Creating a digital map of COVID-19 virus for discovery of new treatment compounds

Quantum physics and high-performance computing have slashed research times for a consortium of researchers led by Qubit Pharmaceuticals. The discovery of chemical substances that may lead to new COVID-19 treatments was completed in only six months using cloud technology.

A global need for speed

The need for fast, accurate, and scalable medical research has never been more apparent. Since the start of the COVID-19 pandemic nearly two years ago, medical researchers have rushed to learn more about the virus, hoping to discover new pharmaceuticals to both prevent and treat the respiratory illness caused by the SARS-CoV-2 virus. While vaccinations against COVID-19 are jabbing arms at record rates, drug treatments remain limited for patients that do become infected, with only four treatments approved for use in the EU.

When will COVID-19 infections no longer be a serious threat to our health? A research company based in France have discovered two novel compounds – chemical substances that react in the presence of the COVID-19 virus – that may help reduce the severity and length of COVID-19 infections.

Using technology and services from AWS, Paris-based Qubit Pharmaceuticals—an international biopharma research company founded by researchers from the French National Centre for Scientific Research (CNRS), University of Texas, Sorbonne University, National Conservatory of Arts and Trades (CNAM), and Washington University—has made a significant step toward developing antiviral treatments for patients infected by COVID-19. Its recent discovery of SARS-CoV-2 main protease inhibitors is currently being assessed in the development of new COVID-19 drugs that stop the virus from replicating in patients, reducing the time and severity of infections.

Researchers discovered these compounds in only six months. Prior to using cloud computing resources, this undertaking would have required years of work and heavy upfront investments. Instead, quantum physics-driven molecular dynamics software accelerated by HPC in the cloud achieved what was previously impossible for smaller research teams who lacked the resources the major players in the biopharma research industry have access to.

An opportunity to go … really big

In May of 2020, the Laboratory of Theoretical Chemistry at Sorbonne University created an ultra-high resolution, digital map of the SARS-CoV-2 virus, relying on the power of the national supercomputing center (GENCI) and AWS services to perform HPC simulations, modeling different proteins involved in the SARS-Cov-2 and Cov-1 viruses. Building on the virus map, Qubit Pharmaceuticals, in collaboration with researchers from Italy, France, Switzerland, and chemistry company Enamine, began their research. The team tested libraries of data against the map to identify compounds that show high levels of activity when exposed to COVID-19. Additional simulations were performed in the AWS Cloud thanks to an AWS COVID-19 special grant. The team used Tinker-HP, a high-performance molecular dynamics software deployed on AWS.

The discovery of two novel drug candidates paves the way for the development of new treatments against the virus, bringing hope to millions of infected patients worldwide. Qubit Pharmaceuticals is now taking its research to the next stage by bringing the drug candidates into the synthesis and testing stage. They will continue their research to include the COVID-19 variants that have exacerbated the global health crisis.

Using compute hours provided by AWS and GENCI, Qubit ran its proprietary algorithms in the cloud to analyze large volumes of data while keeping an eye on cost-efficiency. Qubit used several AWS tools and instances to power their groundbreaking research methods. Amazon EC2 instances with NVIDIA GPUs allowed Qubit to go through vast data libraries looking for potential COVID-19 reactive compounds. Qubit was also able to take advantage of Spot Instances to keep research costs low. The cloud storage service Amazon EFS provided the researchers scalable and encrypted file storage. Finally, Qubit used a plugin for Slurm, a widely-used HPC job scheduler, to make efficient use of resources within AWS. Using these tools, the team dove deep into data libraries, testing compounds that might have previously been overlooked in the past, given time and cost restraints. You can read the full details of the research in their published peer-reviewed article here.

Conclusion

Being able to dramatically increase the speed, accuracy, and scale of virtual screening, quantum mechanics, molecular dynamics, and 3D structure solutions has proven to be a crucial tool in medical research. Deploying HPC technology in the cloud means researchers can scale their compute and storage as needed, providing resources for modeling and simulation to support medical research, inform clinical trial design, and predict trial outcomes without the heavy upfront investment for on-premises HPC.

Read more about High Performance Computing for Healthcare & Life Sciences and learn how AWS can accelerate the pace of medical research.