In this expert commentary, the project DTRIP4H partner University of Helsinki is presented as a contributor to Use Case 2 – “Virtual reality drug combiner.”
Associate Professor Jing Tang outlines the main challenge in the current field – where AI drug discovery often focuses on single drugs despite resistance and toxicity issues—and explains how this use case will address it by developing accurate models for predicting promising cancer drug combinations and integrating them into an interactive virtual reality platform for intuitive exploration and experimental validation by end users:

"My name is Jing Tang, and I am currently an associate professor of medical bioinformatics at the University of Helsinki. My main research focus is drug discovery, and in this particular project (DTRIP4H – Enabling Decentralised Digital Twin Era in existing Research Infrastructures for Predictive, Preventive, Personalised, and Participatory Health), we aim to develop a virtual-reality–supported model for predicting effective drug combinations for cancer treatment.

At the moment, most AI tools in drug discovery are designed to build predictive models for single drugs. However, for many diseases—especially cancer—single-drug therapies often lead to drug resistance. In practice, a drug may work only for a short period before cancer cells adapt, after which the treatment loses effectiveness. This is why we need more powerful treatment strategies, such as drug combinations, which can improve therapeutic efficacy while also minimizing toxicity.

So that's why we are aiming to develop more accurate models for predicting which combinations will be the most promising to achieve maximum efficacy while minimising toxicity. And also, we want to develop these models one step further by incorporating virtual reality technologies, to make the models more interactive and more intuitive for end users, such as pharmaceutical companies, and to enable more interaction between human input and the AI model predictions. So that's why we also want to incorporate experimental validation, to make sure that our model predictions are as accurate as possible using the microfluid-based technologies.

And in the end, we are creating a virtual reality platform to visualize all the predictions, so that users will have a more intuitive understanding of why such a model has been utilized and what the interpretation of those model predictions is, which is also one of the major aims for AI development: to make everything interpretable and reusable in a sense."

We sincerely thank Professor for sharing his valuable insights and for his contribution to the DTRIP4H project.

Find more information on our project:
🔗https://www.dtrip4h.eu/about