Suzanne Bowness, June 11, 2020
Contributing to curing COVID: Two Carleton labs join forces to add to international race to halt COVID-19 pandemic
When you’re searching for a needle in a haystack, ruling out 285,000 straws to narrow the pile to 279 offers a great head start. Even better news when the straws are protein combinations and the needle is the antidote to COVID-19.
This COVID contribution is an initiative of the combined research labs of Dr. James Green (above), professor in the department of Systems and Computer Engineering, and Dr. Kyle Biggar (below), professor in the Institute of Biochemistry. Both researchers credit their student, Biomedical Engineering PhD candidate Kevin Dick, with the initiative to start tackling the problem. Away in Florida when the pandemic became global, Dick knew by the time he returned in mid-March that he wanted to take action. “I felt a responsibility to contribute to this effort,” says Dick, who was part of a team in 2017 that made similar contributions to the scientific community to combat the Zika virus. “My experience from that project enabled me to rapidly take on this work to combat the spread of COVID-19.”
Dick applied software that Green’s group had co-developed over the last 10 years with Carleton University’s Bioinformatics Research Group to help predict physical interactions between human and viral proteins. Since virus proteins need to interact with human cell receptor proteins in order to gain access to the cell and then interact again once inside the cell to hijack it to produce many copies of the virus, figuring out which proteins are capable of interacting narrows the range of culprits. And with 20,000 human proteins and at least 14 viral proteins, there are a lot of possible combinations. The approach already has a precedent; the Bioinformatics Research Group has previously applied this approach in helping to understand how HIV and Zika viruses interact with human cells.
Dick and Green leveraged high-performance computing to run the 280,000 protein pairs through their machine learning software called PIPE4. They then ran the data through a software program called SPRINT, made available a group from Western University, to confirm the interactions and arrive at a set of interactions that both programs predicted with high confidence to be candidates for further study. The final tally was 279 possible interaction candidates involving 225 human proteins. The next step was to find out which parts of the proteins interact, to provide even more precision at the experimentation level so scientists could develop a drug that disrupts the interaction and prevents the virus from taking root.
After filtering the protein dataset, the team uploaded their results to Carleton’s Dataverse, a research data platform and repository for sharing, preserving and discovering data. Dick, Green, and Biggar then drafted a short research paper to post on open-access repository bioRꭓiv (pronounced bio-archive) as a preprint (an article draft shared before the scientific peer review process). The dataset of predicted interactions was downloaded over 1,000 times in the first week (it’s up to almost 3,000 at this point) by other research groups.
Biggar also began to wrestle with what to do with the data in his protein biochemistry lab, aiming to validate the interactions suggested by the dataset. The Biggar lab’s current work partners with a biotechnology company in Ottawa called NuvoBio Corporation to research peptide-based drugs for cancer treatment (the lab focuses specifically on lung cancer), so they were able to quickly switch focus to COVID-19 and investigate which peptides (protein fragments) might be able to disrupt the virus interactions.
The Biggar lab has the ability to make thousands of the peptides in just a couple of days and screen them for whether they disrupt protein interaction, although Biggar notes that the complexity of these interactions forms something akin to a “spider web” of connections that needs to be systematically navigated. His lab navigates these interactions for biological principles that are shared between the human proteins that are predicted to associated with the virus. These experiments have already revealed links to cellular processes related to inflammation and immunity. Then they start looking at high-priority human and viral proteins, creating them in the lab, analyzing how they interact, and testing how those interactions can be disrupted by the new peptides they have created.
The search continues
Green and Biggar are also excited that sharing their research through bioRꭓiv has afforded some good international connections. The Carleton researchers have been approached by scientists in Poland and Slovakia who are also now using their data. “Kyle and I were on a call for an hour with a researcher from Warsaw this morning, discussing how we could go forward together,” says Green, adding that the international pursuit of COVID-19 research has grown enormously. By the beginning of May there were over 2,900 preprint papers on bioRꭓiv and its health sciences counterpart medRꭓiv specifically about COVID-19.
The researchers also say that Carleton is a supportive place to do this research because of the collaborative nature of the university. “I think it is extremely unique to work at a university where it’s so easy to cross disciplinary boundaries,” says Biggar, who came to Carleton in 2016. “At other universities, I would never have talked with a computer engineer, but here the Carlton Bioinformatics Research Group is an example of that type of collaboration. And that’s where a lot of great and innovative science comes from, talking to people.”
Dick agrees that the ability to share research will help push results forward more quickly. “Our computational work promises to help guide other researchers in their effort to better understand the coronavirus proteins and to inform the design of therapies that might prevent the spread of the virus,” says Dick.
As for jumping on the COVID-19 research, all three researchers say they were glad to apply their skills to this urgent work. “My research group is one of the few that has the tools to make predictions that could actually help make an impact. I’ve never worked on something that felt so important and where I felt uniquely positioned to make a contribution. That was really neat,” says Green.
Share: Twitter, Facebook