PharmaIQ held the Transforming Drug Discovery Through AI virtual event July 6-8th, which was focused on the potential of Artificial Intelligence (AI) to reduce the cost, time to market, and failure rate of the traditional drug discovery process. Strateos was pleased to participate and present a talk entitled Transforming Protein Engineering Through AI Enabled Smart Connected Labs on July 6th highlighting their collaboration with researchers from the University of Wisconsin-Madison.
Ben began the talk by outlining the key challenges in conventional analog R&D laboratory methodologies that contribute to poor traceability, reproducibility and long timelines. To overcome these hurdles and meet the industry need for rapid generation of large, clean data sets suited for machine learning, end-to-end infrastructure is needed with smarter software and automated platforms. Additionally, coupling AI with self-driving robotic platforms means experimental data can be used to refine experimental models iteratively and autonomously for new hypothesis generation or update experimental goals, which increase the speed and accuracy of scientific discovery.
Strateos’ Smartlab Software platform provides a hybrid solution providing researchers with the flexibility to control their own laboratory with the software and/or remotely tap into Strateos’ robotic infrastructure, SmartLab Studios, to expand testing capabilities.
Automating The Protein Engineering Discovery Loop
In the second half of the presentation, Phil Romero, Assistant Professor at UW-Madison, outlined the collaboration between Strateos and his protein engineering lab where the goal is to combine his AI-driven protein design platform based on directed evolution principles with Strateos’ robotic cloud laboratories to fully automate and accelerate the process of protein engineering.
“Protein engineering and synthetic biology fields are undergoing an evolution thanks to advances in high throughput automated experimentation in conjunction with machine learning / AI. Strateos’ capabilities are going to be invaluable to allow many labs to adopt these fully automated experimentation workflows.” -Philip Romero, Ph.D., Assistant Professor UW-Madison
Leveraging a Reinforcement Learning AI model, the researchers from Romero’s Lab remotely execute biochemical experiments through the Strateos SmartLab Studios, which are automated robotic cloud lab facilities. The data returned to the Romero Lab is used to refine the model to design new protein sequences to test with the goal of improving protein fitness, in this case, thermostability. The more iterative cycles that can be performed within a given timeframe, the better the final protein design, which is why automating and accelerating the experimentation is important.
Successes From The Collaboration
- Leveraging Strateos’ SmartLab Studios allowed researchers from the Romero lab to focus their time on the design principles and hypothesis generation rather than infrastructure management and redundant manual scientific experiments
- Coupling automation with an AI design platform increased the scale of testing protein sequences by ~32x
- Leveraging Strateos’ robotic automation reduced the cycle time from 2-3 weeks down to 7 hrs
In the Romero Lab’s experience, partnering with Strateos has been less expensive than service contractors and operating in-house robotic instrumentation, with cleaner data generation. Ben sees the work done in the Romero Lab as a good example of the next wave of life science experimentation where the coupling of AI and robotic infrastructure will be able to drive faster discovery cycles for many applications.
To view the presentation in full, and learn how Strateos’ Smartlab Studios can accelerate your research, please click the video below:
For more information or a demonstration of Strateos’ capabilities, contact email@example.com