Post-Doctoral Research Visit F/M Graph neural networks for predicting allosteric signaling

Inria, the French national research institute for the digital sciences; France; Sur place; EUR 35 000 - 45 000; Plein temps

Résumé du poste

A leading research institute in digital sciences in France is offering a postdoctoral position to develop innovative deep learning frameworks. The successful candidate will work on the DynaNova project, focusing on graph neural networks to uncover communication pathways in macromolecular complexes. A PhD in relevant fields and strong experience in deep learning and Python is required. This role includes opportunities for collaboration and supervision of students, with a flexible work environment and additional benefits.

Prestations

Partial reimbursement of public transport costs

7 weeks annual leave + 10 additional days off

Possibility of teleworking and flexible work hours

Professional equipment provided

Social and cultural activities

Qualifications

• Experience in machine learning with a focus on deep learning.
• Proven ability to work independently and in a team.
• Strong coding practices and knowledge of reproducible workflows.

Responsabilités

• Develop novel graph neural network architectures.
• Integrate features from molecular dynamics simulations.
• Collaborate with students for supervision and guidance.

Connaissances

Deep learning

Graph neural networks

Python

Molecular modeling

Formation

PhD in Computer Science, Machine Learning, Bioinformatics, Computational Biology

Outils

PyTorch

Description du poste

Inria, the French national research institute for the digital sciences

Organisation/Company Inria, the French national research institute for the digital sciences Research Field Computer science Researcher Profile Recognised Researcher (R2) Country France Application Deadline 5 Jan 2026 - 00:00 (UTC) Type of Contract Temporary Job Status Full-time Hours Per Week 38.5 Offer Starting Date 1 Apr 2026 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Reference Number 2025-09574 Is the Job related to staff position within a Research Infrastructure? No

Offer Description

This 2-year postdoctoral position is funded by the prestigious Programme Inria Quadrant (PIQ) for the project DynaNova, which aims to advance our understanding of conformational dynamics and allosteric communication in macromolecular complexes. The successful candidate will develop novel graph neural network (GNN) architectures to learn dynamic information from molecular dynamics (MD) simulations of protein-protein and protein-nucleic-acid complexes.

You will join the Delta team at Inria (Université de Lorraine), working closely with Dr. Yasaman Karami, expert in conformational dynamics, allostery, and deep learning for structural biology. The team is growing and offers a highly interdisciplinary environment that brings together researchers in structural bioinformatics, computational chemistry, biophysics, and machine learning.

We have access to major national HPC facilities (Grid5000, Jean Zay, GENCI allocations), including large-scale GPU resources.

Biomolecular function is driven by both structure and dynamics. Understanding long-range communication within macromolecular complexes is essential for deciphering molecular mechanisms and for developing therapeutic strategies. While deep learning has revolutionized structural prediction (e.g., AlphaFold2), allosteric signaling remains poorly understood, largely due to the scarcity of dynamic data.

Our group recently developed:

DynaRepo,a database of molecular dynamics trajectories of more than 700 macromolecular complexes (~5.5 bilions of frames) [1]. DynaRepo is the first MDDB node in France: https://dynarepo.inria.fr/ .

ComPASS,a graph-based method for identifying communication networks in protein–protein and protein–nucleic-acid assemblies [2].

DynamicGT,a dynamic-aware graph transformer for predicting binding sites in flexible and disordered regions [3].

Building on these foundations, DynaNova will leverage a large MD dataset (DynaRepo) and advanced GNN/Transformer models to uncover long-range communication pathways within macromolecular complexes.

The postdoctoral fellow will lead the development of an innovative deep learning framework to learn conformational heterogeneity and decode allosteric signaling.

[1] Mokhtari, O., Bignon, E., Khakzad, H., & Karami, Y. (2025). DynaRepo: the repository of macromolecular conformational dynamics.Nucleic Acids Research, gkaf1130.

[2]Bheemireddy, S., González-Alemán, R., Bignon, E., & Karami, Y. (2025). Communication pathway analysis within protein-nucleic acid complexes.Journal of Chemical Theory and Computation.

[3]Mokhtari, O., Grudinin, S., Karami, Y., & Khakzad, H. (2025). DynamicGT: a dynamic-aware geometric transformer model to predict protein binding interfaces in flexible and disordered regions.bioRxiv.

Design and implement novel graph neural network architectures (Message-passing GNNs, Graph Transformers, graph community-aware latent spaces).

Integrate dynamic, geometric, and biophysical features extracted from large MD simulations.

Train and benchmark models on state-of-the-art datasets.

Collaborate with PhD and Master students, contributing to supervision and scientific guidance.

Contribute to publications, open-source software development, and conference presentations.

Participate in the preparation of a webserver and tools to disseminate the developed methods.

PhD in Computer Science, Machine Learning, Bioinformatics, Computational Biology, or related fields.

Strong experience in deep learning, ideally with PyTorch.

Proven experience with graph neural networks, geometric deep learning, or transformers is a major advantage.

Practical knowledge of Python, clean coding practices, and reproducible ML workflows.

Familiarity with protein structure, biophysics, or molecular modeling (MD, docking, etc is highly desirable.

Ability to work independently, collaborate in a multidisciplinary environment, and communicate effectively in English.

Applications without a strong machine learning and/or computer science component cannot be considered.

Languages FRENCH Level Basic

Languages ENGLISH Level Good

Additional Information

• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking (after 6 months of employment) and flexible organization of working hours
• Professional equipment available (videoconferencing, loan of computer equipment, etc.)
• Social, cultural and sports events and activities