HPE’s cloud-based supercomputer accelerates new drug discovery opportunities with biodiversity to fight COVID-19
JUNE 18, 2020 • BLOG POST • JOSEPH GEORGE, EXECUTIVE DIRECTOR OF HPC STRATEGIC ALLIANCES, HPE
IN THIS ARTICLE
- University of Alabama in Huntsville (UAH) increases outcomes for drug discovery using natural products predicted on HPE’s cloud-based supercomputer
The University of Alabama in Huntsville (UAH) uses HPE’s Sentinel cloud-based supercomputer to predict 125 naturally occurring products that counterattack with the virus’ protein, to be considered for drug testing against COVID-19
It’s been several weeks since we invited you along for the ride in chronicling one scientist’s drug discovery journey to fight COVID-19.
Since then, Dr. Jerome Baudry, a distinguished molecular biophysicist at the University of Alabama in Huntsville (UAH), has been regularly sharing updates in video blogs here on his team’s research journey that involves an HPE cloud-based supercomputer to assess the efficacy of naturally occurring compounds against the coronavirus’ proteins. The goal was to predict which natural compounds, if any, can be used in combination with other chemical compounds to develop a potential treatment and combat the pandemic.
The research is part of close collaboration between HPE and UAH where HPE donated time on its Sentinel system, which is a Cray supercomputer that is accessed through the cloud as part of an alliance with Microsoft Azure, and a dedicated group of its computational scientists. In exchange, we witnessed incredible scientific breakthroughs that occurred on a weekly basis from Dr. Baudry and the UAH team.
Speeding up drug discovery from months to weeks
Dr. Baudry and his team used HPE’s Cray Sentinel system to dramatically speed up discovery, and by applying a molecular docking approach, which is a preferred method in pharmaceutical research, the team studied 50,000 diverse natural products against a single protein of the virus in just weeks.
What was even more tremendous to find was how fast the supercomputer helped accelerate molecular docking calculations: from 20,000 molecular dockings per 24 hours (~86,400 seconds) to just 8 minutes (~470 seconds). This breakthrough helped the team perform dramatically more calculations with roughly 1.2 million molecular dockings per day.
The combined supercomputing and expertise has also helped save the scientists months or years of research and millions of dollars in costs.
Expanding possibilities for COVID-19 drug treatment with biodiversity
Today, I am honored to share a key milestone on progress that Dr. Baudry and the UAH team have made.
By using the massive computing power and modeling and simulation capabilities from HPE’s supercomputer, Dr. Baudry and team identified 125 natural products or close derivatives across plants, fungi, the sea, and microbes, to have chemical elements that bind several key coronavirus’ proteins1.
What does this discovery mean for the pharmaceutical research community? The results, which are sourced from diverse habitats and ecosystems, and have been traditionally known to have medicinal value, broaden the pool of drug candidates for pharmacologists to use as potential sources in drug testing to combat COVID-19.
Let’s further break this down.
The research found that the biodiversity of these natural products offers different strategies for testing and development of novel drugs.
Now, this research does not mean you should forest for these naturally occurring extracts in your backyard, the woods, or even the sea to consume as a treatment against COVID-19.
Dr. Baudry has made it very clear to us: “None of these natural products are expected to single-handedly be used as a treatment or cure to COVID-19. Instead, it is the combination of the chemical structures from this list that are beneficial to drug testing to develop a treatment.”
Molecular dockings using natural products performed on HPE's Sentinel system. Credit: The University of Alabama in Huntsville (UAH)
Extracting new learnings from nature
The research showcases that rather than taking a single chemical from a single plant, it is what these 125 natural chemicals have in common that can hold the key to designing novel drugs based on these parent chemicals of natural origin.
Some of the 125 natural products that are found in species and organisms across the world, include chemicals found in traditional medicine in South-East Asia, Eastern African coast, the Andes region in South America, as well as chemicals and organisms commonly found in North America.
The sources of these chemicals range from plants to fungi and microbes, exemplifying the great biodiversity of natural products of potential pharmacological interest. Examples include Cassiarine E, which is used in traditional medicines in Burma and Thailand and is found in an evergreen tree with yellow flowers, called Cassia siamea, or Beccamarin found in the bark of the tree Mesua beccariana, and has been used in traditional medicine to treat fever, renal diseases, poultice and dyspepsia in Malaysia.
Its doctor’s orders
Just like with any decision to start a new treatment, talk to your doctor first, because this one is telling us: “I repeat: these organisms or chemicals should not be used as treatments against COVID-19.”
The full list of natural products and research details can be found in the following paper: High Performance Computing Prediction of Potential Natural Product Inhibitors of SARS-CoV-2 Key Targets. Here are Dr. Baudry’s thoughts on the results:
“The combination of the powerful supercomputing technology and expertise from HPE, molecular docking methodology, and a large, diverse data-set of natural products, is what makes this research truly unique.
We predicted a range of natural products that can now move forward in clinical drug testing, some of which can be found in extracts that come from organisms that grow in our backyards, and some that can only be discovered in remote areas of the world such as in jungles or in the deep sea. We are inspired by these results and it is only the beginning of our journey to unlock more insight that will bring us even closer to conquering COVID-19.”
Dr. Jerome Baudry, Mrs. Pei-Ling Chan Chair and Professor in the Department of Biological Sciences Department of Biological Sciences, The University of Alabama in Huntsville (UAH)
Banding together to be a force for good
It has been an enlightening experience to be a part of Dr. Baudry’s journey. By extending the Sentinel system, which is a Cray XC50 supercomputer that provides extreme scalability and massive computing power in just a single cabinet, we aim to further advance and accelerate his vision for COVID-19 drug design using natural products.
To stay up to date on the progress of this research journey using HPE’s Sentinel supercomputer, tune into Dr. Baudry’s upcoming chat where he will share more about his experiences and the opportunities to advance drug discovery, at the HPE Discover Virtual Experience (DVE). Register for free here to attend the following session on-demand starting Tuesday, June 23:
- Session ID: B503 – From the Frontline: One Scientist’s Journey to Accelerate Drug Discovery
1Natural products listed as potential sources for drug design to combat COVID-19 are not proven to be a cure and should not be used as treatment for the virus