Federico Melograna


After completing a Bachelor’s and a Master’s degree in the field of Statistics and Data Science, I pursued my inclination for advance modelling with a one-year master in Deep Learning. Contextually, I started working in the industry as a hybrid Software developer / Data scientist role to fill my knowledge gap in computer science while developing, maintaining and enhancing a Computer Vision project.  Then, I found myself more and more passionate about networks and their structure, wanting to apply the learned techniques in a different environment.  The omics data were immediately my choice, given the richness of information and the endless integration opportunity. Although I do not have the standard career path and background, I feel that the combination of a mathematical background with modelling skills can make a tangible difference in the omics world.

Ku Leuven

Department of Humans Genetics, Leuven, Belgium

Supervisor: Prof. Dr. Dr. Kristel Van Steen

Secondment 1: Apply state-of-the art methods to detect and compare modules in node/edge weighted personalized networks

Secondment 2: Learn to evolve from individual-specific gene-signatures to risk/disease management

ESR 1 Project

Development of individual-specific molecular networks

Describing a system implies describing its behaviour and important control mechanisms that regulate this behaviour. Crucial in this process are interactions, which may occur at different levels or scales, and thus network theory and network visualization are increasingly being used to understand biological mechanisms operating in human systems. However, an individual, especially when in poor health, is likely to deviate from the “norm” in human systems. In this project, we wish to develop omics data integrative gene-based networks to enhance PM. Such a network would enable the identification of gene modules that are subject-specific (in network nodes/edges) and comprise multi-layer cellular information. It goes beyond existing work in that genes are considered to be complex multi-omics systems, and that statistical significance is assessed for individual-specific nodes/edges (in contrast to f.i. Menche et al. 2017 and Kuijjer et al. 2018). We aim to achieve our goal by building upon the aforementioned references and our work on gene representations using diffusion kernels and network theory (Fouladi et al. 2018). Personalized gene omics-integrative signatures will primarily be derived by combining genome, transcriptome and epigenome data for complex diseases with an inflammatory component.


Scientific Interests

  • Machine learning and Deep learning
  • Network science and Statistics
  • Precision medicine

General Interests

  • Skiing
  • History
  • Sports