AI to improve rare disease diagnosis
Two separate sets of European researchers have developed their own artificial intelligence methods to identify rare diseases, for which obtaining a definitive diagnosis can be difficult and time-consuming.
Almost 80% of rare diseases are genetically determined, so it is important for doctors to be able to predict which genetic variants in the patient’s genome may be the cause of the disease. Predicting the cause of the error is not easy, but predicting whether a combination of errors in different genes has the potential to cause a rare disease is even more difficult.
With this in mind, Belgian researchers led by Professor Tom Lenaerts from ULB-VUB’s Interuniversity Institute of Bioinformatics (IB²), developed an AI algorithm that makes it possible to identify combinations of genetic variants or abnormalities that cause rare diseases through computer analysis. Named VarCoPP (Variant Combinations Pathogenicity Predictor), the algorithm makes it possible to simultaneously test the combinations of different variants in gene pairs and to predict their potential pathogenicity.
The AI that underlies VarCoPP is driven by a database of rare diseases called DIDA, which was developed by the same researchers in 2015. The researchers successfully tested the effectiveness and reliability of the algorithm on 23 independent pathogenic gene combinations, and deliver confidence intervals of 95% and 99% to help doctors zoom in on the most important predictions.
The team is now attempting to use these results to identify the genetic causes of rare diseases in patients for whom no cause could previously be identified. They have also introduced a new online diagnostic platform for researchers and clinicians, named ORVAL, based on the algorithm.
VarCoPP and ORVAL provide a novel approach to study variant combinations for rare diseases for which causal genes are known or unknown, such as, for example, the hundreds of autism or epilepsy genes or the 20 genes of the rare Bardet-Biedl syndrome (a genetic disorder which presents blindness, obesity and motor disorders amongst others) where different combinations of genetic variations are likely to be the cause. VarCoPP has been described in Proceedings of the National Academy of Sciences, with ORVAL in the journal Nucleic Acids Research.
Separate to this, researchers from the University of Bonn and the Charité – Universitätsmedizin Berlin have demonstrated how AI can be used to make comparatively quick and reliable diagnoses via facial analysis, enabling patients to move on to early therapy in order to avert progressive damage.
The team’s study used data of 679 patients with 105 different diseases caused by a change in a single gene, which results in abnormalities in the facial features of those affected. For example, mucopolysaccharidosis (MPS) leads to bone deformation, learning difficulties and stunted growth. Kabuki syndrome is meanwhile reminiscent of the make-up of a traditional Japanese form of theatre, resulting in arched eyebrows, a wide distance between the eyes and long spaces between the eyelids.
The researchers deployed software that can automatically detect these characteristic features from a photo. Together with the clinical symptoms of the patients and genetic data, it is possible to calculate with high accuracy which disease is most likely to be involved.
The key AI tool used in the study was the neural network DeepGestalt, developed by AI and digital health company FDNA. The scientists trained this computer program with around 30,000 portrait pictures of people affected by rare syndromal diseases.
“In combination with facial analysis, it is possible to filter out the decisive genetic factors and prioritise genes,” said Prof Dr Peter Krawitz, from the Institute for Genomic Statistics and Bioinformatics at the University Hospital Bonn. “Merging data in the neuronal network reduces data analysis time and leads to a higher rate of diagnosis.
“This is of great scientific interest to us and also enables us to find a cause in some unsolved cases. Many patients are currently still looking for an explanation for their symptoms.”
Nanostructured layers boast countless potential properties — but how can the most suitable...
The goal was to design and deploy a decision-support tool using AI capabilities — mostly...
Sydney start-ups DetectED-X and LENS Immersive have each launched tools designed to take the...