A project around the use of artificial intelligence, smart computing power and genetics to make more diagnoses in patients with rare diseases gets funding. Half a million euros will be made available from the European Regional Development Fund (ERDF) and the province of Gelderland.
This is a project by RadboudUMC, Bio-Prodict and Vartion aimed at diagnosing rare diseases. For although these diseases are individually relationship rare, in total some 7,000 rare diseases are known. Adding up the number of patients with such diseases, the total equals about 8% of the population.
Cause is by no means always clear
In far from all cases, these patients know exactly what causes their disease. In about half the cases, they manage to identify the causes of their diseases. In the other half, this is not the case. Not only does this leave them in limbo, it also means the diseases cannot be targeted.
Radboud's project in collaboration with regional companies Bio-Prodict and Vartion should change this. The project is called 'Diagnostics-in-3D' and focuses on the human genetic code. This code consists of about three billion letters. Many of these are the same, but a significant proportion differ from patient to patient. Laying genomes side by side reveals all kinds of differences.
Financial support
The project is financially supported by the European Regional Development Fund (ERDF) and OP Oost in particular. By doing so, the fund aims to contribute to a structural strengthening of the economy in the provinces of Gelderland and Overijssel. The project will be led by Martijn Huijnen, professor of bioinformatics at the CMBI (Centre for Molecular and Biomolecular Informatics CMBI) at Radboudumc.
Huijnen: "Nijmegen-based Bio-Prodict has developed software that has automated the prediction of the effect of mutations on proteins. By using anonymised patient data and AI from the Radboudumc, great strides can be made in predicting the effect of mutations on the 3D structure of a protein, and thus on its function. For this, the representation of the 3D protein data, the deep learning architecture to be used and the computing power will be optimised. In the latter field, Malden-based Vartion is a specialist. This collaboration of knowledge institution and companies should soon further boost the number of diagnoses in rare diseases. Not only because a correct diagnosis is essential for the development of a treatment, but also because it can be important when deciding on possible future pregnancies."
'Exactly what difference makes you sick?'
"The problem when looking for the cause of a rare inherited disease is that you cannot easily pinpoint exactly which difference in the genome is causing the disease. Suppose you find 1,000 genetic differences in someone compared to the standard, which of these differences exactly makes you sick? Most differences have no effect at all. So we often look at the translation from gene to protein. Proteins do the work in our bodies. So it is interesting to see if a genetic difference also leads to significant changes in the corresponding protein, making it work less well. That could be the cause of the rare disease."
A protein is a string of folded amino acids. This folding is essential for the protein's function and is compared by researchers to a biological origami. The folding determines the protein's contact with other proteins, which determines its work. A genetic difference that causes a change in the shape of a protein could be a cause of a rare disease. "In this way, we are increasingly - and successfully - searching for the cause of rare diseases," Huijnen explains. "In our 'Diagnostics-in-3D' project, Bio-Prodict, Vartion and the Radboudumc now want to automate and accelerate this approach. In doing so, we focus mainly on the rare diseases that cause intellectual disabilities, a subject in which we have already built up a lot of expertise in Nijmegen."
VarCoPP
There are more developments around the use of AI to diagnose the causes of rare diseases. For example, a team of Belgian researchers led by the ULB-VUB 'Interuniversity Institute of Bioinformatics in Brussels' announced in June 2019 that they had developed an AI method to identify potential genetic causes for rare diseases based on computational analysis.
This is a project by a research team led by Professor Tom Lenaerts. The algorithm can use computer analysis to identify combinations of genetic variants or abnormalities that cause rare diseases. The algorithm was developed by the research team in collaboration with Prof Guillaume Smits, affiliated to theCentre for Human Heredity at ULB, Erasmus Hospital and Queen Fabiola University Children's Hospital. The method is called VarCoPP (Variant Combinations Pathogenicity Predictor).
Author: Wouter Hoeffnagel
