AI-powered drug discovery specialist agrees 15 research collaborations

American startup Atomwise has partnered with no fewer than fifteen universities and health centers looking to use AI and predictive models to explore therapies for COVID-19.

Atomwise was established in 2012 to commercialize deep learning for structure-based, small molecule drug discovery.

In 2015, it launched AtomNet: a deep convolutional neural network for bioactivity prediction.

Each of the new research project aims at developing drug candidates with demonstrable broad-spectrum capability, with the potential for long-term benefits for future coronavirus outbreaks, the company said.

Strength in numbers

Several of the projects are part of the Atomwise Artificial Intelligence Screening (AIMS) program, designed to accelerate research into new treatments.

Research partners include the Dana-Farber Cancer Institute, the universities of Kentucky, Manitoba, South Australia, Connecticut, Toledo, and Texas Health Science Center at San Antonio. Columbia University and Jazan University in Saudi Arabia also are partners.

“Atomwise’s AI screening technology is used to predict the binding of more than 10 million small molecules to a protein of interest, and far exceeds what could be accomplished through traditional laboratory screening methods,” said Dr. Stacie Calad-Thomson, vice president and head of AIMS partnerships at Atomwise.

“Our hope is to democratize access to AI during the early stages of preclinical drug development and enable academics to contribute to the pandemic response.”

The collaborative research is focused on treatments for COVID-19, as well as other strains of coronaviruses that may emerge in the future.