PhD position AI-driven functional DNA interpretation for molecular diagnostics
In this role, you will contribute to cutting-edge computational methods that can help provide answers to millions of undiagnosed patients. Your work will focus on revolutionizing the classification of genetic variants and diseases using functional evidence and explainable AI, bridging the gap between advanced machine learning and clinical application.
Your tasks are:
Project AI-driven functional DNA interpretation for molecular diagnostics
Rare genetic diseases affect approximately 1 in 10 people, yet the majority never receive a molecular diagnosis. Without a diagnosis, patients are left without a prognosis, effective treatment options, or access to the right support groups. A molecular diagnosis hinges on classifying DNA variants as pathogenic (i.e. disease-causing) or benign and assignment to disease phenotypes. A significant portion of these variants—single base-pair changes known as missense variants—alter protein structures and functions. Despite decades of research and hundreds of predictive tools, about half of these missense variants are classified as Variants of Unknown Significance (VUS), meaning there is not enough evidence for a classification as either benign or pathogenic. As a result, much DNA variation with remains untapped, preventing life-changing diagnoses for countless individuals.
The field is currently undergoing a paradigm shift, moving beyond traditional prediction models based on indirect evidence (such as evolutionary conservation and allele frequencies) toward direct functional evidence, including protein stability and activity. This shift is driven by recent breakthroughs such as AlphaFold and AlphaMissense, opening new avenues for variant interpretation. In addition, better phenotypes can be extracted from health records using AI.
In this PhD project, you will work on developing computational methods that generate close approximations of functional evidence and meaningful predictions for variant classification, with a strong focus on explainability. This means not only building AI models but also carefully selecting and interpreting functional features to ensure clinical relevance. Once developed and validated, your methods will be implemented in a clinical setting, directly impacting patient care.
The position is part of the Genomics Coordination Centre (GCC), the ‘big data science’ research & service hub of the University Medical Centre Groningen (UMCG) and University of Groningen (rank 66 worldwide, 3rd best place to work in EU), hosted by the Department of Genetics. Our mission is to accelerate scientific discovery in health data with innovative methods and tools that expedite medical research and improve people's lives, using open source software and large computer ‘clouds’, in particular the MOLGENIS software that we lead, but also DataSHIELD, Singularity, RedCap, XNAT, OpenStack etc.
This is a full-time PhD contract for 4 years and an excellent environment for further development. First, a temporary one-year position will be offered with the option of renewal for another 3 years. Your salary will be a minimum of € 3.017,- gross per month in the first year and a maximum of € 3.824,- (PhD salary scale) in the final (4th) year, based on a full-time appointment. In addition, the UMCG will offer you 8% holiday pay, and 8.3% end-of-year bonus. The conditions of employment comply with the Collective Labour Agreement.
If you’re excited about pushing the boundaries of AI in precision medicine and making a tangible difference in rare disease diagnostics, we encourage you to apply!
Please use the the digital application form at the bottom of this page - only these will be processed. You can apply until 23 March 2025. Within half an hour after sending the digital application form you will receive an email- confirmation with further information.
The UMCG has a preventive Hepatitis B policy. The UMCG can provide you with the vaccination, should it be required for your position. In case of specific professions a ‘Certificate of Good Conduct’ is required.
© BSL Media & Learning, onderdeel van Springer Nature