Resume

Summary

PhD candidate in Biostatistics & Genetic Epidemiology at INSERM, Paris. I develop graph-based machine learning methods, e.g. knowledge graphs and heterogeneous graph neural networks, to extract genotype-phenotype associations from large-scale cohort data. Committed to open science, reproducibility, and supporting women in machine learning.

Research Experience

PhD Student in Biostatistics & Genetic Epidemiology (September 2024 – present)
Université Paris-Saclay, UVSQ, Inserm, CESP / HiDiBiostat team, Villejuif, France Supervised by Anne-Louise Leutenegger, Anaïs Baudot (MMG, Aix-Marseille Université), Anne-Sophie Jannot (APHP/HeKA)
Funded by ANR PEPR Santé Numérique — M4DI project

  • Developing graph-based machine learning methods to extract genotype-phenotype associations from heterogeneous, multi-scale cohort data (10,000 participants)
  • Integrating clinical, pharmacological, and genomic data using network representations and deep learning approaches
  • Conducting genome-wide association studies on quantitative biological traits and benchmarking against existing catalogs
  • Regular research visits to Marseille Medical Genetics (MMG), Aix-Marseille Université
  • Active participant in the M4DI ANR-PEPR Santé Numérique consortium

Biostatistics Consultant (September 2022 – September 2024)
IT&M Stats (L’Oréal client), Boulogne-Billancourt

  • Statistical analysis of cohort studies linking skin biomarkers with clinical signs (PLS, factorial analyses, linear mixed models)
  • Project management and monitoring of a 10,000-participant cohort

Statistician Intern (March – September 2022)
Chanel, Paris

  • Clinical results database design, data visualisation, structural equation modelling

Education

M.Sc. in Data Sciences for Biology (2019 – 2022)
Institut Agro, Rennes
Machine learning, multivariate statistics, bioinformatics, perception data analysis
Thesis: Construction and Evaluation of Clinical Performance Scores

Diploma of Technological Studies, Biological Engineering (2017 – 2019)
Université d’Avignon
Statistics, bioinformatics, genetics, plant and animal biology

Skills

Machine Learning & Statistics — Graph neural networks, contrastive learning, knowledge graph embeddings, GWAS, high-dimensional data integration, clustering, dimensionality reduction, network analysis.

Programming & Tools — Python (PyTorch, Polars, NetworkX, scikit-learn), R (Shiny, tidyverse), Bash, Git, Docker, Quarto, PLINK, GCTA, VEP, CADD, SQL (BigQuery), GCP.

Languages — French (native), English (C1 — TOEIC 975), Spanish (B2).

Scientific Output

Conference Poster — Knowledge graph to dissect genotype-phenotype associations
Ghestem F., Marenne G., Genin E., Jannot A-S., Baudot A., Leutenegger A-L.
IGES, Cologne, Germany, September 2025. HAL

Flash Talk & Poster — BIOTIC Annual Day, Marseille, May 2025.

Journal Article — Novel indices reveal that pollinator exposure to pesticides varies across biological compartments and crop surroundings
Laurent M., Bougeard S., Caradec L., Ghestem F., et al.
Science of The Total Environment, 927, 172118, 2024. DOI

Software — Rnem (R Shiny) — simulation app for nematode-plant dynamics and virulence evolution
Ghestem F., Touzeau S., Calcagno V., Mailleret L. — INRAE, 2021. HAL · App

Community

Co-organiser — StatGen Annual Meeting (December 2025, Campus Odéon, Paris)
Built the conference website, coordinated speakers, managed participant communications and mailing list, contributed to programme design..

Co-organiser — M4DI PhD Student Annual Retreat (October 2025)
Designed workshops, coordinated programme, and facilitated sessions for the M4DI ANR consortium PhD student retreat.

Mission Doctorale — StatGen Network Launch (May 2025)
Contributed to the launch and scientific animation of the INSERM StatGen thematic network.

Career Talk — Journée des Métiers (March 2025, Institut Agro Rennes)
Presented industry-to-academia trajectory to prospective students.