Computational Biology

TBA

The construction of gene regulatory networks based on machine learning algorithms is a powerful method to generate hypotheses about functional elements, be it DNA sequences in cis or transcription factors, histone mark readers, and mediators in trans. A large array of possible methods, random forest decision trees, (convolutional) neural networks, variational autoencoders among others are available. In this project, we explore different approaches to predicting plant gene expression with variation in tissue and in stress responses using RNA-seq, ATAC-seq, single cell RNA-seq, and single cell ATAC-seq data as input and output. We connect our results to genome variation data to predict responses of varieties and test the predictions against reality.

Specialized metabolism (also known as secondary metabolism) produces metabolites which are specific to plant species or groups of plant species. Photosynthesis is part of the primary metabolism, is present in nearly all plant species and is highly conserved with regard to its genes and its regulation. In contrast, specialized metabolism is highly variable in all points, providing insights into recent evolutionary events. Our target organisms are tansy (Tancetum vulgare) and the bittersweet nightshade (Solanum dulcamara).

For both organisms, we generated de novo genome sequence assemblies. We use RNA-seq data to produce hypotheses about transcription factors controlling terpenoid biosynthesis in tansy and steroidal glycoalkaloids in the bittersweet nightshade. DAP-seq is used to test whether the candidate TFs indeed bind the respective target genes in the genomes and comparative evolutionary analyses are used to test the evolutionary origins of the regulon.

Photosynthesis in plants is the basis to all agricultural productivity, either directly or indirectly. While its post-transcriptional regulation is well studied, our knowledge about its transcriptional regulation remains limited. Improving this knowledge is crucial, since photosynthesis has become an active target for traditional and modern breeding methods in recent years. The climate change marked by extreme weather conditions, such as excessive drought, heat, or excess water periods, massively affects photosynthesis and, consequently, crop yield.

We use gene regulatory networks generated by machine learning to predict, which transcription factors (TFs) control photosynthetic target genes in the model organism Arabidopsis thaliana. Currently, we study candidate TFs through knock-out and overexpression lines using plant physiology methods, RNA-seq, and protein biochemistry.

Further, we have used gene regulatory networks generated by machine learning to predict TFs controlling photosynthetic genes in the liverwort, Marchantia polymorpha. Candidate TFs and closely related protein family members were expressed and functionally induced. The resulting RNA-seq data determined the candidate’s involvement in photosynthetic regulation.

As part of this work, we sequenced the genome of M. polymorpha var. BoGa (Beaulieu et al., 2025; genome available under Laker et al., 2024) and studied the evolutionary conservation of TF binding motifs (Zenker et al., 2024).

Additionally, we are developing a novel promoter bashing tool using synthetic biology. In vivo recombination is used to test the importance of binding sites and their positions. A laboratory project is available for an ambitious master student.

Plant seed development and germination are complex processes involving the programmed shut-down and re-establishment of photosynthesis. It requires the formation of storage compounds and their use upon germination. Precise control and timing are critical for seed development and germination. Both processes are of agricultural interest, since seed filling affects yield and germination ability and timing impacts field performance. Seeds must remain dormant when needed and germinate at nearly 100% efficiency when in the field.

We have generated gene regulatory networks of seed filling identifying known transcription factors (TFs) involved in seed filling and novel candidates. We study their expression patterns, target genes, affinity, and genetic variations in binding sites. Polyploid species like Brassica napus are particularly interesting as they offer the opportunity to observe patterns in two, recently combined genomes at the same time.

Crassulacean acid metabolism (CAM) photosynthesis is a key adaptation that enables plants to survive seasonal or prolonged water limitation. CAM operates by temporally uncoupling CO2 uptake from CO2 assimilation into sugars: CO2 is taken up during times of limited transpiration and therefore limited water loss and either assimilated at dusk or dawn or stored as malate in the vacuole at night. During the day, CO2 is released and assimilated using the solar energy. Introducing CAM ability into crop plants could improve their drought tolerance.

In many plant species, CAM photosynthesis is induced or reinforced by drought. We are currently testing which drought inducible transcription factors bind to genes controlling the CAM trait and their binding affinities. Additionally, we study evolutionary conservation of drought transcription factor binding sites.

Cyanobacteria invented oxygenic photosynthesis more than 3 billion years ago and thus enabled complex life on earth. Furthermore, they are the evolutionary ancestors of plant chloroplasts. Studying photosynthesis in cyanobacteria is to the best advantage, since the prokaryotes are easily amenable to genetic engineering. Besides studying manganese management, a basic requirement for efficient oxygenic photosynthesis, we are also interested in transcriptional regulation of photosynthetic genes.

Gene regulatory networks have identified candidates for photosynthetic regulation. To test these predictions, we use DNA affinity purification sequencing (DAP-seq), mutant and overexpression lines. Initial experiments with single transcription factors (TFs) confirm multiple binding sites per TF. Ultimately, we plan to analyze all TFs amenable to DAP-seq.

In nature, bacteria typically live in communities. We study the metabolic and regulatory interactions within microbial consortia that include a photosynthetic primary producer like cyanobacteria. To explore these relationships, we sequence the metagenomes of such consortia. Selected consortia are cultured under defined conditions and genome-sequenced to characterize their members. Based on genome data, we generate hypotheses about the nature of the interactions, mutual, parasitic, or symbiotic, and test the hypotheses using synthetic ecology.

We expect the primary producers to defend themselves or to selectively foster beneficial collaborations for defense through specialized metabolites, nutrient sequestration, adjustments to CO2 assimilation, or nutrient exchange. For selected strains, we characterize their contributions in detail including gene regulation circuits.

This project is integrated in our teaching. Researching project and parts are addressed within the course “Basics in Molecular Microbiology” (202113).