Mots-Clés

Genomic prediction maize drought regulatory networks AI

Description

Work environment, missions and activities

The postdoctoral fellow will be welcomed in the GEvAD team (Evolutionary Genetics and
Crops Adaptation, http://moulon.inrae.fr/en/equipes/gevad/) at the UMR Quantitative
genetics and Evolution (GQE) – Le Moulon (Gif-sur-Yvette, France). GQE is part of IDEEV
(the Institute for the Diversity, Ecology and Evolution of the Living World,
https://www.ideev.universite-paris-saclay.fr/en/), located on the Paris-Saclay campus. The
GEvAD team combines various approaches, including field and greenhouses experiments,
theoretical (models, stat development) and applied population genetics, genomics, systems
genomics, to understand the evolutionary mechanisms behind the domestication and
environmental adaptation of crops. The postdoctoral fellow will also be collaborating with
members of the GQMS team (Quantitative genetics and Plant Breeding Methodology), whose
research focus on developping experimental and theoretical approaches, statistical methods
and decision support tools to understand maize diversity and decipher the architecture of
quantitative traits to optimize maize breeding schemes.
The postdoctoral fellow will participate to the ANR JCJC NETWITS project, led by Maud
Fagny, which aims at exploring the role of gene regulatory networks structure in maize
response to drought. GQE has an important expertise on the molecular bases of maize
drought response, including association studies and the inference of gene regulatory
networks. We thus have identified numerous loci, both genes and regulatory elements,
potentially involved in determining maize yield in response to drought.
The postdoctoral fellow, specialist in quantitative genetics, will develop a yield prediction
method for maize in drought condition that will leverage the available information. The aim
will be to integrate prior biological information about gene expression regulation and natural
selection within the model. Working with data generated by the members of the NETWITS
project and others, the postdoctoral fellow will pursue the following integration steps:
1/ Classify the polymorphisms into different categories according to their expected
importance in the regulatory network, GWAS, eQTL and population genetics analyses.
2/ Use bio-informed methods such as (Bertolini 2025), GFBLUP (Edwards 2016, Fang
2017), or BayesRC+ (Fikere 2018, Mollandin 2022) to directly integrate the regulatory
network based classification in the model.
3/ Eventually implement a bio-informed neural network to directly integrate the regulatory
interactions in the model (NetGP, Zhao 2025; DLGBLUP, Shokor 2025).
The predictive abilities of these models will be compared to reference models such as
GBLUP in different prediction scenarios potentially involving genotype x environment
interactions. For this, the postdoctoral fellow will design cross-validation scenarios based on
a multi-environment trial of 250 maize hybrids evaluated in 25 environments. They will
particularly focus on the prediction of genetically distant material. All datasets are already
curated and ready to use.
The postdoctoral fellow will be supervised by M. Fagny (GEvAD), Renaud Rincent and
Tristan Mary-Huard (GQMS). They will collaborate closely with the other participants of the
NETWITS project, in order to integrate their results in the model. The postdoctoral fellow
will also be responsible for supervising interns (licence or master students) and to help
training the PhD students of the team in quantitative genetics.
The work can be performed partly remotely (2 days/week maximum).

Required skills and knowledges

• The candidate is required to hold a PhD.
• Academic knowledges: Advanced knowledges in quantitative genetics are required; an
  experience with deep neural network models would be preferred, but not indispensable.
  Knowledges in systems genomics/gene regulatory networks or in population genetics will be
  appreciated but are not necessary.
• Bioinformatics: programming skills are required in at least one of the following languages:
  python or R. Skills in shell and SLURM-based computational clusters will be appreciated.
  Basic knowledge of the FAIR principles and about git usage are required as all scripts will be
  developed and made publicly available following the FAIR management standards.
• Communications skills: writing scientific articles, giving poster presentations and talks are
  required skills to valorize the scientific results. Spoken & written English: B1 to B2 level
  (Common European Framework of Reference for Languages) is required.
• Interest in supervising students will be appreciated.

Bibliography

Bertolini E., et al., 2024. Genomic prediction of cereal crop architectural traits using models
informed by gene regulatory circuitries from maize. Genetics 228(4): iyae162.
https://doi.org/10.1093/genetics/iyae162
Edwards, S. M., Sorensen, I. F., Sarup, P., Mackay, T. F., & Sorensen, P. (2016). Genomic
prediction for quantitative traits is improved by mapping variants to gene ontology categories
in Drosophila melanogaster. Genetics, 203(4), 1871–1883.
https://doi.org/10.1534/genetics.116.187161
Fang L., et al., 2017. Use of biological priors enhances understanding of genetic architecture
and genomic prediction of complex traits within and between dairy cattle breeds. BMC
Genomics 18:604. https://doi.org/10.1186/s12864-017-4004-z
Fikere M., et al., 2018. Genomic Prediction Using Prior Quantitative Trait Loci Information
Reveals a Large Reservoir of Underutilised Blackleg Resistance in Diverse Canola (Brassica
napus L.) Lines. The Plants Genome 11:170100.
https://doi.org/10.3835/plantgenome2017.11.0100
Zhao L., et al., 2025. Genomic Prediction with NetGP Based on Gene Network and Multi-
omics Data in Plants. Plant Biotechnology Journal 23:1190-1201.
https://doi.org/10.1111/pbi.14577
Mollandin F., et al., 2022. Accounting for overlapping annotations in genomic prediction
models of complex traits. BMC Bioinformatics 23(1):365. https://doi.org/10.1186/s12859-
022-04914-5
Shokor, F., Croiseau, P., Gangloff, H., Saintilan, R., Tribout, T., Mary-Huard, T., &
Cuyabano, B. C. D. (2025). Deep learning and genomic best linear unbiased prediction
integration: An approach to identify potential nonlinear genetic relationships between traits.
Journal of Dairy Science 108(6) :6174-6189. https://doi.org/10.3168/jds.2024-2