RNANetMotif: Identifying sequence-structure RNA network motifs in RNA-protein binding sites

RNA molecules can adopt stable secondary and tertiary structures, which are essential in mediating physical interactions with other partners such as RNA binding proteins (RBPs) and in carrying out their cellular functions. In vivo and in vitro experiments such as RNAcompete and eCLIP have revealed in vitro binding preferences of RBPs to RNA oligomers and in vivo binding sites in cells. Analysis of these binding data showed that the structure properties of the RNAs in these binding sites are important determinants of the binding events; however, it has been a challenge to incorporate the structure information into an interpretable model. Here we describe a new approach, RNANetMotif, which takes predicted secondary structure of thousands of RNA sequences bound by an RBP as input and uses a graph theory approach to recognize enriched subgraphs. These enriched subgraphs are in essence shared sequence-structure elements that are important in RBP-RNA binding. To validate our approach, we performed RNA structure modeling via coarse-grained molecular dynamics folding simulations for selected 4 RBPs, and RNA-protein docking for LIN28B. The simulation results, e.g., solvent accessibility and energetics, further support the biological relevance of the discovered network subgraphs.     

Integrated sequence-structure motifs suffice to identify microRNA precursors

Background

Upwards of 1200 miRNA loci have hitherto been annotated in the human genome. The specific features defining a miRNA precursor and deciding its recognition and subsequent processing are not yet exhaustively described and miRNA loci can thus not be computationally identified with sufficient confidence.

Results

We rendered pre-miRNA and non-pre-miRNA hairpins as strings of integrated sequence-structure information, and used the software Teiresias to identify sequence-structure motifs (ss-motifs) of variable length in these data sets. Using only ss-motifs as features in a Support Vector Machine (SVM) algorithm for pre-miRNA identification achieved 99.2% specificity and 97.6% sensitivity on a human test data set, which is comparable to previously published algorithms employing combinations of sequence-structure and additional features. Further analysis of the ss-motif information contents revealed strongly significant deviations from those of the respective training sets, revealing important potential clues as to how the sequence and structural information of RNA hairpins are utilized by the miRNA processing apparatus.

Conclusion

Integrated sequence-structure motifs of variable length apparently capture nearly all information required to distinguish miRNA precursors from other stem-loop structures.     

Local gapped subforest alignment and its application in finding RNA structural motifs.

RNA molecules whose secondary structures contain similar substructures often have similar functions. Therefore, an important task in the study of RNA is to develop methods for discovering substructures in RNA secondary structures that occur frequently (also referred to as motifs). In this paper, we consider the problem of computing an optimal local alignment of two given labeled ordered forests F1 and F2. This problem asks for a substructure of F1 and a substructure of F2 that exhibit a high similarity. Since an RNA molecule's secondary structure can be represented as a labeled ordered forest, the problem we study has a direct application to finding potential motifs. We generalize the previously studied concept of a closed subforest to a gapped subforest and present the first algorithm for computing the optimal local gapped subforest alignment of F1 and F2. We also show that our technique can improve the time and space complexity of the previously most efficient algorithm for optimal local closed subforest alignment. Furthermore, we prove that a special case of our local gapped subforest alignment problem is equivalent to a problem known in the literature as the local sequence-structure alignment problem (lssa) and modify our main algorithm to obtain a much faster algorithm for lssa than the one previously proposed. An implementation of our algorithm is available at www.comp.nus.edu.sg/~bioinfo/LGSFAligner/. Its running time is significantly faster than the original lssa program.     

Analysis of RNA motifs

RNA motifs are directed and ordered stacked arrays of non-Watson-Crick base pairs forming distinctive foldings of the phosphodiester backbones of the interacting RNA strands. They correspond to the 'loops' - hairpin, internal and junction - that intersperse the Watson-Crick two-dimensional helices as seen in two-dimensional representations of RNA structure. RNA motifs mediate the specific interactions that induce the compact folding of complex RNAs. RNA motifs also constitute specific protein or ligand binding sites. A given motif is characterized by all the sequences that fold into essentially identical three-dimensional structures with the same ordered array of isosteric non-Watson-Crick base pairs. It is therefore crucial, when analyzing a three-dimensional RNA structure in order to identify and compare motifs, to first classify its non-Watson-Crick base pairs geometrically.     

Accurate multiple sequence-structure alignment of RNA sequences using combinatorial optimization

Background  

The discovery of functional non-coding RNA sequences has led to an increasing interest in algorithms related to RNA analysis. Traditional sequence alignment algorithms, however, fail at computing reliable alignments of low-homology RNA sequences. The spatial conformation of RNA sequences largely determines their function, and therefore RNA alignment algorithms have to take structural information into account.     

The annotation of RNA motifs

The recent deluge of new RNA structures, including complete atomic-resolution views of both subunits of the ribosome, has on the one hand literally overwhelmed our individual abilities to comprehend the diversity of RNA structure, and on the other hand presented us with new opportunities for comprehensive use of RNA sequences for comparative genetic, evolutionary and phylogenetic studies. Two concepts are key to understanding RNA structure: hierarchical organization of global structure and isostericity of local interactions. Global structure changes extremely slowly, as it relies on conserved long-range tertiary interactions. Tertiary RNA-RNA and quaternary RNA-protein interactions are mediated by RNA motifs, defined as recurrent and ordered arrays of non-Watson-Crick base-pairs. A single RNA motif comprises a family of sequences, all of which can fold into the same three-dimensional structure and can mediate the same interaction(s). The chemistry and geometry of base pairing constrain the evolution of motifs in such a way that random mutations that occur within motifs are accepted or rejected insofar as they can mediate a similar ordered array of interactions. The steps involved in the analysis and annotation of RNA motifs in 3D structures are: (a) decomposition of each motif into non-Watson-Crick base-pairs; (b) geometric classification of each basepair; (c) identification of isosteric substitutions for each basepair by comparison to isostericity matrices; (d) alignment of homologous sequences using the isostericity matrices to identify corresponding positions in the crystal structure; (e) acceptance or rejection of the null hypothesis that the motif is conserved.     

Synthetic biology with RNA motifs

Structural motifs in naturally occurring RNAs and RNPs can be employed as new molecular parts for synthetic biology to facilitate the development of novel devices and systems that modulate cellular functions. In this review, we focus on the following: (i) experimental evolution techniques of RNA molecules in vitro and (ii) their applications for regulating gene expression systems in vivo. For experimental evolution, new artificial RNA aptamers and RNA enzymes (ribozymes) have been selected in vitro. These functional RNA molecules are likely to be applicable in the reprogramming of existing gene regulatory systems. Furthermore, they may be used for designing hypothetical RNA-based living systems in the so-called RNA world. For the regulation of gene expressions in living cells, the development of new riboswitches allows us to modulate the target gene expression in a tailor-made manner. Moreover, recently RNA-based synthetic genetic circuits have been reported by employing functional RNA molecules, expanding the repertory of synthetic biology with RNA motifs.     

A structural database for k-turn motifs in RNA

The kink-turn (k-turn) is a common structural motif in RNA that introduces a tight kink into the helical axis. k-turns play an important architectural role in RNA structures and serve as binding sites for a number of proteins. We have created a database of known and postulated k-turn sequences and three-dimensional (3D) structures, available via the internet. This site provides (1) a database of sequence and structure, as a resource for the RNA community, and (2) a tool to enable the manipulation and comparison of 3D structures where known.     

RNA bulges as architectural and recognition motifs

RNA bulges constitute versatile structural motifs in the assembly of RNA architectures. Three-dimensional structures of RNA molecules and their complexes reveal the role of bulges in RNA architectures and illustrate the molecular mechanisms by which they confer intramolecular interactions and intermolecular recognition.     

RNA loop-loop interactions as dynamic functional motifs

RNA loop-loop interactions are frequently used to trigger initial recognition between two RNA molecules. In this review, we present selected well-documented cases that illustrate the diversity of biological processes using RNA loop-loop recognition properties. The first one is related to natural antisense RNAs that play a variety of regulatory functions in bacteria and their extra-chromosomal elements. The second one concerns the dimerization of HIV-1 genomic RNA, which is responsible for the encapsidation of a diploid RNA genome. The third one concerns RNA interactions involving double-loop interactions. These are used by the bicoid mRNA to form dimers, a property that appears to be important for mRNA localization in drosophila embryo, and by bacteriophage phi29 pRNA which forms hexamers that participate in the translocation of the DNA genome through the portal vertex of the capsid. Despite the high diversity of systems and mechanisms, some common features can be highlighted. (1) Efficient recognition requires rapid bi-molecular binding rates, regardless of the RNA pairing scheme. (2) The initial recognition is favored by particular conformations of the loops enabling a proper presentation of nucleotides (generally a restricted number) that initiate the recognition process. (3) The fate of the initial reversible loop-loop complex is dictated by both functional and structural constraints. RNA structures have evolved either to "freeze" the initial complex, or to convert it into a more stable one, which involves propagation of intermolecular interactions along topologically feasible pathways. Stabilization of the initial complex may also be assisted by proteins and/or formation of additional contacts.     

Modeling RNA tertiary structure motifs by graph-grammars

A new approach, graph-grammars, to encode RNA tertiary structure patterns is introduced and exemplified with the classical sarcin-ricin motif. The sarcin-ricin motif is found in the stem of the crucial ribosomal loop E (also referred to as the sarcin-ricin loop), which is sensitive to the alpha-sarcin and ricin toxins. Here, we generate a graph-grammar for the sarcin-ricin motif and apply it to derive putative sequences that would fold in this motif. The biological relevance of the derived sequences is confirmed by a comparison with those found in known sarcin-ricin sites in an alignment of over 800 bacterial 23S ribosomal RNAs. The comparison raised alternative alignments in few sarcin-ricin sites, which were assessed using tertiary structure predictions and 3D modeling. The sarcin-ricin motif graph-grammar was built with indivisible nucleotide interaction cycles that were recently observed in structured RNAs. A comparison of the sequences and 3D structures of each cycle that constitute the sarcin-ricin motif gave us additional insights about RNA sequence-structure relationships. In particular, this analysis revealed the sequence space of an RNA motif depends on a structural context that goes beyond the single base pairing and base-stacking interactions.     

A conditional random fields method for RNA sequence-structure relationship modeling and conformation sampling

Accurate tertiary structures are very important for the functional study of non-coding RNA molecules. However, predicting RNA tertiary structures is extremely challenging, because of a large conformation space to be explored and lack of an accurate scoring function differentiating the native structure from decoys. The fragment-based conformation sampling method (e.g. FARNA) bears shortcomings that the limited size of a fragment library makes it infeasible to represent all possible conformations well. A recent dynamic Bayesian network method, BARNACLE, overcomes the issue of fragment assembly. In addition, neither of these methods makes use of sequence information in sampling conformations. Here, we present a new probabilistic graphical model, conditional random fields (CRFs), to model RNA sequence-structure relationship, which enables us to accurately estimate the probability of an RNA conformation from sequence. Coupled with a novel tree-guided sampling scheme, our CRF model is then applied to RNA conformation sampling. Experimental results show that our CRF method can model RNA sequence-structure relationship well and sequence information is important for conformation sampling. Our method, named as TreeFolder, generates a much higher percentage of native-like decoys than FARNA and BARNACLE, although we use the same simple energy function as BARNACLE. CONTACT: zywang@ttic.edu; j3xu@ttic.edu SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.     

A Bayes-optimal sequence-structure theory that unifies protein sequence-structure recognition and alignment

A rigorous Bayesian analysis is presented that unifies protein sequence-structure alignment and recognition. Given a sequence, explicit formulae are derived to select (1) its globally most probable core structure from a structure library; (2) its globally most probable alignment to a given core structure; (3) its most probable joint core structure and alignment chosen globally across the entire library; and (4) its most probable individual segments, secondary structure, and super-secondary structures across the entire library. The computations involved are NP-hard in the general case (3D-3D). Fast exact recursions for the restricted sequence singleton-only (1D-3D) case are given. Conclusions include: (a) the most probable joint core structure and alignment is not necessarily the most probable alignment of the most probable core structure, but rather maximizes the product of core and alignment probabilities; (b) use of a sequence-independent linear or affine gap penalty may result in the highest-probability threading not having the lowest score; (c) selecting the most probable core structure from the library (core structure selection or fold recognition only) involves comparing probabilities summed over all possible alignments of the sequence to the core, and not comparing individual optimal (or near-optimal) sequence-structure alignments; and (d) assuming uninformative priors, core structure selection is equivalent to comparing the ratio of two global means.     

Detection of common motifs in RNA secondary structures.

We describe a novel computerized system for comparison of RNA secondary structures and demonstrate its use for experimental studies. The system is able to screen a very large number of structures, to cluster similar structures and to detect specific structural motifs. In particular, the system is useful for detecting mutations with specific structural effects among all possible point mutations, and for predicting compensatory mutations that will restore the wild type structure. The algorithms are independent of the folding rules that are used to generate the secondary structures.     

Arrangement of 3D structural motifs in ribosomal RNA

Structural 3D motifs in RNA play an important role in the RNA stability and function. Previous studies have focused on the characterization and discovery of 3D motifs in RNA secondary and tertiary structures. However, statistical analyses of the distribution of 3D motifs along the RNA appear to be lacking. Herein, we present a novel strategy for evaluating the distribution of 3D motifs along the RNA chain and those motifs whose distributions are significantly non-random are identified. By applying it to the X-ray structure of the large ribosomal subunit from Haloarcula marismortui, helical motifs were found to cluster together along the chain and in the 3D structure, whereas the known tetraloops tend to be sequentially and spatially dispersed. That the distribution of key structural motifs such as tetraloops differ significantly from a random one suggests that our method could also be used to detect novel 3D motifs of any size in sufficiently long/large RNA structures. The motif distribution type can help in the prediction and design of 3D structures of large RNA molecules.     

The distribution of RNA motifs in natural sequences.

Functional analysis of genome sequences has largely ignored RNA genes and their structures. We introduce here the notion of 'ribonomics' to describe the search for the distribution of and eventually the determination of the physiological roles of these RNA structures found in the sequence databases. The utility of this approach is illustrated here by the identification in the GenBank database of RNA motifs having known binding or chemical activity. The frequency of these motifs indicates that most have originated from evolutionary drift and are selectively neutral. On the other hand, their distribution among species and their location within genes suggest that the destiny of these motifs may be more elaborate. For example, the hammerhead motif has a skewed organismal presence, is phylogenetically stable and recent work on a schistosome version confirms its in vivo biological activity. The under-representation of the valine-binding motif and the Rev-binding element in GenBank hints at a detrimental effect on cell growth or viability. Data on the presence and the location of these motifs may provide critical guidance in the design of experiments directed towards the understanding and the manipulation of RNA complexes and activities in vivo.