Mots-Clés

deep learning population genetics generative models interpretability inference evolution

Description

PhD projects on Deep Learning for Population Genetics

Supervisors: Flora Jay (CR CNRS), Guillaume Charpiat (CR INRIA),

Direct collaborators: Burak Yelmen (postdoc U Paris-Saclay), Cyril Furtlehner (CR INRIA),

Aurélien Decelle (Complutense University of Madrid)

Contacts: flora.jay@lri.fr ; guillaume.charpiat@inria.fr

web: https://flora-jay.blogspot.com/ ; https://www.lri.fr/~gcharpia/

Location: LISN (Paris-Saclay University)

machine learning and bioinformatics groups
1, rue Raimond Castaing

91190 GIF-SUR-YVETTE

When/Duration: September 2022, 3 years.
 

We are looking for one highly motivated candidate to do a PhD in our lab and suggest two potential projects. They are follow-up research of two of the lab papers:

- "Creating artificial human genomes using generative neural networks", Yelmen et al 2021
- “Deep learning for population size history inference: design, comparison and combination with approximate Bayesian computation”, Sanchez et al 2020

Please note that we would happily talk with candidates that have an alternative project in mind, as long as it falls in the scope of deep learning for population genetics.

In their motivation letter, the candidates should explain which of these topics they are interested in (it could be both).

Keywords

deep learning, population genetics, generative models, interpretability, inference, evolution

apprentissage profond, génétique des populations, modèles génératifs, interprétabilité, inférence, évolution

Projects
Details at https://www.lri.fr/~fjay/archive/phd2022_DL_popgen.pdf

* Subject 1: Creating artificial human genomes using generative neural networks

* Subject 2: Inferring the evolutionary past of populations from genomic data and interpreting neural networks.

* Subject 3: Your project

Feel free to contact us if you have a strong opinion on the PhD project that you would like to pursue, as long as it remains in the scope of machine learning and population genetics!

Requirements

The ideal candidate should be good at python scripting and machine learning/statistics concepts. Experience with deep learning is a plus. Familiarity with some of the following topics: genomics, population genetics, generative models, bash scripting and high-performance computing is not mandatory but a clear plus. Being curious and autonomous is highly recommended for any PhD. Being able to communicate, read and write in English (French not required).

Salary: regular PhD stipend in academia. The PhD fellowship is fully funded (ANR grant) for 3 years from September 2022.

Application: Ideally the candidate should provide a CV, motivation letter, past scores/ranks, names and emails of previous mentors that can be contacted