorded as reports, not as proof that the drug caused the event.

  5. 05

    Rank the targets

    The OneShot agent uses its prediction tool to rank the candidates against the team’s assessment criteria and its own past program outcomes, and shows which evidence moved each score. It compiles one dossier per target with every claim linked to its source; the ranking is a proposal and does not drop a target.

  6. 06

    Route, decide and recheck

    Send each dossier to the scientist who owns the program and chase an answer before the portfolio review. Record the pursue, hold or drop decision with its reasoning. When a new trial, a reclassified variant or a new safety signal changes the evidence, update the dossier and send the owner what changed, by text when a trial is stopped for safety.

What counts as complete

Every candidate target has a sourced dossier and a recorded pursue, hold or drop decision from its owning scientist, with any later change in the evidence sent to that owner.

Decisions and exceptions

  • The prediction ranks targets; the owning scientist decides which to pursue.
  • Variants of uncertain significance and conflicting classifications go to the owner as open questions, not as evidence for or against.
  • A trial stopped for safety or a new serious adverse event signal reopens the dossier at once, whatever its last decision.

What to measure

  • Candidate targets with a current dossier
  • Time from candidate list to owner decision
  • Dossiers reopened by new evidence
  • Pursued targets that reach the next stage gate

Establish the baseline and review period before launch. Compare completed cases, unresolved work and reviewer corrections against the same scope.

The agent’s tool basket

OneShot gives its agents the tools to analyze records, communicate, research and act on authorized business data. The platform carries the workflow from the first action through to the recorded result.

  • Target portfolio and program owners

    Read the candidate target list, assessment criteria, past program outcomes and each program's owning scientist; record the owner's pursue, hold or drop decision.

Workflow pricing

Scope the full workflow around your volume, business systems and required outcome.

Discuss workflow pricing →

Bring a workflow
like this one.

Discuss your pilotRead the docs