CAMBRIDGE, Mass.--(BUSINESS WIRE)--For the National Center for Advancing Translational Sciences (NCATS), it’s all about delivering new treatments and cures for diseases faster.

“If you ask a patient when they want their cure, they’ll say yesterday,” said Sam Michael, NCATS chief information officer. “Generating a new therapeutic can take 10 to 15 years and billions of dollars. We need to find better drug development methods that are efficient.”

Looking toward a better tomorrow for patients, the NCATS scientist reached out to Kebotix, a technology platform company for new chemicals and materials, to enhance productivity of NCATS’ high-throughput experimentation (HTE) via artificial intelligence-driven optimization of assay conditions for biosynthesis inhibitors.

“We’re running systems in a very open loop fashion,” Michael explained. “We have our assay, we have our fixed chemical libraries, we run our experiment, we get our results, and then we iterate, but we’re really not taking advantage of some of the technologies out there, especially when it comes to machine learning.”

By collaborating with NCATS, part of the National Institutes for Health (NIH), Kebotix demonstrates fully autonomous optimization of the assay’s performance for biosynthesis inhibition. Kebotix software has processed and learned from previous experimental results, leading to submission of new AI-generated assay conditions via remote access to the HTE facility at NCATS for experimental testing.

“The goal of the multi-objective optimization in this case study was to reduce time and cost and R&D, reduce lab consumables – enzyme and substrate – while maintaining a sufficiently high signal-to-noise ratio defined in terms of statistical effect size,” said Dr. Jill S. Becker, Kebotix CEO. “This fits perfectly into our mission to usher in a new age of high-speed innovation using AI and robotics to increase the rate of discovery at a reduced cost.”

Prior to collaborating with Kebotix, NCATS had been using a high-throughput approach and factorial design of experiment (DOE) that required the team to run a full set of 294 experiments in a brute force fashion. Despite the moderate complexity of the optimization task (three free parameters), executing the factorial DOE kept an expensive HTE facility occupied for more than two days. This approach often wastes testing materials and time, and for an organization focused on speeding up the development of treatments for rare diseases, of which there are over 7,000, there needed to be a better way, according to Michael.

Saving Lives, Time and Money

Instead of testing a set of predefined parameters as in the traditional DOE approach, Kebotix uses adaptive optimization to gain more information using fewer experiments and less materials. By employing an advanced proprietary AI-driven optimization algorithm, Kebotix found suitable conditions and optimal assay performance with only 55 measurements with up to 20 running simultaneously. This output was sufficient for the Kebotix algorithm to identify and test the global optimum of the full 294 DOE experiments with a high probability (>95%). NCATS achieved a five-fold reduction in costs for lab supplies and HTE run-time, reduced from 49 hours to just 9 hours.

Kebotix provides NCATS with the ability to organize its operation and automate at a level where it has all the data in a very structured format ideally suited for machine learning and ready for running experiments autonomously.

“Our collaboration so far has resulted in a 5x reduction in costs of high-throughput

experimentation and a 5x increase in speed through fully autonomous assay optimization.” said Christoph Kreisbeck, Kebotix chief product officer. “The fact that these achievements effectively create efficiencies in the drug development process is a good thing for patients and saving lives.”

About Kebotix

Kebotix (www.kebotix.com) is a platform company for chemicals and materials ushering in a new age of high-speed innovation using artificial intelligence and robotic automation. Kebotix has built the world’s first self-driving lab for materials discovery powered by AI and robotics. Kebotix, founded in 2017, is accelerating the exploration, discovery, use and production of new molecules and materials that can solve some of the world’s most urgent problems.

Source:https://www.businesswire.com/news/home/20200212005239/e