Using AI for a green thumb, Lumen teams with Google to help grow drug-producing algae

Artificial intelligence has promised to shave years off the pursuit of small-molecule drugs, and, now, Lumen Bioscience has enlisted researchers at Google to help do the same with large-molecule biologics.

Their goal is to wield machine-learning techniques to optimize the environment necessary for the growth of spirulina—the bluish-green algae Lumen has genetically engineered to produce therapeutic compounds.

With Google’s help and funding from the Bill & Melinda Gates Foundation, the company hopes to scale up its biomanufacturing capacity and the ability of algae to deliver cells filled with medicines that can be harvested, dried and sealed into a capsule to be taken orally.

In addition, the U.S. Department of Energy has thrown a $2 million grant behind Lumen’s efforts, which ultimately aim to package a variety of biologic drugs in an easy-to-take dose at a low cost, instead of the traditional intravenous infusion.

So far, the team’s machine-learning algorithms were able to crunch as many as 17 individual variables to find the best way to help algae grow—including water pH and temperature as well as the amount of light used and the proper spectrum. The number of possible combinations made it financially impossible to experiment with one change at a time.

RELATED: Lumen bags U.S. grant for GI-focused COVID-19 antibody cocktail

AI also offered an opportunity to leapfrog the improvement processes that took decades for previous biologic production methods, such as those centered around bacteria like yeast and E. coli as well as the mammalian Chinese hamster ovary, or CHO, cells, used to mass-produce proteins.

The team’s work, illustrating 16 rounds of experiments among 96 separate half-liter bioreactors, was published on the preprint server bioRxiv and showed adaptive machine-learning methods could help fine-tune production and potentially increase yield volumes by 70% to 100%.

In one example, researchers varied the intensity of light to find the sweet spot where it drove efficient photosynthesis but without causing cellular damage. In another, lower pH levels helped antibody-expressing spirulina strains grow at a larger scale.

“We believe this paper is the first to describe the application of AI techniques to biologics manufacturing,” said Lumen’s co-founder and chief scientific officer Jim Roberts.

“We look forward to the future implementation of these practices, as supported with funding from the Department of Energy, to provide mucosally and topically delivered biologics for highly prevalent diseases that, until now, have been infeasible due to the cost and scaling challenges of traditional biomanufacturing platforms,” Roberts said.

Meanwhile, the grant from the Department of Energy will help fund a project that builds on the AI researchers’ work, increasing its complexity and scale.

The goal will be to support Lumen’s clinical-stage pipeline of biologic drugs, including algae-produced treatments for Clostridium difficile infections, diarrhea and COVID-19 infections affecting the gastrointestinal tract.