The pandemic has raged worldwide and still fluctuates. Progress on emergency COVID-19 therapeutics has become a public focus.

Led by Professor 罗海彬, the team developed a method for precise prediction of drug–target affinity. By optimizing parallel-computing efficiency, they greatly accelerated the path from laboratory to clinic and successfully created the COVID-19 emergency drug dipyridamole, with encouraging clinical results.

Advanced computing raises drug-design speed by 30×

Raising the ability to predict drug–target affinity is a core problem in drug design. Traditional methods based on commercial software have a limited prediction range and, because of licensing and other constraints, are slow.

When speed is everything in a pandemic, Professor 罗海彬’s team, working on advanced computing platforms including Sugon, used independently developed free-energy-perturbation software for precise drug–target affinity prediction for the first time to create a COVID-19 emergency drug, shortening the prediction cycle for a single compound to 1–2 days. Compared with the 1–2 months per compound required by traditional FEP methods, speed rose 30–60×.

Clinical comparison: a 7:3 result after three weeks of treatment

Dipyridamole, created with the independently developed software, has moved from the laboratory into the clinic (154 COVID-19 patients). Versus an 11-day average discharge time on umifenovir (Arbidol), dipyridamole combined with umifenovir averaged 7 days to discharge—four days earlier.

Clinical comparison also showed that the emergency drug developed with the independently developed software improved coagulation in COVID-19 patients. Among severe patients sampled after three weeks of treatment, the discharge rate in the dipyridamole group reached 87.5%, versus only 33.3% in the contemporaneous control group.

Advanced computing: carrying the fight against the pandemic through to the end

This study, backed by advanced computing and independently developed software, offers a successful model for emergency drug R&D when the next outbreak comes.

Beyond emergency drugs, advanced computing has contributed in many other ways. When the outbreak hit at the start of last year, Sugon moved quickly into a scientific response. In early January 2020, the National Genomics Data Center released a novel-coronavirus resource library to help researchers analyze genomic variation; the center’s computing resources were provided by Sugon.

Sugon also launched an emergency computing-capacity program for national key research institutions, announcing more than 100 PFlops of free computing resources for the research community to support COVID-19 prevention and treatment. It has provided more than 2,000 nodes of free computing to 18 organizations including Huazhong University of Science and Technology, Huazhong Agricultural University, and the University of Science and Technology of China, strongly supporting gene and genome studies of SARS-CoV-2, viral mechanisms of action, structural evolution, drug mechanisms, drug screening, transmission, and early warning for severe COVID-19 patients.

Going forward, Sugon will put still more resources and effort into carrying this scientific fight against the pandemic through to the end.