What was the impact of the AI solution?

Shortened target identification and validation cycles by 40% Increased identification of viable oncology targets that were previously considered undruggable Reduced computational analysis time from weeks to days Improved efficiency of computational drug discovery workflows

Source

MicrosoftLive sourceScaled productionEvidence: High75/100

AI-Enabled Collaboration for First-in-Class Oncology Targets

Use case typeOncology discoveryUpdated Jun 13, 2026

Ideaya Biosciences partnered with ATTMOS to engineer a physics-based computational discovery platform targeting undruggable oncology molecules, leveraging AI/ML and high-performance computing methods.

Organization: Ideaya Biosciences
Industry: Pharma
Location: Global
Published: March 2025

Reported outcomes

40%

quantified impactOther quantified impact

Strategic outcomes

New product / capabilityDeveloped a physics-based discovery platformNew product / capabilityEnabled target prioritization with machine learningBetter decisions & insightAccelerated R&D decision-makingNew product / capabilityIncreased viable oncology target identification

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Quantified impact: 40%

PharmabizMar 5, 2025News articleInferred claimHigh evidence strength

Shortened target identification and validation cycles by 40%

Last evidence check: Jul 22, 2026

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Ideaya Biosciences
Provider: Microsoft
Maturity: Scaled Production
Linked source: Pharmabiz

High failure rate for novel oncology drug targets due to 'undruggable' molecules Traditional R&D methods are slow and costly, stretching drug development timelines Limited ability to predict molecular interactions at scale Difficulty in analyzing large, complex chemical datasets efficiently Developed a physics-based computational platform powered by AI/ML and HPC Leveraged Amber Molecular Dynamics Suite for large-scale molecular simulations Integrated advanced machine learning models for priorit

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 5

1AI-driven molecular simulation for undruggable oncology targets
2Automated prioritization of drug target candidates using ML
3High-throughput virtual screening with physics-based AI models

High failure rate for novel oncology drug targets due to 'undruggable' molecules
Traditional R&D methods are slow and costly, stretching drug development timelines
Limited ability to predict molecular interactions at scale
Difficulty in analyzing large, complex chemical datasets efficiently

Developed a physics-based computational platform powered by AI/ML and HPC
Leveraged Amber Molecular Dynamics Suite for large-scale molecular simulations
Integrated advanced machine learning models for prioritization of druggable targets
Automated analysis and prediction processes to accelerate R&D decision-making

Technologies

Amber Molecular Dynamics Suite

Shortened target identification and validation cycles by 40%
Increased identification of viable oncology targets that were previously considered undruggable
Reduced computational analysis time from weeks to days
Improved efficiency of computational drug discovery workflows

Implementation partners1

ATTMOS

Sources & evidence1

Evidence: High75/100Evidence strength