Harnessing Benchling AI/ML to Double Functional Titer and Reduce Experimental Runs

Affinia Therapeutics (“Affinia”) is an innovative gene therapy company with a pipeline of first-in-class and/or best-in-class adeno-associated virus (AAV) gene therapies initially for devastating cardiovascular diseases. The company’s programs are enabled by its proprietary AAV capsid engineering platform, which supports precise tissue targeting, and its cutting-edge plasmid system technology, which delivers industry-leading AAV production yields. 

When the team at Affinia Therapeutics set out to optimize their AAV production process, they had a clear goal: increase functional titer while reducing the number of experimental runs needed to reach that optimal outcome.

Partnering with Benchling, Affinia became an early adopter of Experiment Optimization in Benchling. Built on industry best practices in machine learning like Bayesian optimization, this feature helps scientists quickly and confidently identify the best set of conditions to hit their experimental goals, whether it’s maximizing titer, improving yield, or reducing cost.

“Machine learning capabilities in Benchling Insights helped me efficiently and cost effectively identify optimal experimental conditions during our process optimization efforts.”

 — Paul Freeman, Head of Upstream Process Science @ Affinia Therapeutics

By following Benchling’s machine learning recommendations, Affinia achieved the same functional titer using only half the DNA. This preserved quality and enabled potential cost savings.

Expected impact:

●  Significant cost savings per 200L and 1,000L GMP batch

●  Fewer experimental runs to find optimal conditions

●  Reduced analyst time through guided experimentation

Scaling smarter science

While this story is about AAV, Experiment Optimization is designed to apply across a wide range of scientific use cases:

●  Optimizing media formulations

●  Fine-tuning process parameters

●  Improving reaction conditions

●  Navigating complex purification workflows

●  Even guiding genetic construct design

With the release of Experiment Optimization in Fall 2026, any R&D team using Benchling Insights can use it to work smarter, faster, and more cost-effectively.