Analysing peptides for antiviral and anti-inflammatory properties

SARS-CoV-2 Main Protease

Irish company Nuritas has carved itself a niche in the world of drug discovery with its AI platform that analyses the therapeutic potential of thousands of naturally-occurring peptide sequences. With the onset of the COVID-19 pandemic, they are now putting all of their efforts into discovering peptides that can be used to mitigate the disease’s progression.

Since its inception in 2014, Dublin-based company Nuritas has been changing the landscape of drug discovery. Using its proprietary artificial intelligence platform, it analyses vast quantities of peptides quickly and accurately and identifies top candidates for addressing some of the most difficult drug discovery challenges.

By leveraging the inherently high target specificity and low risk of immunogenicity that peptides possess, Nuritas creates therapeutics that meet or surpass industry standards across a broad range of diseases. Founded in 2014 by mathematician and bioinformatician Nora Khaldi, Nuritas is a leader in the identification of therapeutic peptides to prevent or treat disease.

In April, the company received an allocation from PRACE to identify therapeutic peptides for the treatment of patients with COVID-19. It is using its artificial intelligence platform to identify peptides with antiviral properties as well as peptides with cytokine regulatory properties, with the goal of creating a therapeutic ‘peptide cocktail.’ If successful, this approach has the potential to slow or stop disease progression by both mitigating viral replication and modifying the cytokine-based inflammatory response known to drive respiratory damage in patients with COVID-19.

Speaking earlier this year, Khaldi said: “COVID-19 is a devastating infectious disease, and there is an enormous unmet need to diminish viral replication and tackle the underlying inflammatory response that, if left unchecked, can lead to lung damage and death. It is obvious that more than one solution may be needed to mitigate the impact of the pandemic, and the Nuritas team is eager to help identify therapeutic peptides with the potential to address these two key drivers of the disease’s progression.”

Peptides are naturally suited to addressing infectious diseases as they can disrupt protein-protein interactions. Hansel Gómez Martínez, a research scientist who leads the molecular modelling research at Nuritas, has been leading the PRACE project that is looking for bioactive peptides that can bind to targets on SARS-CoV-2.

Hansel Gómez Martínez

Hansel Gómez Martínez

“The main goal of this project is to find peptides that show activity against the virus responsible for the COVID-19 pandemic,” he explains. “More specifically, we are ideally looking for peptides that are already in the market or at the stage of advanced clinical trials so that they can quickly be repurposed for antiviral therapies.”

The team used a molecular modelling pipeline that involved carrying out docking simulations, molecular dynamics simulations and free energy calculations of all of the peptides being studied against protein targets forming the virus-host cell interactome. At the same time, they trained machine learning predictors to predict if any of the peptide sequences showed any potential for antiviral or anti-inflammatory behaviour. This work generated a massive dataset of characteristics of the peptide sequences examined, which the team is now statistically analysing in order to create an overall ranking.

“The pipeline itself isn’t particularly complicated in terms of methodology,” says Gómez. “What was challenging for us was the volume of simulations we were doing. We considered the entire interactome of the viral infection, meaning every single protein of SARS-CoV-2 that we knew the structure of, combined with every single protein from the human host cells that are thought to be important in the infection process. This meant that the number of targets totalled over 200, which we then had to cross reference against over 700 peptide sequences, giving a total of over 175 000 separate systems.”

Proposed workflow for the Nuritas project

The schematic shows the proposed workflow for the Nuritas project.

This huge undertaking generated terabytes of data, and it is only thanks to access to the Piz Daint supercomputer in Switzerland, using hundreds of GPUs simultaneously, that the team were able to complete their work. “PRACE’s allocation to us was absolutely essential to this project,” says Gómez. “If you look at the alternatives, it would have cost us in the order of several hundreds of thousands of dollars to carry this work out on the cloud, so we are really appreciative of this opportunity.”

The deployment of such an unprecedentedly large pipeline also meant that the team had to create an automation process for its implementation. Many steps had to be connected and the process required a lot of coding and code wrapping on the fly, but Gómez believes that this work will be of great value to the company moving forwards in future projects.

a view of a hit candidate peptide binding to the SARS-CoV-2 Main Protease

A view of a hit candidate peptide binding to the SARS-CoV-2 Main Protease.

“At the beginning of the project, we also had to deal with the fact that there were only very few peptide sequences available publicly, most of them in fact being owned by pharmaceutical companies who are unwilling to disclose them due to their potential value,” explains Gómez.

“To get around this issue, we developed state-of-the-art NLP techniques to crawl through thousands of papers and patents online and find peptide sequences to give us a point to start from. This is another outcome of the project that will be very useful to Nuritas in the future.”

When the final ranking is completed and the best candidates have been identified, Nuritas will be sharing the results of the study and will be open to collaborations with the companies that manufacture the most promising peptides in order to push towards having them used therapeutically in the fight against the COVID-19 pandemic.

This article was also published in PRACE Digest 2020.

More information:

Resources awarded:
40 million core hours (i.e. 588 000 node hours) on Piz Daint GPU hosted by ETH Zurich/CSCS, Switzerland


For all questions about PRACE Communications, promotional and press materials, social media, and publications, email: or phone us on +32 2 613 09 28.

Share This

Share on twitter
Share on linkedin
Share on facebook
Share on email


For all questions about PRACE Communications, promotional and press materials, social media, and publications, email: or phone us on +32 2 613 09 28.

Latest News

Recent Tweets

Let's Stay Connected

Stay up to date with all the latest PRACE news and HPC access call announcements by subscribing to our newsletter: