How Our R&D Team Is Using AI To Transform Drug Discovery

Why AI and Computational Science Matter

For Yves Fomekong Nanfack, Head of AI/ML Research, AI is a catalyst for better science.

“The best outcomes happen when computational and experimental scientists are working together from day one. AI gives scientists the ability to explore more hypotheses and focus experiments where they matter most," shares Yves.

Yves’ team develops foundational and agentic AI models that mathematically represent biology and chemistry. These models help Takeda teams:

• Explore vast hypothesis spaces that cannot be tested manually
• Screen and design small molecules and biologics with greater precision
• Reduce experimental burden while improving molecule quality and safety

Here, AI success is measured by impact, not metrics.

“We do not optimize for model accuracy. We optimize for getting molecules into the clinic and to patients,” explains Yves.

Turning Human Data Into Decision

In neuroscience, AI supports earlier and more confident decisions across discovery and development. Daria Prilutsky, Director of Computational Biology and Neuroscience Lead, manages a global team that works end to end across the pipeline.

Her team integrates genetics, biomarkers, multi-omics data, functional data and scientific literature to inform:

• Target identification and prioritization
• Disease and pathway relevance
• Biomarker strategy and patient segmentation

According to Daria, what use to take months or more can now be done in weeks. As new data comes in, her team can iterate and compare biological hypotheses systematically. This reduction in uncertainty upstream means improved outcomes downstream.

Where Wet Labs and AI Converge

Automation plays a critical role in how data is generated earlier in development. Hilary Ranson, Associate Director of High-Throughput Mass Spectrometry, leads work that helps teams access higher quality data sooner. 

Her wet lab work enables scientists to:

• Screen more compounds earlier in development
• Generate richer in vitro data to inform later in vivo studies
• Identify false positives and liabilities sooner

She shares, “The earlier we get high-quality data, the better we can predict what will happen downstream.”

That data feeds directly into computational models, strengthening the connection between experiment and prediction.

Cellular Lead Profiling and Closed-Loop Discovery

Cellular lead profiling sits at a key decision point in the discovery process. In this space, Sangram Parelkar, Associate Director and Head of Cellular Lead Profiling, focuses on integrating automation, biology and analytics.

His teams run automated cellular assays at scale and pair them with real-time data analysis to inform go or no-go decisions.

“Labs of Tomorrow is about closed‑loop systems,” he explains. “Data from one cycle directly shapes the next. We are making smarter decisions with higher confidence.” 

By reducing Design‑Make‑Test‑Analyze cycle time, Takeda improves both productivity and molecule quality.