Data Integration using Bayesian Credal Networks

 CDD · Postdoc  · 12 mois (renouvelable)    Bac+8 / Doctorat, Grandes Écoles   Computational Systems Biomedicine Lab, Institut Pasteur · Paris (France)

Mots-Clés

Machine Learning, Clinical Research, Bayesian Networks, Credal Networks

Description

Project

The European DECIDER research project (http://deciderproject.eu) aims to develop diagnostic tools and treatments for high-grade serous ovarian cancer with the help of AI methods—in particular, to identify earlier those patients who do not respond well to first-line treatments, and to find effective treatments to patients with a drug-resistant cancer.

In this project, we team with several European research groups to develop a decision support system that will integrate genomics and clinical data along with inferences made by AI models. Together with practitioners of the Molecular Tumor Board, we develop a visualization of the relationships between the data produced by the project, from sequencing data up to treatment suggestions made by AI models.

To this end, we plan to use an extension of Bayesian Networks that can model uncertain probabilities: Credal Networks. Our goal is to create a simple directed graph model of high-level biological observations (gene expression, mutations, biological pathways, drug effects, etc.) that also works in cases of conflicting or missing data.

The candidate is expected to

  • Study the project context and work with the project team(s) to identify all the sources of relevant data,
  • Implement a proof-of-concept, integrated with the existing codebase,
  • Evaluate how well the Credal Network formalism is able to solve the task at hand,
  • Design and implement a reproducible validation protocol,
  • Work with the project team to implement a visualization web application,
  • Present their work in team meetings, and to
  • Document their work in a related publication.

Research Group

As part of the Computational Biology department of Institut Pasteur, the Computational Systems Biomedicine lab (http://bit.ly/ip-csb) focuses on data-driven identification of computational models that capture biomedical complexity in the context of complex disease. We are excited to leverage our computational skills to advance cancer precision medicine in a multi-disciplinary academic environment. The lab values creative thinking, teamwork, open-mindedness, inclusivity and kindness.

Position

This is a postdoctoral full-time work contract with an initial fixed term of 1 year. Institut Pasteur offers a competitive compensation, depending on work experience, and associated benefits. The department allows 2–3 days of remote work per week. Postdoctoral fellows are eligible to multiple training programs, individual career coaching, and various occasions to participate to the institute life. Institut Pasteur provides an international and dynamic research environment with access to a broad range of state-of-the-art computing equipment.

Key competences

Essential

  • Ph.D. in a relevant field related to Bayesian/Credal Networks or Data Integration
  • Self-driven and able to work independently in an organized manner while being part of a team
  • Strong will to learn
  • Good communication skills in English (the team’s primary language)
  • Good programming skills in Python
  • Scientific rigor on data management and version control systems for reproducibility

Desirable

  • Background in bioinformatics or precision medicine
  • Knowledge of the aGrUM library
  • Knowledge of Git, Javascript, and/or Django

Candidature

Procédure : To apply or inquire, please e-Mail benno.schwikowski@pasteur.fr

Date limite : None

 https://research.pasteur.fr/en/job/postdoctoral-position-data-integration-and-bayesian-credal-networks/

Offre publiée le 4 août 2023, affichage jusqu'au 31 décembre 2023