Dr. Barbara Gravel of the AI Laboratory at the Vrije Universiteit Brussel (VUB) and the Interuniversity Institute of Bioinformatics at the Université Libre de Bruxelles (ULB), has developed innovative bioinformatics techniques that shed new light on the genetic causes of complex diseases. Her PhD research focused on understanding how combinations of genetic variants, rather than single mutations, contribute to conditions such as cardiac disorders, hearing loss, visual impairment, and male infertility.

Dr. Gravel explained: “Genetic variants are small changes in the DNA that can affect how proteins function, leading to diseases. While much of genetics research has focused on how a single variant in one gene can cause a disease, I explored how two variants across different genes can combine to create a disorder. It’s a complex challenge, because individually these variants might not cause any harm, but when inherited together from two parents, they can lead to disease in their children.”

Gravel’s work uses advanced machine learning algorithms to analyze vast datasets and identify these complex variant combinations. To power her research, she and a team of fellow PhD students created a comprehensive database called "OLIDA," which compiles genetic data from over 300 scientific studies, covering more than 1,800 variant combinations linked to 200 diseases.

“The first part of my PhD involved collecting data on variant combinations from the scientific literature. We meticulously gathered information on which variants were associated with which diseases, scoring the strength of these associations to ensure accuracy,” Gravel said. “This effort led to the creation of OLIDA, a freely available resource that researchers can now use to develop their own predictive algorithms.”

Once the database was completed, Gravel turned her focus to developing an algorithm that could be used by clinicians to better understand and diagnose rare genetic diseases. This tool allows doctors to prioritize potential genetic combinations responsible for a patient’s symptoms, significantly narrowing down the possibilities.

"When a clinician sequences a patient's DNA, they are often faced with tens of thousands of genetic variants, most of which are irrelevant. My algorithm helps them filter through the noise, ranking variant combinations that are most likely causing the patient’s disease,” she said.

Gravel's algorithm was tested using data from patients in Estonia suffering from male infertility. Her results confirmed that the tool was highly effective, consistently placing disease-causing combinations within the top 50 rankings, thereby saving clinicians valuable time and effort.

Prof. Dr. Lenaerts of the AI Lab (VUB) and Interuniversity Institute of Bioinformatics (ULB), who supervised Dr. Gravel’s PhD research added: “Barbara’s research not only advances the field of bioinformatics but also opens the door to the discovery of new disease-associated genes, which in turn may have an immense impact on healthcare.”

Reference:

Barbara Gravel, Alexandre Renaux, Sofia Papadimitriou, Guillaume Smits, Ann NowĂ©, Tom Lenaerts 2024 Prioritization of oligogenic variant combinations in whole exomes, Bioinformatics, Volume 40, btae184, https://doi.org/10.1093/bioinformatics/btae184

Contact:

Barbara Gravel: barbara.gravel@vub.be +32456299366