Researchers from the Cantabrian Institute of Biomedicine and Biotechnology (IBBTEC) have developed a tool that predicts pathologies in patients with congenital malformations.
Specifically, as reported by the University of Cantabria (UC), researchers Víctor Sánchez and Álvaro Rada, of the ‘Trasscriptional Regulation in Development and Congenital Diseases Group’, have developed a computational tool, which they have called DOSTRE (Prediction Of STRuctural variant Effects ), able to predict the pathological consequences of the structural variants identified in the DNA of patients with congenital defects.
DESSERT is a public utility tool that can be used through a simple application that does not require advanced IT knowledge. Therefore, it can represent a very valuable and useful tool for the medical community, as it could facilitate the diagnosis and interpretation of a large number of congenital malformations, highlighted the UC.
The results of this work, due to their relevance, have been published in the scientific journal Nucleic Acids Research.
A large number of human diseases are due to DNA alterations. Traditionally it was thought that most of these alterations directly affected some of the genes contained in the human genome. However, in recent years, it has become increasingly clear that many of these alterations occur outside of genes and appear to affect other types of DNA sequences that control gene expression and are called enhancers.
Despite the great medical relevance of structural variants, there are still very few computational tools that allow us to predict their possible consequences, especially if they do not directly alter genes. As a result, there are hundreds of patients with congenital malformations whose genetic diagnosis remains uncertain.
DESSERT was born from the union of the forces of the two researchers, on the one hand, informatics and bioinformatics and, on the other, all the years of experience in understanding the functioning of enhancers, as reported by the principal investigator of the group, Álvaro Rada. .
The tool uses different types of genomic data available in public databases. After integrating them, it is able to predict not only whether a structural variant can cause a congenital disease, but also the underlying mechanism and the genes whose function or expression levels may be affected.
It can currently help in the diagnosis of developmental diseases of the brain, heart, extremities and craniofacial structures, indicated Sánchez, noting that they are also working “to improve it based on the interest of other researchers who want to exploit the tool “.
Sánchez and Rada explained that this tool performs a similar activity to what a geneticist would do with their own criteria. DESSERT implements an algorithm (evaluates a set of rules that must be satisfied) created on the basis of existing knowledge in the field of interpretation of structural variants, which is automatically applied when evaluating the impact of the detected genetic alterations.
The logic followed for the creation of the DESSERT prediction algorithm differs from the one typically used today in the field of artificial intelligence, where the importance and weight of each variable is automatically assigned after training the predictive model with a lot of real data. This typical strategy in the field of artificial intelligence could not be applied because there is no data with sufficient resolution to train a model in the field of interpretation of structural variants that alter the communication between genes and enhancers.
