SHIELD
News & Updates
Since its launch in January 2025, the SHIELD project has been tackling the elusive mechanisms viruses use to enter cells and escape immune defenses. Follow our latest updates and discoveries.
Besides the below news blog, stay connected and follow SHIELD also on social media (LinkedIn, Bluesky) for insights, behind-the-scenes moments, and highlights from our research journey!
Can Less Data Be More? The Key to Smarter AI in Drug Discovery
30 March 2026
Based on the publication by Behnam et al. (2026) in ACS Pharmacology & Translational Science.
We’ve always been told that when it comes to AI, more data is better—but what if the secret to discovering the next life-saving drug is actually less data?
Artificial Intelligence (AI) is set to revolutionize the discovery of medicines, but it needs massive amounts of data. To feed it, some researchers pool small, public datasets together into a single dataset. If this data targets the same protein, it seems like a clever shortcut, right?
Unfortunately, no!
Testing the exact same drug on the same target protein can yield drastically different results if the assay setup varies. In one test for a dengue protease inhibitor, altering testing conditions caused a staggering 100-fold difference in inhibition!
One issue is that proteins are essentially shapeshifters—a phenomenon known as conformational plasticity. Depending on their immediate environment, such as temperature, pH, or salt levels, they slightly alter their form. Consequently, key metrics like IC50 or Kd values depend entirely on the specific experimental conditions used to acquire them.
And that's just one problem. Furthermore, as shown in the figure, cellular EC50 values can swing wildly due to numerous other assay factors.
Feeding this "apples-and-oranges" data into a single AI model creates confusion, rendering predictions practically useless.
And how to solve this? Sort the data first. By grouping results by specific conditions, we train a highly accurate Machine Learning (ML) model for each setup. We then cluster these "expert" models into a smarter, more reliable AI strategy.
In the SHIELD consortium, we are tracking down these hidden variables by comparing different experimental setups. Feeding AI endless amounts of mismatched data won’t improve its output. Higher data quality through consistency is the true key to unlocking the next generation of life-saving drugs.
What Is the Viral Glycan Shield – and Why Does It Matter?
20 November 2025
Have you ever wondered how viruses manage to slip past our immune system’s defenses? One of their strategies is a clever disguise known as the viral glycan shield – and it’s one of the focus areas of the SHIELD project, which even takes its name from this fascinating feature.
In simple terms, the glycan shield is formed of sugar molecules that coat certain positions, called glycosylation sites, on the virus’s surface proteins. You can think of it as a protective armor that helps the virus hide from the body’s immune system. Not every virus wears the same kind of armor: some have only a few sugar “decorations,” while others are covered in a dense layer of them. For instance, within SHIELD’s research scope, the Nipah virus fusion protein (F) has only a few of such sugar sites, whereas the Lassa virus glycoprotein complex (GPC) is almost completely wrapped in a protective layer.
Interestingly, viruses don’t make these sugars on their own. They hijack the glycosylation machinery of the cells they infect, using the host’s own building blocks to mimic the cell surface glycans and avoid immune detection.
This sugar coating is not only a form of camouflage – it also stabilizes the viral particle and helps the virus attach to glycan-binding proteins to enter our cells.
At SHIELD, we’re working to uncover insights about this glycan shield and ultimately, how we might overcome it. Understanding this viral armor could open new paths toward preventing or treating viral infections.
SHIELD at DEFENDER Symposium, Hamburg
6 November 2025
Today, Christian Klein (Heidelberg University) and Leo Hanke (Karolinska Institutet) presented the SHIELD project at the DEFENDER 2025: International Symposium in Hamburg.
A big thanks to Stephanie Pfänder and Luca Bertzbach for the initiative and organisation of this inspiring event - bringing together the projects funded under the call: HORIZON-HLTH-2024-DISEASE-08-20 – Pandemic preparedness and response: Host-pathogen interactions of infectious diseases with epidemic potential.
4 April 2025
We are now online on LinkedIn and Bluesky! Follow us and learn more about the project!
Follow us on LinkedIn!Follow us on Bluesky!
27 February 2025
On the occasion of the project start and the kick-off event, Heidelberg University has published a press release on the SHIELD project. If you want to learn more about SHIELD's objectives and scientific approaches, this is a good and comprehensibly written piece to start with!
Read the press release and find out more!
25 February 2025
The first meeting of SHIELD's Steering Committee took place online. The Steering Committee is comprising the PIs and other representatives of the partners leading the seven work packages of the project. The Steering Committee will prepare consortium decisions and serve as a platform for the overall monitoring of the project progress.
17 and 18 February 2025
The kick-off event for the SHIELD project took place in Heidelberg at the Internationales Wissenschaftsforum. For two days, the SHIELD members discussed the various work packages and administrative aspects.
10 February 2025
The offical logo for the SHIELD projekt is finished! You can see it, for example, here on the SHIELD website. A big thank you to UCgraphics for their great work!
1 January 2025
After months of preperation, the SHIELD project officially started on 1 January 2025. We are excited about what is to come. Particulary, we are looking forward to our kick-off meeting at Heidelberg University in February.
Funding information
Programme: Horizon Europe
Call: HORIZON-HLTH-2024-DISEASE-08-20
Grant agreement no: 101191794
Overall funding: 7 978 165,75 €
Period: 1 January 2025 to 31 December 2029
Co-Funded by the Swiss State Secretariat for Education, Research and Innovation (SERI)
