AI Genomics

Scientific Blog

AI

Montpellier Genomics: Where AI Meets Genetic

AI pest detection on grapevines.

Genomics in Montpellier: Why the City Is Becoming Europe's Hub for AI-Driven Genetic Research

Genomics is one of the most data-intensive sciences on Earth. Sequencing a genome is now fast and affordable. Making sense of what that sequence means for a disease, a crop, a population, an ecosystem remains extraordinarily hard.

Montpellier has quietly become one of Europe's most important cities for that second challenge. Not just for generating genomic data, but for building the scientific and computational infrastructure to interpret it. And within that ecosystem, Bionomeex has developed a technology that addresses one of the deepest unsolved problems in genetics published in Genome Biology and actively deployed in research environments across Europe and North America.

This article maps the genomics landscape of Montpellier and explains where Bionomeex sits within it.

Montpellier's Genomics Infrastructure - A Concentration Rare in Europe

Few cities outside major capitals can match the density of genomics research institutions that Montpellier has assembled. Understanding this landscape is essential for understanding why Bionomeex's AI-driven approach to genetic analysis is grounded in a uniquely rich scientific environment.

IGMM - Institut de Génétique Moléculaire de Montpellier

The Institute of Human Genetics (IGH, CNRS and University of Montpellier) is a leading research center whose work spans genome regulation, chromosome biology, and the genetic basis of human disease publishing consistently in top-tier journals tracked by the Nature Index. Its proximity to clinical environments through the CHU de Montpellier creates a direct pipeline from molecular discovery to medical application.

IGF - Institut de Génomique Fonctionnelle

The Institute of Functional Genomics specializes in the molecular and cellular mechanisms underlying brain function and neurological disease one of the domains where complex genetic interactions, of the kind Bionomeex's GWAS-2D is designed to map, are most scientifically consequential.

IPSiM - Institut des Sciences des Plantes de Montpellier

IPSiM is a joint research unit of CNRS, INRAE, Institut Agro, and the University of Montpellier focused on plant water and mineral nutrition and their responses to environmental stress. It is the institutional home of Gabriel Krouk co-founder and CSO of Bionomeex and the scientific environment from which Bionomeex's plant genomics work directly emerged.

This concentration of genomics institutions molecular genetics, functional genomics, sequencing infrastructure, plant science, and human genetics in a single mid-sized city is genuinely unusual. It reflects decades of deliberate scientific investment and creates an environment where collaboration across disciplines happens by proximity as much as by design.

The Central Problem in Genomics - And Why AI Is Now Essential

Generating genomic data has become routine. The hard problem is interpretation specifically, understanding how genetic variants interact to produce observable traits, diseases, and responses to the environment.

During the past decade, GWAS have allowed the discovery of many genetic variants associated with human, plant, and animal phenotypic traits, leading to disruptive insights in biology and impacting both basic knowledge and translational approaches to agronomy and medicine. However, "mono-dimensional GWAS" the study of genetic variation taken one variant at a time is somehow limited. The missing heritability, defined as the unexplained variance of a trait, is probably at least in part attributable to interactions among variants, known as epistasis.

The mystery of missing heritability denotes the gap between the expected heritability of many common diseases as estimated by family and twin studies, and the overall additive heritability obtained by accumulating the effects of all SNPs found significantly associated with these conditions in GWAS. Many diseases that comprise a considerable portion of the healthcare burden display such a gap including type 1 and type 2 diabetes, bipolar disorder, schizophrenia, Alzheimer's disease, multiple sclerosis, and Parkinson's disease.

Solving this problem requires analyzing not individual variants, but the interactions between them a computational challenge so large that it was effectively intractable until AI made it feasible.

Bionomeex's Answer - GWAS-2D, Published in Genome Biology

In their 2024 paper published in Genome Biology, Clément Carré, Gabriel Krouk, André Mas and collaborators from BionomeeX, IMAG, IPSiM, and the University of Bordeaux applied machine learning approaches to GWAS analysis, achieving an acceleration that makes it possible to provide, for the first time, full epistatic interaction maps with over 60 billion pairwise SNP interactions.

This is what makes GWAS-2D technically distinct from every existing approach. It does not sample or approximate the epistatic interaction space. It maps it completely identifying the genetic combinations that drive trait expression, disease risk, and phenotypic prediction at a resolution no prior method could achieve, delivered in hours rather than the years that brute-force computation would require.

The paper was peer-reviewed and published in Genome Biology one of the highest-impact journals in biological science in March 2024. The Luciol visualization software, developed by Bionomeex to make these interaction maps navigable and interpretable, was co-funded by BPI France through an Émergence grant.

→ Read more: A Breakthrough in GeneticsWhat is GWAS?

Where GWAS-2D Is Being Applied - Three Scientific Domains

Human Disease Genetics In collaboration with CHU de Bordeaux and CHU Sainte-Justine in Montreal, Bionomeex is applying GWAS-2D to complex human diseases conditions where the genetic architecture is driven by epistatic interactions that standard GWAS systematically misses. The goal is to identify genetic combinations linked to disease predisposition, progression, and potential therapeutic targets in conditions where individual variant analysis has reached its ceiling.

2D-GWAS in Complex Disease Research - CHU Bordeaux

Plant Science and Crop Genomics In collaboration with HZPC (one of the world's leading potato breeding companies), Michigan State University, and INRAE-affiliated laboratories in Montpellier, Bionomeex applies GWAS-2D to plant traits drought resistance, nutrient efficiency, yield potential where the same epistatic complexity that limits human disease genetics also limits conventional plant breeding. Full interaction maps narrow the genetic search space dramatically, accelerating the identification of breeding targets.

2D-GWAS in Potato Breeding - HZPC

Forest Ecology and Climate Adaptation In collaboration with CEFE (Centre d'Écologie Fonctionnelle et Évolutive) in Montpellier, Bionomeex applies GWAS-2D to beech tree populations analyzing the genetic interactions that determine which tree populations are most resilient to climate-driven environmental stress. With datasets of up to 300 individuals, 450,000 genetic variants, and approximately one terabyte of interaction data per phenotype, this represents one of the most computationally ambitious ecological genomics projects currently running in France.

2D-GWAS in Beech Trees - CEFE Collaboration

Why Montpellier Is the Right Place to Build This

The institutions described above are not just the backdrop to Bionomeex's story they are its origin and its ongoing scientific community. Gabriel Krouk built his career at IPSiM. André Mas works at IMAG. The founding technology was licensed through SATT AxLR. The sequencing infrastructure at MGX is available to researchers across the ecosystem.

For a company developing AI for genomics, being embedded in this institutional fabric provides something that no external company can replicate: direct access to the domain experts, the data infrastructure, the computational resources, and the scientific culture that makes rigorous genomics AI possible.

Montpellier is not a city where genomics AI is being trialed. It is a city where genomics science has been built over decades and Bionomeex is the company translating that depth into deployable AI technology.

Frequently Asked Questions

What genomics institutions are based in Montpellier?

Montpellier hosts IGMM (molecular genetics), IGF (functional genomics), MGX (CNRS sequencing platform), IGH (human genetics), IPSiM (plant sciences), and CEFE (functional ecology) making it one of Europe's most concentrated genomics research cities outside Paris.

What is the difference between standard GWAS and Bionomeex's GWAS-2D?

Standard GWAS analyzes genetic variants one at a time, looking for individual associations with a trait. GWAS-2D maps pairwise interactions between all variants over 80 billion combinations identifying the epistatic relationships that individual variant analysis cannot detect. This recovers a significant portion of the missing heritability that has limited genetic research for over a decade.

Has Bionomeex's genomics technology been independently validated?

Yes. The GWAS-2D methodology was peer-reviewed and published in Genome Biology in March 2024 one of the most rigorous review processes in biological science. The technology has also been applied in active collaborations with two university hospitals and multiple research institutions.

Can GWAS-2D be applied to any species?

The technology is species-agnostic. It has been applied to model plants, potato, beech trees, and human disease datasets. The primary requirements are a sufficiently large population with both genotype data (SNPs) and measured phenotypic traits.

How does Bionomeex relate to the University of Montpellier's genomics ecosystem?

Bionomeex is an official spin-off of CNRS and the University of Montpellier, with co-founders from IPSiM and IMAG. It maintains active research collaborations with multiple University of Montpellier-affiliated laboratories and participates in events like the Montpellier Omics Days.


→ Contact Bionomeex