Computational design of RNAs with complex energy landscapes

Siederdissen, Christian Höner zu; Hammer, Stefan; Abfalter, Ingrid G.; Hofacker, Ivo L.; Stadler, Peter F.

doi:10.1002/bip.22337

Cited by 30 publications

(41 citation statements)

References 71 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Given the ability to quantitatively predict RNA-reporter affinities, optimization algorithms could be applied to design RNAs that sense and interact with other biomolecules in their environments, as demonstrated by our design tests in simulation. [45][46][47] This level of predictive performance would enable us to precisely design interactions in biological systems, including those central to modern biotechnological tools.…”

Section: Discussionmentioning

confidence: 99%

Prospects for recurrent neural network models to learn RNA biophysics from high-throughput data

Andreasson

Kladwang

et al. 2017

Preprint

View full text Add to dashboard Cite

RNA is a functionally versatile molecule that plays key roles in genetic regulation and in emerging technologies to control biological processes. Computational models of RNA secondary structure are well-developed but often fall short in making quantitative predictions of the behavior of multi-RNA complexes. Recently, large datasets characterizing hundreds of thousands of individual RNA complexes have emerged as rich sources of information about RNA energetics. Meanwhile, advances in machine learning have enabled the training of complex neural networks from large datasets. Here, we assess whether a recurrent neural network model, Ribonet, can learn from high-throughput binding data, using simulation and experimental studies to test model accuracy but also determine if they learned meaningful information about the biophysics of RNA folding. We began by evaluating the model on energetic values predicted by the Turner model to assess whether the neural network could learn a representation that recovered known biophysical principles. First, we trained Ribonet to predict the simulated free energy of an RNA in complex with multiple input RNAs. Our model accurately predicts free energies of new sequences but also shows evidence of having learned base pairing information, as assessed by in silico double mutant analysis. Next, we extended this model to predict the simulated affinity between an arbitrary RNA sequence and a reporter RNA. While these more indirect measurements precluded the learning of basic principles of RNA biophysics, the resulting model achieved sub-kcal/mol accuracy and enabled design of simple RNA input responsive riboswitches with high activation ratios predicted by the Turner model from which the training data were generated. Finally, we compiled and trained on an experimental dataset comprising over 600,000 experimental affinity measurements published on the Eterna open laboratory. Though our tests revealed that the model likely did not learn a physically realistic representation of RNA interactions, it nevertheless achieved good . CC-BY-NC-ND 4.0 International license peer-reviewed) is the author/funder. It is made available under aThe copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/227611 doi: bioRxiv preprint first posted online Dec. 1, 2017; 2 performance of 0.76 kcal/mol on test sets with the application of transfer learning and novel sequence-specific data augmentation strategies. These results suggest that recurrent neural network architectures, despite being naïve to the physics of RNA folding, have the potential to capture complex biophysical information. However, more diverse datasets, ideally involving more direct free energy measurements, may be necessary to train de novo predictive models that are consistent with the fundamentals of RNA biophysics. Author SummaryThe precise design of RNA interactions is essential to gaining greater control over RNA-based biotechnology tools, including designer riboswitches and CRISPR-Cas9 gene editing. However, the ...

show abstract

Section: Discussionmentioning

confidence: 99%

Prospects for recurrent neural network models to learn RNA biophysics from high-throughput data

Andreasson

Kladwang

et al. 2017

Preprint

View full text Add to dashboard Cite

show abstract

“…The possibility to encode effective sensors at RNA level makes riboswitches valuable gadgets that can directly interfere with the complex process of gene expression without the need of additional co-factors such as proteins. Here we elucidate the complex process of designing such a 15 ligand-sensing riboswitch that, for simplicity, does not implement a specific regulation mechanism at transcriptional or translational level. The RNA sequence should "simply" adapt two alternative conformations depending on the presence or absence of a ligand.…”

Section: Introductionmentioning

confidence: 99%

“…characterizing the mechanism of the artificial device. Only recently, some published design programs started to allow to compile an objective function from a catalog of predefined functions [15,13]. RNAblueprint [10] went one step further and allowed to formulate the objective utilizing a scripting interface, which gives the user complete control over the optimization procedure.…”

Section: Introductionmentioning

confidence: 99%

In silico design of ligand triggered RNA switches

Findeiß

Hammer

Wolfinger

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

This contribution sketches a work flow to design an RNA switch that is able to adapt two structural conformations in a ligand dependent way. A well characterized RNA aptamer, i. e., knowing its K d and adaptive structural features, is an essential ingredient of the described design process. We exemplify the principles using the well known theophylline aptamer throughout this work. The aptamer in its ligand-binding competent structure represents one structural conformation of the switch while an alternative fold that disrupts the binding competent structure forms the other conformation. To keep it simple we do not incorporate any regulatory mechanism to control transcription or translation. We elucidate a commonly used design process by explicitly dissecting and explaining the necessary steps in detail. We developed a novel objective function which methodically describes the function of this simple, ligand-triggered riboswitch and describe an extensive in silico analysis pipeline to evaluate important kinetic properties of the designed sequences. This protocol and the developed software can be easily extended or adapted to fit novel design scenarios and thus can serve as a template for future needs.

show abstract

“…This functionality is useful because wild-type sequences within living organisms often present medium or low GC-content, presumably to offer better transcription rates and/or structural plasticity. RNAdesign is another tool for designing RNA sequences that fold into multiple target structures [ 21 ]. It uses the graph coloring techniques and heuristic local optimization algorithm to find sequences whose energy landscapes are dominated by the prescribed conformations.…”

Section: Introductionmentioning

confidence: 99%

ERD: a fast and reliable tool for RNA design including constraints

Esmaili-Taheri

Ganjtabesh

2015

BMC Bioinformatics

View full text Add to dashboard Cite

BackgroundThe function of an RNA in cellular processes is directly related to its structure. The free energy of RNA structure in another important key to its function as only some structures with a specific level of free energy can take part in cellular reactions. Therefore, to perform a specific function, a particular RNA structure with specific level of free energy is required. For a given RNA structure, the goal of the RNA design problem is to design an RNA sequence that folds into the given structure. To mimic the biological features of RNA sequences and structures, some sequence and energy constraints should be considered in designing RNA. Although the level of free energy is important, it is not considered in the available approaches for RNA design problem.ResultsIn this paper, we present a new version of our evolutionary algorithm for RNA design problem, entitled ERD, and extend it to handle some sequence and energy constraints. In the sequence constraints, one can restrict sequence positions to a fixed nucleotide or to a subset of nucleotides. As for the energy constraint, one can specify an interval for the free energy ranges of the designed sequences. We compare our algorithm with INFO-RNA, MODENA, NUPACK, and RNAiFold approaches for some artificial and natural RNA secondary structures and constraints.ConclusionsThe results indicate that our algorithm outperforms the other mentioned approaches in terms of accuracy, speedup, divergency, nucleotides distribution, and similarity to the natural RNA sequences. Particularly, the designed RNA sequences in our method are much more reliable and similar to the natural counterparts. The generated sequences are more diverse and they have closer nucleotides distribution to the natural one. The ERD tool and web server are freely available at http://mostafa.ut.ac.ir/corna/erd-cons/.

show abstract

Computational design of RNAs with complex energy landscapes

Cited by 30 publications

References 71 publications

Prospects for recurrent neural network models to learn RNA biophysics from high-throughput data

Prospects for recurrent neural network models to learn RNA biophysics from high-throughput data

In silico design of ligand triggered RNA switches

ERD: a fast and reliable tool for RNA design including constraints

Contact Info

Product

Resources

About