AI discovers ways to recycle chemical waste – News

RESEARCHERS used artificial intelligence (AI) to discover ways to recycle waste into useful products.

The work takes on the challenge of performing a comprehensive analysis to understand what valuable products can be produced from a diverse range of wastes and could help achieve circular chemistry.

Bartosz Grzybowski led a team that used the drug discovery platform Allchemistry discover “tens of thousands of ways” to use commercially produced waste compounds to create about 300 known chemicals used in pharmaceuticals and agriculture. Grzybowski is a professor at the Institute of Organic Chemistry of the Polish Academy of Sciences, Poland.

Allchemistry combines state-of-the-art computational synthesis with AI algorithms to predict molecular properties. The research team fed around 200 waste substrates into the discovery tool and generated a large-scale array comprising “nearly a billion molecules”, from which they identified usable target molecules by focusing on agrochemicals and drugs.

The team used experiments to validate several of the suggested routes, including via an industrially realistic demonstration using the Pharmacy on demand flow chemistry platform. Pharmacy on demand is an advanced, miniaturized and automated suite of pharmaceutical manufacturing systems, owned by technology company On Demand Pharmaceuticals.

They also used an algorithm to evaluate synthetic routes with respect to durability and process parameters.

Although the study focused on agrochemicals and drugs, Grzybowski acknowledged that the network generated could include other “molecules of interest”. He also pointed out that Allchemistry accepts arbitrary substrates and the addition of other industrial waste or raw materials (or intermediates) could allow “to explore completely new molecular spaces”.

In an article about the work, the researchers said that widespread adoption of computerized waste recovery algorithms can accelerate the productive reuse of chemicals that would otherwise be stored, disposed of, or pose environmental risks.

Grzybowski said: “We are talking to many industry players and some government players to deploy Allchemistry to orchestrate circular chemistry and waste recovery on a larger scale.

“Ultimately, I see the software as a ‘chemical brain’ on a platform where some organizations enter the waste they would like to get rid of, the software makes predictions [on] what values ​​can be drawn from this waste, then other companies offer to take back this waste and carry out the syntheses.

He added: “I’m actually very interested in how this is going – everyone is talking about circular chemistry and now there is a tool with unparalleled capability to plan circular syntheses. Will it really have a impact on how we deal with chemical waste or will business-as-usual prevail?”

Nature: https://doi.org/hrxd