Researchers at the University of California, Riverside leveraged machine learning to screen more than 10 million commercially available small molecules for interactions with SARS-CoV-2 viral proteins, and found hundreds of drugs that might offer treatments against COVID-19.
Researchers at the University of California, Riverside (UCR) leveraged machine learning to screen more than 10 million commercially available small molecules for interactions with SARS-CoV-2 viral proteins, and found hundreds of drugs that might offer treatments against COVID-19.
“There is an urgent need to identify effective drugs that treat or prevent COVID-19,” said Anandasankar Ray, a professor of molecular, cell, and systems biology who led the research. “We have developed a drug discovery pipeline that identified several candidates.”
Initially, Joel Kowalewski, a graduate student in Ray’s lab, identified previously-known ligands for 65 human proteins that are known to interact with the virus’ proteins. He then developed machine learning models for each of the human proteins.
“These models are trained to identify new small molecule inhibitors and activators—the ligands—simply from their 3-D structures,” Kowalewski said.
The researchers then developed a database of molecules whose structures were predicted to interact with the 65 protein targets, and also analyzed the compounds for safety.
“The 65 protein targets are quite diverse and are implicated in many additional diseases as well, including cancers,” Kowalewski said. “Apart from drug-repurposing efforts ongoing against these targets, we were also interested in identifying novel chemicals that are currently not well studied.”
Using the machine learning models, they screened more than 10 million molecules from a database of 200 million chemicals and then determined the best-in-class hits for the 65 human proteins that interact with the SARS-CoV-2 proteins. The next step was determining which of those compounds were already approved for human use by the U.S. Food and Drug Administration (FDA), and calculated toxicity.
Ray notes that the “compounds I am most excited to pursue are those predicted to be volatile, setting up the unusual possibility of inhaled therapeutics.”
The study, “Predicting Novel Drugs for SARS-CoV-2 using Machine Learning from a >10 Million Chemical Space” was published in the journal Heliyon.
A few of the identified drugs include Novartis’ Tykerb (lapatinib) used to treat breast cancer; Wyeth’s Sparine (promazine), an older drug now discontinued used to treat schizophrenia; and Karyopharm Therapeutics’ Selinexor used to treat multiple myeloma and B-cell lymphoma.
The researchers are looking for funds and collaborators to move testing cell lines, animal models and eventually clinical trials.
“Historically, disease treatments become increasingly more complex as we develop a better understanding of the disease and how individual genetic variability contributes to the progression and severity of symptoms,” Kowalewski said. “Machine learning approaches like ours can play a role in anticipating the evolving treatment landscape by providing researchers with additional possibilities for further study. While the approach crucially depends on experimental data, virtual screening may help researchers ask new questions or find new insight.”
Although the algorithms are exciting for shining a light on potential drugs and drug combinations that might be tested to fight the ongoing COVID-19 pandemic, they are aware that the potential use of their technology could have a major impact on broader drug screening and development.
“Our database can serve as a resource for rapidly identifying and testing novel, safe treatment strategies for COVID-19 and other diseases where the same 65 target proteins are relevant,” Kowalewski added. “While the COVID-19 pandemic was what motivated us, we expect our predictions from more than 10 million chemicals will accelerate drug discovery in the fight against not only COVID-19 but also a number of other diseases.”
