Powering AI And Synthetic Biology In Therapy Design

By Erin Harris, Editor-In-Chief, Cell & Gene
Follow Me On Twitter @ErinHarris_1

During our recent Cell & Gene Live, “Building Smarter Cell Therapies with AI and Synthetic Biology,” Senti Biosciences’ Tim Lu, M.D., Ph.D., and Claire Aldridge, Ph.D., joined me to unpack data’s make-or-break role in fueling AI and synthetic biology for cell and gene therapies. As Dr. Aldridge put it, “data is the new oil,” and our panel cut straight to what CGT biotechs need: smart infrastructure, sharp priorities, academic strategies, and doable next steps.

Defining Essential Data Infrastructure for AI in Cell Therapy

Dr. Lu laid out the layers that make AI truly useful across discovery, process development, and clinical work. He stressed starting with a diverse set of data, including things that you consider successes and failures, to train models that don’t overfit and capture real biological variety. He stressed that human intuition is key.

Relevance matters too; high-throughput assays must mirror lab realities, not just surrogates for human function, often demanding a diverse set of assays to try to capture a more holistic picture. Beyond sequences and readouts, Dr. Lu pushed digitizing metadata. “Much of the existing metadata is still based on how the experiment was run, the specific protocols that were used, and the actual conditions,” he said. “Machine-readable formats unlock advanced algorithms that weigh context for deeper insights, not just correlations.”

Prioritizing AI Investments with Limited Resources

Dr. Aldridge addressed tight budgets, naming time as the ultimate scarcity. “Time is the one variable you can’t compress,” she warned. “Preclinical gaps such as spotty data can tank IND work, racking up delays that no cash infusion can fix.” Echoing Dr. Lu, she also championed early metadata. “Being able to put all of that data in at the very beginning is going to save you a lot of time when you’re pulling that information together for your IND package.”

For resource-pinched CGT firms, skimping on data systems is a false economy. “Your data management and how you’re going to do your data collection and annotation and that entire data system is not a place to scrimp because it will cost you more money and time later,” Dr. Aldridge advised. “Front-load it to hit milestones, such as first-in-patient dosing or scale-up runs without derailment.”

Tackling Preclinical AI Adoption in Academia

 “There are answers to some of our toughest health questions already sitting in data sets somewhere that we just can’t get to.” She urged leaders to invest in partnerships that monetize data, privacy safeguards and all.

Dr. Lu nodded to AlphaFold’s Nobel Prize as proof: “There’s academic gold here, but it’s sitting in data sets that no one is able to read across right now,” said Dr. Lu. “The first groups that do get their act together are going to win in terms of academic credit or industry.” According to Dr. Aldridge, “If you can be that first and you can capture the prestige of being the one to do that, that's something that is gold.”

Practical Tools, Governance Advice around AI and Synbio

Dr. Aldridge shared that proprietary data shines with targeted tools. “You don’t necessarily need to get access to those very powerful, large language models,” she said. “You just need tools that can help you gather insights that are more complicated than the human brain can keep together.”

“The non-trivial thing is actually making sure you have that data collected in a comprehensive way so you can actually apply the tools; tackle the legacy data pile-up now,” Dr. Lu said. Dr. Aldridge recommended matching data to solvable problems without bloating teams or budgets.

Data has delivered AI and synthetic biology success from Senti’s gene circuits to CGT at large. Dr. Lu and Dr. Aldridge, battle-tested in clinic-stage realities, stressed digitizing early and thoroughly, prioritizing time-saving systems, chasing prestige through data mining, and nailing discrete use cases. CGT leaders who act now will build smarter therapies faster without the strain.