Understanding the genetic regulatory code governing gene expression is an important challenge in molecular biology. However, how individual coding and non-coding regions of the gene regulatory structure interact and contribute to mRNA expression levels remains unclear. Here we apply deep learning on over 20,000 mRNA datasets to examine the genetic regulatory code controlling mRNA abundance in 7 model organisms ranging from bacteria to Human. In all organisms, we can predict mRNA abundance directly from DNA sequence, with up to 82% of the variation of transcript levels encoded in the gene regulatory structure. By searching for DNA regulatory motifs across the gene regulatory structure, we discover that motif interactions could explain the whole dynamic range of mRNA levels. Co-evolution across coding and non-coding regions suggests that it is not single motifs or regions, but the entire gene regulatory structure and specific combination of regulatory elements that define gene expression levels.

Utilization of synthetic biology approaches to enhance current plastic degradation processes is of crucial importance, as natural plastic degradation processes are very slow. For instance, the predicted lifetime of a polyethylene terephthalate (PET) bottle under ambient conditions ranges from 16 to 48 years.

Design of de novo synthetic regulatory DNA is a promising avenue to control gene expression in biotechnology and medicine. Using mutagenesis typically requires screening sizable random DNA libraries, which limits the designs to span merely a short section of the promoter and restricts their control of gene expression. Here, we prototype a deep learning strategy based on generative adversarial networks (GAN) by learning directly from genomic and transcriptomic data. Our ExpressionGAN can traverse the entire regulatory sequence-expression landscape in a gene-specific manner, generating regulatory DNA with prespecified target mRNA levels spanning the whole gene regulatory structure including coding and adjacent non-coding regions. Despite high sequence divergence from natural DNA, in vivo measurements show that 57% of the highly-expressed synthetic sequences surpass the expression levels of highly-expressed natural controls. This demonstrates the applicability and relevance of deep generative design to expand our knowledge and control of gene expression regulation in any desired organism, condition or tissue.

Potato (Solanum tuberosum) is highly water and space efficient but susceptible to abiotic stresses such as heat, drought, and flooding, which are severely exacerbated by climate change. Our understanding of crop acclimation to abiotic stress, however, remains limited. Here, we present a comprehensive molecular and physiological high-throughput profiling of potato (Solanum tuberosum, cv. Désirée) under heat, drought, and waterlogging applied as single stresses or in combinations designed to mimic realistic future scenarios. Stress responses were monitored via daily phenotyping and multi-omics analyses of leaf samples comprising proteomics, targeted transcriptomics, metabolomics, and hormonomics at several timepoints during and after stress treatments. Additionally, critical metabolites of tuber samples were analyzed at the end of the stress period. We performed integrative multi-omics data analysis using a bioinformatic pipeline that we established based on machine learning and knowledge networks. Waterlogging produced the most immediate and dramatic effects on potato plants, interestingly activating ABA responses similar to drought stress. In addition, we observed distinct stress signatures at multiple molecular levels in response to heat or drought and to a combination of both. In response to all treatments, we found a downregulation of photosynthesis at different molecular levels, an accumulation of minor amino acids, and diverse stress-induced hormones. Our integrative multi-omics analysis provides global insights into plant stress responses, facilitating improved breeding strategies toward climate-adapted potato varieties.

Understanding the molecular mechanisms behind plant response to stress can enhance breeding strategies and help us design crop varieties with improved stress tolerance, yield, and quality. To investigate resource redistribution from growth- to defense-related processes in an essential tuber crop, potato, here we generate a large-scale compartmentalized genome-scale metabolic model (GEM), potato-GEM. Apart from a large-scale reconstruction of primary metabolism, the model includes the full known potato secondary metabolism, spanning over 566 reactions that facilitate the biosynthesis of 182 distinct potato secondary metabolites. Constraint-based modeling identifies that the activation of the largest amount of secondary (defense) pathways occurs at a decrease of the relative growth rate of potato leaf, due to the costs incurred by defense. We then obtain transcriptomics data from experiments exposing potato leaves to two biotic stress scenarios, a herbivore and a viral pathogen, and apply them as constraints to produce condition-specific models. We show that these models recapitulate experimentally observed decreases in relative growth rates under treatment as well as changes in metabolite levels between treatments, enabling us to pinpoint the metabolic rewiring underlying growth–defense trade-offs. Potato-GEM thus presents a useful resource to study and broaden our understanding of potato and general plant defense responses under stress conditions.