New web tool for the clinic facilitates the interpretation of genetic information

Genome sequencing is increasingly being used in clinical practice. Nevertheless, the interpretation of rare genetic mutations remains difficult, even for well-studied disease-causing genes. Current prognostic models can already help in the interpretation of these mutations, but tend to classify benign mutations as disease-causing and thus as false positives. Research teams from the Max Planck Institute of Molecular Cell Biology and Genetics (MPI-CBG) in Dresden, the Centre for Systems Biology Dresden (CSBD) in Germany and Harvard Medical School in Boston, USA, have developed a tool called Deciphering Mutations in Actionable Genes (DeMAG) and published it in the journal Nature Communications. The open-source web server DeMAG (demag.org) provides an interpretation of the effects of all potential single amino acid mutations in 316 clinically relevant genes that cause diseases for which there are already predictive diagnoses and treatments. DeMAG reduces the rate of false positives, allowing clinicians to more accurately assess the impact of mutations in these genes, as fewer benign mutations are categorised as disease-causing. As a result, DeMAG can support clinical decision-making.

In recent years, genome sequencing has become cheaper and more advanced. This enables doctors to increasingly use sequencing for diagnostic purposes on the one hand and scientists to investigate more research questions on the other. At the same time, there is no clear clinical interpretation for many discovered mutations. Uncertainty about whether a mutation causes a disease can be stressful for patients and lead to psychological impairment, morbidity and treatment costs due to under- and overdiagnosis. Although tools already exist to predict the functional impact of these variants, their performance is limited due to limited clinical data. This makes it difficult to distinguish between pathogenic (disease-causing) and benign (neutral) variants within a given gene and often leads to mutations that do not cause disease being categorised as pathogenic. Solving these difficulties is crucial for the development of a reliable prediction method for clinical applications.

The research group of Agnes Toth-Petroczy at the MPI-CBG and the CSBD, together with Christopher Cassa, Assistant Professor of Medicine in the Department of Genetics at the Brigham College of Medicine;Department of Genetics at Brigham and Women's Hospital at Harvard Medical School, and Ivan Adzhubei, research associate in the Department of Biomedical Informatics at Harvard Medical School, developed a statistical model and the DeMAG web server that achieve high accuracy in the interpretation of genetic mutations in disease genes. To train the model, the researchers carefully selected known pathogenic and benign mutations. „We used clinical and other population databases. We only selected mutations for which there was a consensus on the clinical interpretation of the data among several investigators, such as physicians and genetics laboratories. We also included data from populations that are underrepresented in current databases, such as Korean or Japanese populations, to make the model more representative and accurate," explains Federica Luppino, first author of the research and a PhD student in Toth-Petroczy's group. DeMAG has a new function, the „partner score“, which identifies groups of amino acids in a protein that have the same clinical effect. With the partner score, DeMAG exploits relationships between amino acids, relying on evolutionary information from the genomes of many organisms and the recent AI (artificial intelligence) revolution in predicting the 3D shapes of proteins using the AlphaFold algorithm developed by Google DeepMind.

Agnes Toth-Petroczy, who led the study, summarises: „With the tool, we provide a basic framework for the integration of clinical and protein data that supports the assessment of effects due to mutations. We hope that our tool and web server will facilitate clinical decision making. The newly developed functions can also be used for genes and organisms beyond humans.“ The DeMAG code is available on GitLab (https://git.mpi-cbg.de/tothpetroczylab/DeMAG) and all data is available on the web server at https://demag.org/ freely accessible.

&about the MPI-CBG
The Max Planck Institute of Molecular Cell Biology and Genetics (MPI-CBG) is one of over 80 institutes of the Max Planck Society, an independent non-profit organisation in Germany. 600 people from 50 countries from a wide range of disciplines work at the MPI-CBG and are driven by their thirst for research to clarify the question: How do cells organise themselves into tissues? Research at the MPI-CBG covers the widest possible range of different levels of complexity: at the level of molecular networks, cell organelles, cells, tissues, organs or even entire organisms. www.mpi-cbg.de

&about the CSBD
The Centre for Systems Biology Dresden (CSBD) is a cooperation between the Max Planck Institute of Molecular Cell Biology and Genetics (MPI-CBG), the Max Planck Institute for the Physics of Complex Systems (MPI-PKS) and the TU Dresden. The interdisciplinary centre brings together physicists, computer scientists, mathematicians and biologists under one roof. The scientists work together to develop computer-aided and theoretical methods to better understand biological systems. www.csbdresden.de

Press release by "idw - Informationsdienst Wissenschaft" from 21 April 2023