Objective: The scientific objective of this Dagstuhl seminar is to bring together computational scientists and experimental biologists who are experts in the problem of protein-RNA recognition to identify, discuss and address major computational challenges in this field.

Challenges: Two broad challenges of fundamentally computational nature related to protein-RNA interactions will provide the foundation for the seminar. The first challenge is the mining of extensive genome-wide information on regulatory interactions on RNA and the associated proteins. The second is the extension of atomic level models from narrow paradigmatic systems to broad genome-wide quantitative prediction of cellular networks. There is an urgent need for developing innovative computational approaches to model RNA structure, RNA-protein complexes, predict their interactions at multiple levels and integrate this information into genome-wide modelling of gene expression networks. A significant challenge is the tight interaction between experimentalists and computationalists, to make use of achievable experimental results for the questions most relevant to biological researchers.

Discussion topics: Specific items to be discussed are listed next under five categories from which scientific sessions will be generated.

1. Identifying new RNA-binding proteins – Great advances have recently been made in the development of high-throughput screens to identify novel RNA Binding Proteins (RBPs) in cultured cell lines and in tissues. While the technology has increased the number of identified RBPs and suggested novel cellular mechanisms for these proteins, these high throughput technologies tend to generate many false predictions. A great challenge in validating novel RBPs is the fact that the classical computational approaches for predicting protein functions are not applicable as the latter rely on sequence and structural homology to known proteins in the databases. In the seminar, we shall discuss ways to overcome these challenges, such as developing and employing non-homology based modeling and deep learning approaches for protein function predictions.

2. Modeling RBP interaction networks – The development of genome-wide approaches to investigate the RNA population targeted by an RBP and to establish its specificity has been a tremendous recent advance. The combinatorial complexity of RNA-based regulation is stunning and daunting. Computational challenges range from hierarchical mixture models for individual interactions to integrated models that make use of supervised and unsupervised machine learning methods. The seminar will discuss ways by which the experimental results on RBP specificity should be fed into computational models of gene expression networks.

3. Modelling the impact of RBPs on gene expression – The development of high-throughput sequencing protocols has dramatically advanced our knowledge on RNA regulation in steady state. However, reconstructing the transcriptome at the different stages of the gene expression cascade and how regulation by RBPs affect gene expression networks is still in early phases. New protocols typically require dedicated algorithms. We will discuss developments in analysis and reconstruction of the status of the transcriptome in its different subcellular and molecular states, and the development of algorithms to link changes in RNA sequence or structure to functional molecular outcomes.

4. Development and extension of predictive atomic models of protein-RNA interactions – The structural and biochemical analysis of protein-RNA complexes remains a mainstay of the field and provides the most detailed source of information on the physical chemical basis of gene regulation. However, these are necessarily limited by the slow pace of the experimental methods. In the seminar we will discuss challenges and new computational approaches to bridge the gap between the extensive information derived from high throughput experimental technology and the very detailed atomic structural information.

5. Design of RNA-binding proteins and RNA scaffolds – Newly engineered proteins would allow the control of cellular RNA regulatory network. Furthermore, the design of RNA scaffolds for bringing together different RBPs into close proximity would have applications in drug discovery and synthetic biology. This is a significant computational challenge, believed to be NP-complete. In the seminar, we shall discuss algorithms for the design of new RBPs and RNA scaffolds.

Creative Commons BY 3.0 DE

Rolf Backofen, Yael Mandel-Gutfreund, Uwe Ohler, and Gabriele Varani

• Frédéric Allain (ETH Zürich, CH)
• Rolf Backofen (Universität Freiburg, DE) [dblp]
• Benedikt Beckman (HU Berlin, DE)
• Janusz Bujnicki (Int. Inst. of Molecular & Cell Biology - Warsaw, PL) [dblp]
• Tomaž Curk (University of Ljubljana, SI) [dblp]
• Philipp Drewe (Max-Delbrück-Centrum - Berlin, DE) [dblp]
• André Gerber (University of Surrey - Guildford, GB)
• Mahsa Ghanbari (Max-Delbrück-Centrum - Berlin, DE) [dblp]
• Tim Hughes (University of Toronto, CA) [dblp]
• Eckhard Jankowsky (Case Western Reserve University - Cleveland, US) [dblp]
• Hilal Kazan (Antalya International University, TR) [dblp]
• Grzegorz Kudla (University of Edinburgh, GB) [dblp]
• Markus Landthaler (Max-Delbrück-Centrum - Berlin, DE) [dblp]
• Martin Lewinski (Universität Bielefeld, DE) [dblp]
• Yael Mandel-Gutfreund (Technion - Haifa, IL) [dblp]
• Hannah Margalit (The Hebrew University of Jerusalem, IL) [dblp]
• Annalisa Marsico (MPI für Molekulare Genetik - Berlin, DE) [dblp]
• Daniel Maticzka (Universität Freiburg, DE)
• Irmtraud Meyer (Max-Delbrück-Centrum - Berlin, DE)
• Quaid Morris (University of Toronto, CA) [dblp]
• Uwe Ohler (Max-Delbrück-Centrum - Berlin, DE) [dblp]
• Teresa Przytycka (National Center for Biotechnology - Bethesda, US) [dblp]
• Andres Ramos (University College London, GB) [dblp]
• Guido Sanguinetti (University of Edinburgh, GB) [dblp]
• Michael Sattler (TU München, DE) [dblp]
• Peter F. Stadler (Universität Leipzig, DE) [dblp]
• Gabriele Varani (University of Washington - Seattle, US) [dblp]
• Gene Yeo (UC - San Diego, US) [dblp]