Principles for Predicting RNA Secondary Structure Design Difficulty

OpenASO: RNA Rescue - designing splice-modulating antisense oligonucleotides through community science

Splice-modulating antisense oligonucleotides (ASOs) are precision RNA-based drugs that are becoming an established modality to treat human disease. Previously, we reported the discovery of ASOs that target a novel, putative intronic RNA structure to rescue splicing of multiple pathogenic variants of F8 exon 16 that cause hemophilia A. However, the conventional approach to discovering splice-modulating ASOs is both laborious and expensive. Here, we describe an alternative paradigm that integrates data-driven RNA structure prediction and community science to discover splice-modulating ASOs. Using a splicing-deficient pathogenic variant of F8 exon 16 as a model, we show that 25% of the top-scoring molecules designed in the Eterna OpenASO challenge have a statistically significant impact on enhancing exon 16 splicing. Additionally, we show that a distinct combination of ASOs designed by Eterna players can additively enhance the inclusion of the splicing-deficient exon 16 variant. Together, our data suggests that crowdsourcing designs from a community of citizen scientists may accelerate the discovery of splice-modulating ASOs with potential to treat human disease.

Pioneering role of RNA in the early evolution of life

The catalytic, regulatory and structural properties of RNA, combined with their extraordinary ubiquity in cellular processes, are consistent with the proposal that this molecule played a much more conspicuous role in heredity and metabolism during the early stages of biological evolution. This review explores the pivotal role of RNA in the earliest life forms and its relevance in modern biological systems. It examines current models that study the early evolution of life, providing insights into the primordial RNA world and its legacy in contemporary biology.

Differentiable partition function calculation for RNA

Ribonucleic acid (RNA) is an essential molecule in a wide range of biological functions. In 1990, McCaskill introduced a dynamic programming algorithm for computing the partition function of an RNA sequence. McCaskill's algorithm is widely used today for understanding the thermodynamic properties of RNA. In this work, we introduce a generalization of McCaskill's algorithm that is well-defined over continuous inputs. Crucially, this enables us to implement an end-to-end differentiable partition function calculation. The derivative can be computed with respect to the input, or to any other fixed values, such as the parameters of the energy model. This builds a bridge between RNA thermodynamics and the tools of differentiable programming including deep learning as it enables the partition function to be incorporated directly into any end-to-end differentiable pipeline. To demonstrate the effectiveness of our new approach, we tackle the inverse folding problem directly using gradient optimization. We find that using the gradient to optimize the sequence directly is sufficient to arrive at sequences with a high probability of folding into the desired structure. This indicates that the gradients we compute are meaningful.

DNA Structure Design Is Improved Using an Artificially Expanded Alphabet of Base Pairs Including Loop and Mismatch Thermodynamic Parameters

We show that in silico design of DNA secondary structures is improved by extending the base pairing alphabet beyond A-T and G-C to include the pair between 2-amino-8-(1'-β-d-2'-deoxyribofuranosyl)-imidazo-[1,2-a]-1,3,5-triazin-(8H)-4-one and 6-amino-3-(1'-β-d-2'-deoxyribofuranosyl)-5-nitro-(1H)-pyridin-2-one, abbreviated as P and Z. To obtain the thermodynamic parameters needed to include P-Z pairs in the designs, we performed 47 optical melting experiments and combined the results with previous work to fit free energy and enthalpy nearest neighbor folding parameters for P-Z pairs and G-Z wobble pairs. We find G-Z pairs have stability comparable to that of A-T pairs and should therefore be included as base pairs in structure prediction and design algorithms. Additionally, we extrapolated the set of loop, terminal mismatch, and dangling end parameters to include the P and Z nucleotides. These parameters were incorporated into the RNAstructure software package for secondary structure prediction and analysis. Using the RNAstructure Design program, we solved 99 of the 100 design problems posed by Eterna using the ACGT alphabet or supplementing it with P-Z pairs. Extending the alphabet reduced the propensity of sequences to fold into off-target structures, as evaluated by the normalized ensemble defect (NED). The NED values were improved relative to those from the Eterna example solutions in 91 of 99 cases in which Eterna-player solutions were provided. P-Z-containing designs had average NED values of 0.040, significantly below the 0.074 of standard-DNA-only designs, and inclusion of the P-Z pairs decreased the time needed to converge on a design. This work provides a sample pipeline for inclusion of any expanded alphabet nucleotides into prediction and design workflows.

RNA design via structure-aware multifrontier ensemble optimization

MOTIVATION

RNA design is the search for a sequence or set of sequences that will fold to desired structure, also known as the inverse problem of RNA folding. However, the sequences designed by existing algorithms often suffer from low ensemble stability, which worsens for long sequence design. Additionally, for many methods only a small number of sequences satisfying the MFE criterion can be found by each run of design. These drawbacks limit their use cases.

RESULTS

We propose an innovative optimization paradigm, SAMFEO, which optimizes ensemble objectives (equilibrium probability or ensemble defect) by iterative search and yields a very large number of successfully designed RNA sequences as byproducts. We develop a search method which leverages structure level and ensemble level information at different stages of the optimization: initialization, sampling, mutation, and updating. Our work, while being less complicated than others, is the first algorithm that is able to design thousands of RNA sequences for the puzzles from the Eterna100 benchmark. In addition, our algorithm solves the most Eterna100 puzzles among all the general optimization based methods in our study. The only baseline solving more puzzles than our work is dependent on handcrafted heuristics designed for a specific folding model. Surprisingly, our approach shows superiority on designing long sequences for structures adapted from the database of 16S Ribosomal RNAs.

AVAILABILITY AND IMPLEMENTATION

Our source code and data used in this article is available at https://github.com/shanry/SAMFEO.

NanoAdventure: Development of a Text-Based Adventure Game in English, Spanish, and Chinese for Communicating about Nanotechnology and the Nanoscale

Video games and immersive, narrative experiences are often called upon to help students understand difficult scientific concepts, such as sense of scale. However, the development of educational video games requires expertise and, frequently, a sizable budget. Here, we report on the use of an interactive text-style video game, NanoAdventure, to communicate about sense of scale and nanotechnology to the public. NanoAdventure was developed on an open-source, free-to-use platform with simple coding and enhanced with free or low-cost assets. NanoAdventure was launched in three languages (English, Spanish, Chinese) and compared to textbook-style and blog-style control texts in a randomized study. Participants answered questions on their knowledge of nanotechnology and their attitudes toward nanotechnology before and after reading one randomly assigned text (textbook, blog, or NanoAdventure game). Our results demonstrate that interactive fiction is effective in communicating about sense of scale and nanotechnology as well as the relevance of nanotechnology to a general public. NanoAdventure was found to be the most "fun" and easy to read of all text styles by participants in a randomized trial. Here, we make the case for interactive "Choose Your Own Adventure" style games as another effective tool among educational game models for chemistry and science communication.

DNA Structure Design Is Improved Using an Artificially Expanded Alphabet of Base Pairs Including Loop and Mismatch Thermodynamic Parameters

We show that in silico design of DNA secondary structures is improved by extending the base pairing alphabet beyond A-T and G-C to include the pair between 2-amino-8-(1'-β-D-2'-deoxyribofuranosyl)-imidazo-[1,2- a ]-1,3,5-triazin-(8 H )-4-one and 6-amino-3-(1'-β-D-2'-deoxyribofuranosyl)-5-nitro-(1 H )-pyridin-2-one, simply P and Z. To obtain the thermodynamic parameters needed to include P-Z pairs in the designs, we performed 47 optical melting experiments and combined the results with previous work to fit a new set of free energy and enthalpy nearest neighbor folding parameters for P-Z pairs and G-Z wobble pairs. We find that G-Z pairs have stability comparable to A-T pairs and therefore should be considered quantitatively by structure prediction and design algorithms. Additionally, we extrapolated the set of loop, terminal mismatch, and dangling end parameters to include P and Z nucleotides. These parameters were incorporated into the RNAstructure software package for secondary structure prediction and analysis. Using the RNAstructure Design program, we solved 99 of the 100 design problems posed by Eterna using the ACGT alphabet or supplementing with P-Z pairs. Extending the alphabet reduced the propensity of sequences to fold into off-target structures, as evaluated by the normalized ensemble defect (NED). The NED values were improved relative to those from the Eterna example solutions in 91 of 99 cases where Eterna-player solutions were provided. P-Z-containing designs had average NED values of 0.040, significantly below the 0.074 of standard-DNA-only designs, and inclusion of the P-Z pairs decreased the time needed to converge on a design. This work provides a sample pipeline for inclusion of any expanded alphabet nucleotides into prediction and design workflows.

MncR: Late Integration Machine Learning Model for Classification of ncRNA Classes Using Sequence and Structural Encoding

Non-coding RNA (ncRNA) classes take over important housekeeping and regulatory functions and are quite heterogeneous in terms of length, sequence conservation and secondary structure. High-throughput sequencing reveals that the expressed novel ncRNAs and their classification are important to understand cell regulation and identify potential diagnostic and therapeutic biomarkers. To improve the classification of ncRNAs, we investigated different approaches of utilizing primary sequences and secondary structures as well as the late integration of both using machine learning models, including different neural network architectures. As input, we used the newest version of RNAcentral, focusing on six ncRNA classes, including lncRNA, rRNA, tRNA, miRNA, snRNA and snoRNA. The late integration of graph-encoded structural features and primary sequences in our MncR classifier achieved an overall accuracy of >97%, which could not be increased by more fine-grained subclassification. In comparison to the actual best-performing tool ncRDense, we had a minimal increase of 0.5% in all four overlapping ncRNA classes on a similar test set of sequences. In summary, MncR is not only more accurate than current ncRNA prediction tools but also allows the prediction of long ncRNA classes (lncRNAs, certain rRNAs) up to 12.000 nts and is trained on a more diverse ncRNA dataset retrieved from RNAcentral.

Design Approaches to Expand the Toolkit for Building Cotranscriptionally Encoded RNA Strand Displacement Circuits

Cotranscriptionally encoded RNA strand displacement (ctRSD) circuits are an emerging tool for programmable molecular computation, with potential applications spanning in vitro diagnostics to continuous computation inside living cells. In ctRSD circuits, RNA strand displacement components are continuously produced together via transcription. These RNA components can be rationally programmed through base pairing interactions to execute logic and signaling cascades. However, the small number of ctRSD components characterized to date limits circuit size and capabilities. Here, we characterize over 200 ctRSD gate sequences, exploring different input, output, and toehold sequences and changes to other design parameters, including domain lengths, ribozyme sequences, and the order in which gate strands are transcribed. This characterization provides a library of sequence domains for engineering ctRSD components, i.e., a toolkit, enabling circuits with up to 4-fold more inputs than previously possible. We also identify specific failure modes and systematically develop design approaches that reduce the likelihood of failure across different gate sequences. Lastly, we show the ctRSD gate design is robust to changes in transcriptional encoding, opening a broad design space for applications in more complex environments. Together, these results deliver an expanded toolkit and design approaches for building ctRSD circuits that will dramatically extend capabilities and potential applications.

Community science designed ribosomes with beneficial phenotypes

Functional design of ribosomes with mutant ribosomal RNA (rRNA) can expand opportunities for understanding molecular translation, building cells from the bottom-up, and engineering ribosomes with altered capabilities. However, such efforts are hampered by cell viability constraints, an enormous combinatorial sequence space, and limitations on large-scale, 3D design of RNA structures and functions. To address these challenges, we develop an integrated community science and experimental screening approach for rational design of ribosomes. This approach couples Eterna, an online video game that crowdsources RNA sequence design to community scientists in the form of puzzles, with in vitro ribosome synthesis, assembly, and translation in multiple design-build-test-learn cycles. We apply our framework to discover mutant rRNA sequences that improve protein synthesis in vitro and cell growth in vivo, relative to wild type ribosomes, under diverse environmental conditions. This work provides insights into rRNA sequence-function relationships and has implications for synthetic biology.

RNA secondary structure packages evaluated and improved by high-throughput experiments

Despite the popularity of computer-aided study and design of RNA molecules, little is known about the accuracy of commonly used structure modeling packages in tasks sensitive to ensemble properties of RNA. Here, we demonstrate that the EternaBench dataset, a set of more than 20,000 synthetic RNA constructs designed on the RNA design platform Eterna, provides incisive discriminative power in evaluating current packages in ensemble-oriented structure prediction tasks. We find that CONTRAfold and RNAsoft, packages with parameters derived through statistical learning, achieve consistently higher accuracy than more widely used packages in their standard settings, which derive parameters primarily from thermodynamic experiments. We hypothesized that training a multitask model with the varied data types in EternaBench might improve inference on ensemble-based prediction tasks. Indeed, the resulting model, named EternaFold, demonstrated improved performance that generalizes to diverse external datasets including complete messenger RNAs, viral genomes probed in human cells and synthetic designs modeling mRNA vaccines.

aRNAque: an evolutionary algorithm for inverse pseudoknotted RNA folding inspired by Lévy flights

BACKGROUND

We study in this work the inverse folding problem for RNA, which is the discovery of sequences that fold into given target secondary structures.

RESULTS

We implement a Lévy mutation scheme in an updated version of aRNAque an evolutionary inverse folding algorithm and apply it to the design of RNAs with and without pseudoknots. We find that the Lévy mutation scheme increases the diversity of designed RNA sequences and reduces the average number of evaluations of the evolutionary algorithm. Compared to antaRNA, aRNAque CPU time is higher but more successful in finding designed sequences that fold correctly into the target structures.

CONCLUSION

We propose that a Lévy flight offers a better standard mutation scheme for optimizing RNA design. Our new version of aRNAque is available on GitHub as a python script and the benchmark results show improved performance on both Pseudobase++ and the Eterna100 datasets, compared to existing inverse folding tools.

Cotranscriptionally encoded RNA strand displacement circuits

Engineered molecular circuits that process information in biological systems could address emerging human health and biomanufacturing needs. However, such circuits can be difficult to rationally design and scale. DNA-based strand displacement reactions have demonstrated the largest and most computationally powerful molecular circuits to date but are limited in biological systems due to the difficulty in genetically encoding components. Here, we develop scalable cotranscriptionally encoded RNA strand displacement (ctRSD) circuits that are rationally programmed via base pairing interactions. ctRSD circuits address the limitations of DNA-based strand displacement circuits by isothermally producing circuit components via transcription. We demonstrate circuit programmability in vitro by implementing logic and amplification elements, as well as multilayer cascades. Furthermore, we show that circuit kinetics are accurately predicted by a simple model of coupled transcription and strand displacement, enabling model-driven design. We envision ctRSD circuits will enable the rational design of powerful molecular circuits that operate in biological systems, including living cells.

Deep learning models for predicting RNA degradation via dual crowdsourcing

Medicines based on messenger RNA (mRNA) hold immense potential, as evidenced by their rapid deployment as COVID-19 vaccines. However, worldwide distribution of mRNA molecules has been limited by their thermostability, which is fundamentally limited by the intrinsic instability of RNA molecules to a chemical degradation reaction called in-line hydrolysis. Predicting the degradation of an RNA molecule is a key task in designing more stable RNA-based therapeutics. Here, we describe a crowdsourced machine learning competition ('Stanford OpenVaccine') on Kaggle, involving single-nucleotide resolution measurements on 6,043 diverse 102-130-nucleotide RNA constructs that were themselves solicited through crowdsourcing on the RNA design platform Eterna. The entire experiment was completed in less than 6 months, and 41% of nucleotide-level predictions from the winning model were within experimental error of the ground truth measurement. Furthermore, these models generalized to blindly predicting orthogonal degradation data on much longer mRNA molecules (504-1,588 nucleotides) with improved accuracy compared with previously published models. These results indicate that such models can represent in-line hydrolysis with excellent accuracy, supporting their use for designing stabilized messenger RNAs. The integration of two crowdsourcing platforms, one for dataset creation and another for machine learning, may be fruitful for other urgent problems that demand scientific discovery on rapid timescales.

Deep learning models for predicting RNA degradation via dual crowdsourcing

Messenger RNA-based medicines hold immense potential, as evidenced by their rapid deployment as COVID-19 vaccines. However, worldwide distribution of mRNA molecules has been limited by their thermostability, which is fundamentally limited by the intrinsic instability of RNA molecules to a chemical degradation reaction called in-line hydrolysis. Predicting the degradation of an RNA molecule is a key task in designing more stable RNA-based therapeutics. Here, we describe a crowdsourced machine learning competition ("Stanford OpenVaccine") on Kaggle, involving single-nucleotide resolution measurements on 6043 102-130-nucleotide diverse RNA constructs that were themselves solicited through crowdsourcing on the RNA design platform Eterna. The entire experiment was completed in less than 6 months, and 41% of nucleotide-level predictions from the winning model were within experimental error of the ground truth measurement. Furthermore, these models generalized to blindly predicting orthogonal degradation data on much longer mRNA molecules (504-1588 nucleotides) with improved accuracy compared to previously published models. Top teams integrated natural language processing architectures and data augmentation techniques with predictions from previous dynamic programming models for RNA secondary structure. These results indicate that such models are capable of representing in-line hydrolysis with excellent accuracy, supporting their use for designing stabilized messenger RNAs. The integration of two crowdsourcing platforms, one for data set creation and another for machine learning, may be fruitful for other urgent problems that demand scientific discovery on rapid timescales.

MoiRNAiFold: a novel tool for complex in silico RNA design

Novel tools for in silico design of RNA constructs such as riboregulators are required in order to reduce time and cost to production for the development of diagnostic and therapeutic advances. Here, we present MoiRNAiFold, a versatile and user-friendly tool for de novo synthetic RNA design. MoiRNAiFold is based on Constraint Programming and it includes novel variable types, heuristics and restart strategies for Large Neighborhood Search. Moreover, this software can handle dozens of design constraints and quality measures and improves features for RNA regulation control of gene expression, such as Translation Efficiency calculation. We demonstrate that MoiRNAiFold outperforms any previous software in benchmarking structural RNA puzzles from EteRNA. Importantly, with regard to biologically relevant RNA designs, we focus on RNA riboregulators, demonstrating that the designed RNA sequences are functional both in vitro and in vivo. Overall, we have generated a powerful tool for de novo complex RNA design that we make freely available as a web server (https://moiraibiodesign.com/design/).

Pepblock Builder VR - An Open-Source Tool for Gaming-Based Bio-Edutainment in Interactive Protein Design

Proteins mediate and perform various fundamental functions of life. This versatility of protein function is an attribute of its 3D structure. In recent years, our understanding of protein 3D structure has been complemented with advances in computational and mathematical tools for protein modelling and protein design. 3D molecular visualisation is an essential part in every protein design and protein modelling workflow. Over the years, stand-alone and web-based molecular visualisation tools have been used to emulate three-dimensional view on computers. The advent of virtual reality provided the scope for immersive control of molecular visualisation. While these technologies have significantly improved our insights into protein modelling, designing new proteins with a defined function remains a complicated process. Current tools to design proteins lack user-interactivity and demand high computational skills. In this work, we present the Pepblock Builder VR, a gaming-based molecular visualisation tool for bio-edutainment and understanding protein design. Simulating the concepts of protein design and incorporating gaming principles into molecular visualisation promotes effective game-based learning. Unlike traditional sequence-based protein design and fragment-based stitching, the Pepblock Builder VR provides a building block style environment for complex structure building. This provides users a unique visual structure building experience. Furthermore, the inclusion of virtual reality to the Pepblock Builder VR brings immersive learning and provides users with "being there" experience in protein visualisation. The Pepblock Builder VR works both as a stand-alone and VR-based application, and with a gamified user interface, the Pepblock Builder VR aims to expand the horizons of scientific data generation to the masses.

Biomolecular Modeling and Simulation: A Prospering Multidisciplinary Field

We reassess progress in the field of biomolecular modeling and simulation, following up on our perspective published in 2011. By reviewing metrics for the field's productivity and providing examples of success, we underscore the productive phase of the field, whose short-term expectations were overestimated and long-term effects underestimated. Such successes include prediction of structures and mechanisms; generation of new insights into biomolecular activity; and thriving collaborations between modeling and experimentation, including experiments driven by modeling. We also discuss the impact of field exercises and web games on the field's progress. Overall, we note tremendous success by the biomolecular modeling community in utilization of computer power; improvement in force fields; and development and application of new algorithms, notably machine learning and artificial intelligence. The combined advances are enhancing the accuracy andscope of modeling and simulation, establishing an exemplary discipline where experiment and theory or simulations are full partners.

RNA Engineering for Public Health: Innovations in RNA-Based Diagnostics and Therapeutics

RNA is essential for cellular function: From sensing intra- and extracellular signals to controlling gene expression, RNA mediates a diverse and expansive list of molecular processes. A long-standing goal of synthetic biology has been to develop RNA engineering principles that can be used to harness and reprogram these RNA-mediated processes to engineer biological systems to solve pressing global challenges. Recent advances in the field of RNA engineering are bringing this to fruition, enabling the creation of RNA-based tools to combat some of the most urgent public health crises. Specifically, new diagnostics using engineered RNAs are able to detect both pathogens and chemicals while generating an easily detectable fluorescent signal as an indicator. New classes of vaccines and therapeutics are also using engineered RNAs to target a wide range of genetic and pathogenic diseases. Here, we discuss the recent breakthroughs in RNA engineering enabling these innovations and examine how advances in RNA design promise to accelerate the impact of engineered RNA systems.

Inverse RNA Folding Workflow to Design and Test Ribozymes that Include Pseudoknots

Ribozymes are RNAs that catalyze reactions. They occur in nature, and can also be evolved in vitro to catalyze novel reactions. This chapter provides detailed protocols for using inverse folding software to design a ribozyme sequence that will fold to a known ribozyme secondary structure and for testing the catalytic activity of the sequence experimentally. This protocol is able to design sequences that include pseudoknots, which is important as all naturally occurring full-length ribozymes have pseudoknots. The starting point is the known pseudoknot-containing secondary structure of the ribozyme and knowledge of any nucleotides whose identity is required for function. The output of the protocol is a set of sequences that have been tested for function. Using this protocol, we were previously successful at designing highly active double-pseudoknotted HDV ribozymes.

Engineered RNA-Interacting CRISPR Guide RNAs for Genetic Sensing and Diagnostics

CRISPR guide RNAs (gRNAs) can be programmed with relative ease to allow the genetic editing of nearly any DNA or RNA sequence. Here, we propose novel molecular architectures to achieve RNA-dependent modulation of CRISPR activity in response to specific RNA molecules. We designed and tested, in both living Escherichia coli cells and cell-free assays for rapid prototyping, cis-repressed RNA-interacting guide RNA (igRNA) that switch to their active state only upon interaction with small RNA fragments or long RNA transcripts, including pathogen-derived mRNAs of medical relevance such as the human immunodeficiency virus infectivity factor. The proposed CRISPR-igRNAs are fully customizable and easily adaptable to the majority if not all the available CRISPR-Cas variants to modulate a variety of genetic functions in response to specific cellular conditions, providing orthogonal activation and increased specificity. We thereby foresee a large scope of application for therapeutic, diagnostic, and biotech applications in both prokaryotic and eukaryotic systems.

Evidence of an Unusual Poly(A) RNA Signature Detected by High-Throughput Chemical Mapping

Homopolymeric adenosine RNA plays numerous roles in both cells and noncellular genetic material. We report herein an unusual poly(A) signature in chemical mapping data generated by the Eterna Massive Open Laboratory. Poly(A) sequences of length seven or more show unexpected results in the selective 2'-hydroxyl acylation read out by primer extension (SHAPE) and dimethyl sulfate (DMS) chemical probing. This unusual signature first appears in poly(A) sequences of length seven and grows to its maximum strength at length ∼10. In a long poly(A) sequence, the substitution of a single A by any other nucleotide disrupts the signature, but only for the 6 or so nucleotides on the 5' side of the substitution.

Designing RNA Secondary Structures Is Hard

A ribonucleic acid (RNA) sequence is a word over an alphabet on four elements [Formula: see text] called bases. RNA sequences fold into secondary structures where some bases pair with one another, while others remain unpaired. The two fundamental problems in RNA algorithmic are to predict how sequences fold within some models of energy and to design sequences of bases that will fold into targeted secondary structures. Predicting how a given RNA sequence folds into a pseudoknot-free secondary structure is known to be solvable in cubic time since the eighties and in truly subcubic time by a recent result of Bringmann et al. (FOCS, 2016), whereas Lyngsø has shown it is computationally hard if pseudoknots are allowed (ICALP, 2004). As a stark contrast, it is unknown whether or not designing a given RNA secondary structure is a tractable task; this has been raised as a challenging open question by Condon (ICALP, 2003). Because of its crucial importance in a number of fields such as pharmaceutical research and biochemistry, there are dozens of heuristics and software libraries dedicated to the RNA secondary structure design. It is therefore rather surprising that the computational complexity of this central problem in bioinformatics has been unsettled for decades. In this article, we show that in the simplest model of energy, which is the Watson-Crick model, the design of secondary structures is computationally hard if one adds natural constraints of the form: index i of the sequence has to be labeled by base b. This negative result suggests that the same lower bound holds for more realistic models of energy. It is noteworthy that the additional constraints are by no means artificial: they are provided by all the RNA design pieces of software and they do correspond to the actual practice (e.g., the instances of the EteRNA project).

Predicting RNA secondary structure via adaptive deep recurrent neural networks with energy-based filter

Background

RNA secondary structure prediction is an important issue in structural bioinformatics, and RNA pseudoknotted secondary structure prediction represents an NP-hard problem. Recently, many different machine-learning methods, Markov models, and neural networks have been employed for this problem, with encouraging results regarding their predictive accuracy; however, their performances are usually limited by the requirements of the learning model and over-fitting, which requires use of a fixed number of training features. Because most natural biological sequences have variable lengths, the sequences have to be truncated before the features are employed by the learning model, which not only leads to the loss of information but also destroys biological-sequence integrity.

Results

To address this problem, we propose an adaptive sequence length based on deep-learning model and integrate an energy-based filter to remove the over-fitting base pairs.

Conclusions

Comparative experiments conducted on an authoritative dataset RNA STRAND (RNA secondary STRucture and statistical Analysis Database) revealed a 12% higher accuracy relative to three currently used methods.

Donors, authors, and owners: how is genomic citizen science addressing interests in research outputs?

BACKGROUND

Citizen science is increasingly prevalent in the biomedical sciences, including the field of human genomics. Genomic citizen science initiatives present new opportunities to engage individuals in scientific discovery, but they also are provoking new questions regarding who owns the outputs of the research, including intangible ideas and discoveries and tangible writings, tools, technologies, and products. The legal and ethical claims of participants to research outputs become stronger-and also more likely to conflict with those of institution-based researchers and other stakeholders-as participants become more involved, quantitatively and qualitatively, in the research process. It is not yet known, however, how genomic citizen science initiatives are managing the interests of their participants in accessing and controlling research outputs in practice. To help fill this gap, we conducted an in-depth review of relevant policies and practices of U.S.-based genomic citizen science initiatives.

METHODS

We queried the peer-reviewed literature and grey literature to identify 22 genomic citizen science initiatives that satisfied six inclusion criteria. A data collection form was used to capture initiative features, policies, and practices relevant to participants' access to and control over research outputs.

RESULTS

This analysis revealed that the genomic citizen science landscape is diverse and includes many initiatives that do not have institutional affiliations. Two trends that are in apparent tension were identified: commercialization and operationalization of a philosophy of openness. While most initiatives supported participants' access to research outputs, including datasets and published findings, none supported participants' control over results via intellectual property, licensing, or commercialization rights. However, several initiatives disclaimed their own rights to profit from outputs.

CONCLUSIONS

There are opportunities for citizen science initiatives to incorporate more features that support participants' access to and control over research outputs, consistent with their specific objectives, operations, and technical capabilities.

Scientific Discovery Games for Biomedical Research

Over the past decade, scientific discovery games (SDGs) have emerged as a viable approach for biomedical research, engaging hundreds of thousands of volunteer players and resulting in numerous scientific publications. After describing the origins of this novel research approach, we review the scientific output of SDGs across molecular modeling, sequence alignment, neuroscience, pathology, cellular biology, genomics, and human cognition. We find compelling results and technical innovations arising in problem-oriented games such as Foldit and Eterna and in data-oriented games such as EyeWire and Project Discovery. We discuss emergent properties of player communities shared across different projects, including the diversity of communities and the extraordinary contributions of some volunteers, such as paper writing. Finally, we highlight connections to artificial intelligence, biological cloud laboratories, new game genres, science education, and open science that may drive the next generation of SDGs.

EternaBrain: Automated RNA design through move sets and strategies from an Internet-scale RNA videogame

Emerging RNA-based approaches to disease detection and gene therapy require RNA sequences that fold into specific base-pairing patterns, but computational algorithms generally remain inadequate for these secondary structure design tasks. The Eterna project has crowdsourced RNA design to human video game players in the form of puzzles that reach extraordinary difficulty. Here, we demonstrate that Eterna participants' moves and strategies can be leveraged to improve automated computational RNA design. We present an eternamoves-large repository consisting of 1.8 million of player moves on 12 of the most-played Eterna puzzles as well as an eternamoves-select repository of 30,477 moves from the top 72 players on a select set of more advanced puzzles. On eternamoves-select, we present a multilayer convolutional neural network (CNN) EternaBrain that achieves test accuracies of 51% and 34% in base prediction and location prediction, respectively, suggesting that top players' moves are partially stereotyped. Pipelining this CNN's move predictions with single-action-playout (SAP) of six strategies compiled by human players solves 61 out of 100 independent puzzles in the Eterna100 benchmark. EternaBrain-SAP outperforms previously published RNA design algorithms and achieves similar or better performance than a newer generation of deep learning methods, while being largely orthogonal to these other methods. Our study provides useful lessons for future efforts to achieve human-competitive performance with automated RNA design algorithms.

A pipeline for computational design of novel RNA-like topologies

Designing novel RNA topologies is a challenge, with important therapeutic and industrial applications. We describe a computational pipeline for design of novel RNA topologies based on our coarse-grained RNA-As-Graphs (RAG) framework. RAG represents RNA structures as tree graphs and describes RNA secondary (2D) structure topologies (currently up to 13 vertices, ≈260 nucleotides). We have previously identified novel graph topologies that are RNA-like among these. Here we describe a systematic design pipeline and illustrate design for six broad design problems using recently developed tools for graph-partitioning and fragment assembly (F-RAG). Following partitioning of the target graph, corresponding atomic fragments from our RAG-3D database are combined using F-RAG, and the candidate atomic models are scored using a knowledge-based potential developed for 3D structure prediction. The sequences of the top scoring models are screened further using available tools for 2D structure prediction. The results indicate that our modular approach based on RNA-like topologies rather than specific 2D structures allows for greater flexibility in the design process, and generates a large number of candidate sequences quickly. Experimental structure probing using SHAPE-MaP for two sequences agree with our predictions and suggest that our combined tools yield excellent candidates for further sequence and experimental screening.

Solving the RNA design problem with reinforcement learning

We use reinforcement learning to train an agent for computational RNA design: given a target secondary structure, design a sequence that folds to that structure in silico. Our agent uses a novel graph convolutional architecture allowing a single model to be applied to arbitrary target structures of any length. After training it on randomly generated targets, we test it on the Eterna100 benchmark and find it outperforms all previous algorithms. Analysis of its solutions shows it has successfully learned some advanced strategies identified by players of the game Eterna, allowing it to solve some very difficult structures. On the other hand, it has failed to learn other strategies, possibly because they were not required for the targets in the training set. This suggests the possibility that future improvements to the training protocol may yield further gains in performance.

Composing RNA Nanostructures from a Syntax of RNA Structural Modules

Natural stable RNAs fold and assemble into complex three-dimensional architectures by relying on the hierarchical formation of intricate, recurrent networks of noncovalent tertiary interactions. These sequence-dependent networks specify RNA structural modules enabling orientational and topological control of helical struts to form larger self-folding domains. Borrowing concepts from linguistics, we defined an extended structural syntax of RNA modules for programming RNA strands to assemble into complex, responsive nanostructures under both thermodynamic and kinetic control. Based on this syntax, various RNA building blocks promote the multimolecular assembly of objects with well-defined three-dimensional shapes as well as the isothermal folding of long RNAs into complex single-stranded nanostructures during transcription. This work offers a glimpse of the limitless potential of RNA as an informational medium for designing programmable and functional nanomaterials useful for synthetic biology, nanomedicine, and nanotechnology.

The academic-industrial complex: navigating the translational and cultural divide

In general, the fruits of academic discoveries can only be realized through joint efforts with industry. However, the poor reproducibility of much academic research has damaged credibility and jeopardized translational efforts that could benefit patients. Meanwhile, journals are rife with articles bemoaning the limited productivity and increasing costs of the biopharmaceutical industry and its resultant predilection for mergers and reorganizations while decreasing internal research efforts. The ensuing disarray and uncertainty has created tremendous opportunities for academia and industry to form even closer ties, and to embrace new operational and financial models to their joint benefit. This review article offers a personal perspective on the opportunities, models and approaches that harness the increased interface and growing interdependency between biomedical research institutes, the biopharmaceutical industry and the technological world.