Toward Increasing the Credibility of RNA Design.

In the problem of RNA design, also known as inverse folding, RNA sequences are predicted that achieve the desired secondary structure at the lowest possible free energy and under certain constraints. The designed sequences have applications in synthetic biology and RNA-based nanotechnologies. There are also known cases of the successful use of inverse folding to discover previously unknown noncoding RNAs. Several computational methods have been dedicated to the problem of RNA design. They differ by algorithm and additional parameters, e.g., those determining the goal function in the sequence optimization process. Users can obtain many promising RNA sequences quite easily. The more difficult issue is to critically evaluate them and select the most favorable and reliable sequence that form1s the expected RNA structure. The latter problem is addressed in this paper. We propose an RNA design protocol extended to include sequence evaluation, for which a 3D structure is used. Experiments show that the accuracy of RNA design can be improved by adding a 3D structure prediction and analysis step.

gRNAde: A Geometric Deep Learning Pipeline for 3D RNA Inverse Design.

Fundamental to the diverse biological functions of RNA are its 3D structure and conformational flexibility, which enable single sequences to adopt a variety of distinct 3D states. Currently, computational RNA design tasks are often posed as inverse problems, where sequences are designed based on adopting a single desired secondary structure without considering 3D geometry and conformational diversity. In this tutorial, we present gRNAde, a geometric RNA design pipeline operating on sets of 3D RNA backbone structures to design sequences that explicitly account for RNA 3D structure and dynamics. gRNAde is a graph neural network that uses an SE (3) equivariant encoder-decoder framework for generating RNA sequences conditioned on backbone structures where the identities of the bases are unknown. We demonstrate the utility of gRNAde for fixed-backbone re-design of existing RNA structures of interest from the PDB, including riboswitches, aptamers, and ribozymes. gRNAde is more accurate in terms of native sequence recovery while being significantly faster compared to existing physics-based tools for 3D RNA inverse design, such as Rosetta.

Generative Modeling of RNA Sequence Families with Restricted Boltzmann Machines.

In this chapter, we discuss the potential application of Restricted Boltzmann machines (RBM) to model sequence families of structured RNA molecules. RBMs are a simple two-layer machine learning model able to capture intricate sequence dependencies induced by secondary and tertiary structure, as well as mechanisms of structural flexibility, resulting in a model that can be successfully used for the design of allosteric RNA such as riboswitches. They have recently been experimentally validated as generative models for the SAM-I riboswitch aptamer domain sequence family. We introduce RBM mathematically and practically, providing self-contained code examples to download the necessary training sequence data, train the RBM, and sample novel sequences. We present in detail the implementation of algorithms necessary to use RBMs, focusing on applications in biological sequence modeling.

Sequence Design for RNA-RNA Interactions.

The design of RNA sequences with desired structural properties presents a challenging computational problem with promising applications in biotechnology and biomedicine. Most regulatory RNAs function by forming RNA-RNA interactions, e.g., in order to regulate mRNA expression. It is therefore natural to consider problems where a sequence is designed to form a desired RNA-RNA interaction and switch between structures upon binding. This contribution demonstrates the use of the Infrared framework to design interacting sequences. Specifically, we consider the regulation of the rpoS mRNA by the sRNA DsrA and design artificial 5 ' UTRs that place a downstream protein coding gene under control of DsrA. The design process is explained step by step in a Jupyter notebook, accompanied by Python code. The text discusses setting up design constraints for sampling sequences in Infrared, computing quality measures, constructing a suitable cost function, as well as the optimization procedure. We show that not only thermodynamic but also kinetic folding features can be relevant. Kinetics of interaction formation can be estimated efficiently using the RRIkinDP tool, and the chapter explains how to include kinetic folding features from RRIkinDP directly in the cost function. The protocol implemented in our Jupyter notebook can easily be extended to consider additional requirements or adapted to novel design scenarios.

Riboswitch Design Using MODENA.

In silico design of artificial riboswitches is a challenging and intriguing task. Since experimental approaches such as in vitro selection are time-consuming processes, computational tools that guide riboswitch design are desirable to accelerate the design process. In this chapter, we describe the usage of the MODENA web server to design ON riboswitches on the basis of a multi-objective genetic algorithm and RNA secondary structure prediction.

Sequence Design Using RNAstructure.

RNA is present in all domains of life. It was once thought to be solely involved in protein expression, but recent advances have revealed its crucial role in catalysis and gene regulation through noncoding RNA. With a growing interest in exploring RNAs with specific structures, there is an increasing focus on designing RNA structures for in vivo and in vitro experimentation and for therapeutics. The development of RNA secondary structure prediction methods has also spurred the growth of RNA design software. However, there are challenges to designing RNA sequences that meet secondary structure requirements. One major challenge is that the secondary structure design problem is likely NP-hard, making it computationally intensive. Another issue is that objective functions need to consider the folding ensemble of RNA molecules to avoid off target structures. In this chapter, we provide protocols for two software tools from the RNAstructure package: "Design" for structured RNA sequence design and "orega" for unstructured RNA sequence design.

Machine Learning for RNA Design: LEARNA.

Machine learning algorithms, and in particular deep learning approaches, have recently garnered attention in the field of molecular biology due to remarkable results. In this chapter, we describe machine learning approaches specifically developed for the design of RNAs, with a focus on the learna_tools Python package, a collection of automated deep reinforcement learning algorithms for secondary structure-based RNA design. We explain the basic concepts of reinforcement learning and its extension, automated reinforcement learning, and outline how these concepts can be successfully applied to the design of RNAs. The chapter is structured to guide through the usage of the different programs with explicit examples, highlighting particular applications of the individual tools.

RNA Design Using incaRNAfbinv Demonstrated with the Identification of Functional RNA Motifs in Hepatitis Delta Virus.

Computational RNA design was introduced in the 1990s by Vienna's RNAinverse, which is a simple inverse RNA folding solver. Further developments and contemporary RNA design techniques, in addition to improved efficiency, offer more precise control over the designed sequences. incaRNAfbinv (incaRNAtion with RNA fragment-based inverse) is one such extension that builds upon RNAinverse and includes coarse-graining manipulations. The idea is that an RNA secondary structure can be decomposed to fragments of RNA motifs, and that a significant number of known natural RNA motifs exhibit a remarkable preservation in particular locations in a variety of genomes. This is taken into consideration by the ability of the user to select motifs that are known to be functional for a precise design, whilst the algorithm is more adaptable on other motifs. The latest version, incaRNAfbinv 2.0, is a free-to-use web-server which deploys the above methodology of fragment-based design. Its control over the decomposed RNA secondary structure motifs includes, among other advanced features, the insertion of constraints in a flexible manner. The resultant RNA designed sequences are ranked by their proximity to classical RNA design. Features and capabilities of incaRNAfbinv 2.0 are hereby illustrated with an example taken from hepatitis delta virus (HDV). The web-server is demonstrated in assisting to locate a known RNA motif that is responsible for HDV-3 RNA editing in more HDV genotypes than thought of before. This shows that computational RNA design by using inverse RNA folding is also a valuable strategy for locating functional RNA motifs in genomic data, in addition to artificially designing synthetic RNAs.

Simulated Annealing for RNA Design with SIMARD.

Ribonucleic acid (RNA) design is the inverse of RNA folding. RNA folding aims to identify the most likely secondary structure into which a given strand of nucleotides will fold. RNA design algorithms, on the other hand, attempt to design a strand of nucleotides that will fold into a specified secondary structure. Despite the apparent NP-hard nature of RNA design, promising results can be achieved when formulated as a combinatorial optimization problem and approached with simple heuristics. The main focus of this paper is to describe an RNA design algorithm based on simulated annealing. Additionally, noteworthy features and results will be presented herein.

3D-Based RNA Function Prediction Tools in rnaglib.

Understanding the connection between complex structural features of RNA and biological function is a fundamental challenge in evolutionary studies and in RNA design. However, building datasets of RNA 3D structures and making appropriate modeling choices remain time-consuming and lack standardization. In this chapter, we describe the use of rnaglib, to train supervised and unsupervised machine learning-based function prediction models on datasets of RNA 3D structures.

A Computational Approach for Designing Synthetic Riboswitches for Next-Generation RNA Therapeutics.

Riboswitches are naturally occurring regulatory segments of RNA molecules that modulate gene expression in response to specific ligand binding. They serve as a molecular 'switch' that controls the RNA's structure and function, typically influencing the synthesis of proteins. Riboswitches are unique because they directly interact with metabolites without the need for proteins, making them attractive tools in synthetic biology and RNA-based therapeutics. In synthetic biology, riboswitches are harnessed to create biosensors and genetic circuits. Their ability to respond to specific molecular signals allows for the design of precise control mechanisms in genetic engineering. This specificity is particularly useful in therapeutic applications, where riboswitches can be synthetically designed to respond to disease-specific metabolites, thereby enabling targeted drug delivery or gene therapy. Advancements in designing synthetic riboswitches for RNA-based therapeutics hinge on sophisticated computational techniques, which are described in this chapter. The chapter concludes by underscoring the potential of computational strategies in revolutionizing the design and application of synthetic riboswitches, paving the way for advanced RNA-based therapeutic solutions.

Monte Carlo Inverse RNA Folding.

The inverse RNA folding problem deals with designing a sequence of nucleotides that will fold into a desired target structure. Generalized Nested Rollout Policy Adaptation (GNRPA) is a Monte Carlo search algorithm for optimizing a sequence of choices. It learns a playout policy to intensify the search of the state space near the current best sequence. The algorithm uses a prior on the possible actions so as to perform non uniform playouts when learning the instance of problem at hand. We trained a transformer neural network on the inverse RNA folding problem using the Rfam database. This network is used to generate a prior for every Eterna100 puzzle. GNRPA is used with this prior to solve some of the instances of the Eterna100 dataset. The transformer prior gives better result than handcrafted heuristics.

SHAPE Probing to Screen Computationally Designed RNA.

RNA design is a major challenge for the future development of synthetic biology and RNA-based therapy. The development of efficient and accurate RNA design pipelines is based on trial and error strategies. The fast progression of such algorithms requires assaying the properties of many RNA sequences in a short time frame. High throughput RNA structure chemical probing technologies such as SHAPE-MaP allow for assaying RNA structure and interaction rapidly and at a very large scale. However, the promiscuity of the designed sequences that may differ only by one nucleotide requires special care. In addition, it necessitates the analysis and evaluation of many experimental results that may reveal to be very tedious. Here we propose an experimental and analytical workflow that eases the screening of thousands of designed RNA sequences at once. In particular, we have developed shapemap tools a customized software suite available at https://github.com/sargueil-citcom/shapemap-tools .

Datasets for Benchmarking RNA Design Algorithms.

RNA molecules play vital roles in many biological processes, such as gene regulation or protein synthesis. The adoption of a specific secondary and tertiary structure by RNA is essential to perform these diverse functions, making RNA a popular tool in bioengineering therapeutics. The field of RNA design responds to the need to develop novel RNA molecules that possess specific functional attributes. In recent years, computational tools for predicting RNA sequences with desired folding characteristics have improved and expanded. However, there is still a lack of well-defined and standardized datasets to assess these programs. Here, we present a large dataset of internal and multibranched loops extracted from PDB-deposited RNA structures that encompass a wide spectrum of design difficulties. Furthermore, we conducted benchmarking tests of widely utilized open-source RNA design algorithms employing this dataset.

Generating Artificial Ribozymes Using Sparse Coevolutionary Models.

RNA ribozyme (Walter Engelke, Biologist (London, England) 49:199-203, 2002) datasets typically contain from a few hundred to a few thousand naturally occurring sequences. However, the potential sequence space of RNA is huge. For example, the number of possible RNA sequences of length 150 nucleotides is approximately 1 0 90 , a figure that far surpasses the estimated number of atoms in the known universe, which is around 1 0 80 . This disparity highlights a vast realm of sequence variability that remains unexplored by natural evolution. In this context, generative models emerge as a powerful tool. Learning from existing natural instances, these models can create artificial variants that extend beyond the currently known sequences. In this chapter, we will go through the use of a generative model based on direct coupling analysis (DCA) (Russ et al., Science 369:440-445, 2020; Trinquier et al., Nat Commun 12:5800, 2021; Calvanese et al., Nucleic Acids Res 52(10):5465-5477, 2024) applied to the twister ribozyme RNA family with three key applications: generating artificial twister ribozymes, designing potentially functional mutations of a natural wild type, and predicting mutational effects.

Supervised Chromatin Loop Detection Using Peakachu Version 2.

Peakachu is a supervised-learning-based approach that identifies chromatin loops from chromatin contact data. Here, we present Peakachu version 2, an updated version that significantly improves extensibility, usability, and computational efficiency compared to its predecessor. It features pretrained models tailored for a wide range of experimental platforms, such as Hi-C, Micro-C, ChIA-PET, HiChIP, HiCAR, and TrAC-loop. This chapter offers a step-by-step tutorial guiding users through the process of training Peakachu models from scratch and utilizing pretrained models to predict chromatin loops across various platforms.

GenerRNA: A generative pre-trained language model for de novo RNA design.

The design of RNA plays a crucial role in developing RNA vaccines, nucleic acid therapeutics, and innovative biotechnological tools. However, existing techniques frequently lack versatility across various tasks and are dependent on pre-defined secondary structure or other prior knowledge. To address these limitations, we introduce GenerRNA, a Transformer-based model inspired by the success of large language models (LLMs) in protein and molecule generation. GenerRNA is pre-trained on large-scale RNA sequences and capable of generating novel RNA sequences with stable secondary structures, while ensuring distinctiveness from existing sequences, thereby expanding our exploration of the RNA space. Moreover, GenerRNA can be fine-tuned on smaller, specialized datasets for specific subtasks, enabling the generation of RNAs with desired functionalities or properties without requiring any prior knowledge input. As a demonstration, we fine-tuned GenerRNA and successfully generated novel RNA sequences exhibiting high affinity for target proteins. Our work is the first application of a generative language model to RNA generation, presenting an innovative approach to RNA design.

Hydroxyl radical-induced C1'-H abstraction reaction of different artificial nucleotides.

CONTEXT: Recently, a few antiviral drugs viz Molnupiravir (EIDD-1931), Favipiravir, Ribavirin, Sofosbuvir, Galidesivir, and Remdesivir are shown to be beneficial against COVID-19 disease. These drugs bind to the viral RNA single strand to inhibit the virus genome replication. Similarly, recently, some artificial nucleotides, such as P, J, B, X, Z, V, S, and K were proposed to behave as potent antiviral candidates. However, their activity in the presence of the most reactive hydroxyl (OH) radical is not yet known. Here, the possibility of RNA strand break due to the OH radical-induced C1'-hydrogen (H) abstraction reaction of the above molecules (except Remdesivir) is studied in detail by considering their nucleotide conformation. The results are compared with those of the natural RNA nucleotides (G, C, A, and U). Due to low Gibbs barrier-free energy and high exothermicity, all these nucleotides (except Remdesivir) are prone to OH radical-induced C1'-H abstraction reaction. As Remdesivir contains a C1'-CN bond, the OH radical substitution reactions at the CN and C1' sites would likely liberate the catalytically important CN group, thereby downgrading its activity. METHOD: Initially, the B3LYP-D3 dispersion-corrected density functional theory method and 6-31 + G* basis set were used to optimize all reactant, transition state, and product complexes in the implicit aqueous medium. Subsequently, the structures of these complexes were further optimized by using the ωB97X-D dispersion-corrected density functional theory method and cc-PVTZ basis set in the aqueous medium. The IEFPCM method was used to model the aqueous medium.

Unraveling the complexity: Advanced methods in analyzing DNA, RNA, and protein interactions.

Exploring the intricate interplay within and between nucleic acids, as well as their interactions with proteins, holds pivotal significance in unraveling the molecular complexities steering cancer initiation and progression. To investigate these interactions, a diverse array of highly specific and sensitive molecular techniques has been developed. The selection of a particular technique depends on the specific nature of the interactions. Typically, researchers employ an amalgamation of these different techniques to obtain a comprehensive and holistic understanding of inter- and intramolecular interactions involving DNA-DNA, RNA-RNA, DNA-RNA, or protein-DNA/RNA. Examining nucleic acid conformation reveals alternative secondary structures beyond conventional ones that have implications for cancer pathways. Mutational hotspots in cancer often lie within sequences prone to adopting these alternative structures, highlighting the importance of investigating intra-genomic and intra-transcriptomic interactions, especially in the context of mutations, to deepen our understanding of oncology. Beyond these intramolecular interactions, the interplay between DNA and RNA leads to formations like DNA:RNA hybrids (known as R-loops) or even DNA:DNA:RNA triplex structures, both influencing biological processes that ultimately impact cancer. Protein-nucleic acid interactions are intrinsic cellular phenomena crucial in both normal and pathological conditions. In particular, genetic mutations or single amino acid variations can alter a protein's structure, function, and binding affinity, thus influencing cancer progression. It is thus, imperative to understand the differences between wild-type (WT) and mutated (MT) genes, transcripts, and proteins. The review aims to summarize the frequently employed methods and techniques for investigating interactions involving nucleic acids and proteins, highlighting recent advancements and diverse adaptations of each technique.

Cryo-EM: A window into the dynamic world of RNA molecules.

RNAs are critical for complex cellular functions, characterized by their structural versatility and ability to undergo conformational transitions in response to cellular cues. The elusive structures of RNAs are being unraveled with unprecedented clarity, thanks to the technological advancements in structural biology, including nuclear magnetic resonance (NMR), X-ray crystallography, cryo-electron microscopy (cryo-EM) etc. This review focuses on examining the revolutionary impact of cryo-EM on our comprehension of RNA structural dynamics, underscoring the technique's contributions to structural biology and envisioning the future trajectory of this rapidly evolving field.