The Network Analysis and Modelling group is devoted to investigate molecular mechanisms of phenotype emergence in crops by means of large-scale data integration on the genomic, transcriptomic, metabolomic, and phenomic level. In practical terms, we develop and implement machine learning approaches and network analysis algorithms that help us (or others) to discover new gene functions, or new interactions between genes, metabolites and phenotypes.

Some of our specific activities include e.g. implementations of Deep Learning for crops phenotype prediction and breeding, motif discovery in large multi-omic genome-to-phenome networks, and “gamification” of plant development. Our favourite crops include cereals but we have special affinity to solanaceous plants too. Furthermore, we provide statistical expertise, machine learning solutions, and data visualisation tools for interpretation of high-throughput data in IPK.

AI for Gene Regulation

We are working on a toolbox of deep learning algorithms that can elucidate predictive relationships between sequence of non-coding gene regulatory elements, protein-DNA binding events, and expression patterns of their target genes. We train our algorithms on large resources of genetic variation, DNA binding and gene expression data across multiple species, tissues and treatments. Our models achieve high accuracy and represent a novel and promising approach to reconstruct mechanism-based gene regulatory networks. We use our toolbox to estimate the regulatory impact of structural genetic variation, highlight gene candidates in GWAS studies and design gene editing strategies for gene expression modulation.

Systems Genetics

Our other focus is are methods for integration of multi-omic data and identification of causal/predictive interactions between genetic variation, gene expression, metabolite levels and emergence of crop quality traits. In GWAS experiments, we integrate multi-omic data to identify the most likely gene-mechanism-phenotype paths. In time series experiments, we characterize the molecular events determing developmental and stress responses events in time. Here, we collaborate with multiple labs interested in elucidation of molecular mechanisms of genetic associations or identification of molecular traits for targeted breeding.

Gamification of Plant Life

In our lab, we explore the life of a plant as a captivating game of survival. We construct intricate molecular networks that serve as the game engines, driving the complex interactions that govern plant life and responses to environmental challenges. These networks become dynamic, interactive models that mirror the plants’ strategies for growth, defense, and adaptation. Harnessing the power of educational computer games, we invite students and researchers to step into this vibrant world. Through engaging, game-like simulations, participants can directly interact with our scientific models, gaining an immersive and intuitive understanding of plant biology. Our goal is to bridge science and education, providing a novel way to explore, learn, and contribute to our understanding of plant systems biology.