Mots-Clés
machine learning
deep learning
explainable AI
XAI
knowledge-guided deep learning
interpretability
genomics
genomic prediction
genomic selection
gene regulatory networks
interpretable AI
Description
Two PhD positions in machine learning techniques applied to the analysis of genomic data
Highly motivated students are invited to apply for PhD positions in a collaborative project that aims to develop interpretable deep learning techniques for modelling genomic data in plants and animals. The project involves three teams at the University of Liège in Belgium: the Department of Electrical Engineering and Computer Science - Montefiore Institute (Professors Pierre Geurts and Vân Anh Huynh-Thu), the Unit of Animal Genomics - GIGA (Dr Tom Druet) and the Translational Plant Biology Lab – InBioS (Professor Marc Hanikenne).
This interdisciplinary project aims to develop innovative deep learning models for genomic applications, with the objective of both advancing machine learning methodologies and addressing real-world applications in genomics. In particular, deep learning methods will be optimized for the genomic modelling of complex traits to improve genomic prediction accuracy and identify genomic features associated with genetic variation (including genetic variants). In addition, deep learning approaches will be used to decipher regulatory responses to environmental stress. To this end, single-cell RNA-seq experiments will be performed to identify differentially expressed genes, gene regulatory networks, and regulatory variants associated with the stress response. The project will use primarily data from cattle and Chlamydomonas reinhardtii, a unicellular green alga. To support these applications, methodological research will focus on areas such as explainable AI, knowledge-guided deep learning, and deep learning-based gene regulatory network inference.
Environment
The Montefiore Institute (https://www.montefiore.uliege.be/) is active in many fundamental and applied research topics from various fields, including artificial intelligence (AI) and machine learning (ML). It consists of about 150 researchers, several of whom are active in machine/deep learning, optimization, dynamical non-linear systems, and the application of these methods in various areas such as computational biology, physics, computer vision, and power systems.
The Unit of Animal Genomics (https://www.gigauag.uliege.be/cms/c_4254739/en/gigauag) focuses on the forward genetic dissection of Mendelian and complex traits in domestic animals and humans. One of the key projects of the group is the development of approaches for the utilization of molecular information in livestock breeding, including genomic selection. The lab is based in the Interdisciplinary Cluster of Applied Genoproteomics (GIGA-Research) which comprises more than 600 researchers (https://www.giga.uliege.be).
The Translational Plant Biology lab is affiliated to the InBioS Research Unit of ULiège (https://www.inbios.uliege.be/), which comprises over 120 researchers working on Integrative Biology. This membership enables the lab to collaborate with researchers from different biology fields, such as protein engineering and systems biology, fostering an interdisciplinary environment. The lab is also a member of PhytoSYSTEMS (https://www.phytosystems.uliege.be).
Applications
Candidates should have strong interest and experience in one or several of the following fields: machine learning and artificial intelligence techniques, computer science, data modelling, animal or plant genomics. The work will be performed in an international environment with communication in English.
PhD positions are fully funded for 4 years, with funding available from October 1st 2025 at the earliest. Applications will be analyzed regularly (starting on September 1st, 2025), and positions will remain open until filled. Applications should be sent per e-mail to p.geurts@uliege.be, vahuynh@uliege.be or tom.druet@uliege.be, and include a C.V., a letter of interest in the position and a list of referees with contact information.