More complete gene silencing by fewer siRNAs: transparent optimized design and biophysical signature

OligoFormer: an accurate and robust prediction method for siRNA design

MOTIVATION

RNA interference (RNAi) has become a widely used experimental approach for post-transcriptional regulation and is increasingly showing its potential as future targeted drugs. However, the prediction of highly efficient siRNAs (small interfering RNAs) is still hindered by dataset biases, the inadequacy of prediction methods, and the presence of off-target effects. To overcome these limitations, we propose an accurate and robust prediction method, OligoFormer, for siRNA design.

RESULTS

OligoFormer comprises three different modules including thermodynamic calculation, RNA-FM module, and Oligo encoder. Oligo encoder is the core module based on the transformer encoder. Taking siRNA and mRNA sequences as input, OligoFormer can obtain thermodynamic parameters, RNA-FM embedding, and Oligo embedding through these three modules, respectively. We carefully benchmarked OligoFormer against six comparable methods on siRNA efficacy datasets. OligoFormer outperforms all the other methods, with an average improvement of 9% in AUC, 6.6% in PRC, 9.8% in F1 score, and 5.1% in PCC compared to the best method among them in our inter-dataset validation. We also provide a comprehensive pipeline with prediction of siRNA efficacy and off-target effects using PITA score and TargetScan score. The ablation study shows RNA-FM module and thermodynamic parameters improved the performance and accelerated convergence of OligoFormer. The saliency maps by gradient backpropagation and base preference maps show certain base preferences in initial and terminal region of siRNAs.

AVAILABILITY AND IMPLEMENTATION

The source code of OligoFormer is freely available on GitHub at: https://github.com/lulab/OligoFormer. Docker image of OligoFormer is freely available on the docker hub at https://hub.docker.com/r/yilanbai/oligoformer.

RNAi-Mediated Silencing in the Insect Cell-Baculovirus Expression System

RNA interference (RNAi) serves as an indispensable tool for gene function studies and has been substantiated through extensive research for its practical applications in the baculovirus expression vector system (BEVS). This chapter expands the RNAi toolkit in insect cell culture by including small interfering RNA (siRNA) in the protocol, in addition to the conventional use of double-stranded RNA (dsRNA). This chapter also brings attention to key design and reporting considerations, based on Minimum Information About an RNAi Experiment (MIARE) guidelines. Recommendations regarding online tools for dsRNA and siRNA design are provided, along with guidance on choosing suitable methods for measuring silencing outcomes.

Cheminformatics Modeling of Gene Silencing for Both Natural and Chemically Modified siRNAs

In designing effective siRNAs for a specific mRNA target, it is critically important to have predictive models for the potency of siRNAs. None of the published methods characterized the chemical structures of individual nucleotides constituting a siRNA molecule; therefore, they cannot predict the potency of gene silencing by chemically modified siRNAs (cm-siRNA). We propose a new approach that can predict the potency of gene silencing by cm-siRNAs, which characterizes each nucleotide (NT) using 12 BCUT cheminformatics descriptors describing its charge distribution, hydrophobic and polar properties. Thus, a 21-NT siRNA molecule is described by 252 descriptors resulting from concatenating all the BCUT values of its composing nucleotides. Partial Least Square is employed to develop statistical models. The Huesken data (2431 natural siRNA molecules) were used to perform model building and evaluation for natural siRNAs. Our results were comparable with or superior to those from Huesken's algorithm. The Bramsen dataset (48 cm-siRNAs) was used to build and test the models for cm-siRNAs. The predictive r2 of the resulting models reached 0.65 (or Pearson r values of 0.82). Thus, this new method can be used to successfully model gene silencing potency by both natural and chemically modified siRNA molecules.

siRNA Design and GalNAc-Empowered Hepatic Targeted Delivery

Small interfering RNA (siRNA) is a clinically approved therapeutic modality, which has attracted widespread attention not only from basic research but also from pharmaceutical industry. As siRNA can theoretically modulate any disease-related gene's expression, plenty of siRNA therapeutic pipelines have been established by tens of biotechnology companies. The drug performance of siRNA heavily depends on the sequence, the chemical modification, and the delivery of siRNA. Here, we describe the rational design protocol of siRNA, and provide some modification patterns that can enhance siRNA's stability and reduce its off-target effect. Also, the delivery method based on N-acetylgalactosamine (GalNAc)-siRNA conjugate that is widely employed to develop therapeutic regimens for liver-related diseases is also recapitulated.

Using the Chou’s Pseudo Component to Predict the ncRNA Locations Based on the Improved K-Nearest Neighbor (iKNN) Classifier

Background:

The non-coding RNA identification at the organelle genome level is a challenging task. In our previous work, an ncRNA dataset with less than 80% sequence identity was built, and a method incorporating an increment of diversity combining with support vector machine method was proposed.

Objective:

Based on the ncRNA_361 dataset, a novel decision-making method-an improved KNN (iKNN) classifier was proposed.

Methods:

In this paper, based on the iKNN algorithm, the physicochemical features of nucleotides, the degeneracy of genetic codons, and topological secondary structure were selected to represent the effective ncRNA characters. Then, the incremental feature selection method was utilized to optimize the feature set.

Results:

The results of iKNN indicated that the decision-making method of mean value is distinctly superior to the traditional decision-making method of majority vote the Increment of Diversity Combining Support Vector Machine (ID-SVM). The iKNN algorithm achieved an overall accuracy of 97.368% in the jackknife test, when k=3.

Conclusion:

It should be noted that the triplets of the structure-sequence mode under reading frames not only contains the entire sequence information but also reflects whether the base was paired or not, and the secondary structural topological parameters further describe the ncRNA secondary structure on the spatial level. The ncRNA dataset and the iKNN classifier are freely available at http://202.207.14.87:8032/fuwu/iKNN/index.asp.

Xylosyltransferase-deficient human HEK293 cells show a strongly reduced proliferation capacity and viability

Human xylosyltransferases-I and -II (XT-I and XT-II) catalyze the initial and rate-limiting step in proteoglycan (PG)-biosynthesis. Because PG are major components of the extracellular matrix (ECM), an alternated XT expression is associated with the manifestation of ECM-related diseases. While Drosophila melanogaster and Caenorhabditis elegans only harbor one XT-isoform, all higher organisms contain two isoforms, which are expressed in a tissue-specific manner. The reason for the appearance of two isoenzymes remains unexplained and remarkable, as all other enzymes involved in the synthesis of the tetrasaccharid linker, which connects the PG core protein with attached glycosaminoglycans, only show one isoform. In human, mutations in the XYLT genes cause diseases affecting the homeostasis of the ECM, such as skeletal dysplasias. We investigated for the first time whether already XT-I-deficient human embryonic kidney (HEK293) cells can compensate for decreased expression levels of both XT-isoforms. A siRNA-mediated XYLT2 mRNA knockdown led to reduced cellular proliferation rates and a partially increased cellular senescence of treated HEK293 cells. These results were verified by conducting a stable CRISPR/Cas9-mediated XYLT2 knockout, which revealed that only cells expressing at least partially functional XT-II proteins remain proliferative. Our study, therefore, shows for the first time that cells lacking both XT-isoforms are not viable and clearly indicates the importance of the XT concerning the cellular metabolism.

RNA Interference and Small RNA Analysis

Small silencing RNAs have provided powerful reverse genetics tools and have opened new areas of research. This introduction describes the use of RNAi to suppress expression of individual genes for loss-of-function analysis. It also summarizes methods for measuring specific and global changes in small RNA expression, as well as methods to inhibit the function of individual endogenous small RNA species such as miRNAs.

Designing Tyrosinase siRNAs by Multiple Prediction Algorithms and Evaluation of Their Anti-Melanogenic Effects

Melanin is a pigment produced from tyrosine in melanocytes. Although melanin has a protective role against UVB radiation-induced damage, it is also associated with the development of melanoma and darker skin tone. Tyrosinase is a key enzyme in melanin synthesis, which regulates the rate-limiting step during conversion of tyrosine into DOPA and dopaquinone. To develop effective RNA interference therapeutics, we designed a melanin siRNA pool by applying multiple prediction programs to reduce human tyrosinase levels. First, 272 siRNAs passed the target accessibility evaluation using the RNAxs program. Then we selected 34 siRNA sequences with ΔG ≥-34.6 kcal/mol, i-Score value ≥65, and siRNA scales score ≤30. siRNAs were designed as 19-bp RNA duplexes with an asymmetric 3' overhang at the 3' end of the antisense strand. We tested if these siRNAs effectively reduced tyrosinase gene expression using qRT-PCR and found that 17 siRNA sequences were more effective than commercially available siRNA. Three siRNAs further tested showed an effective visual color change in MNT-1 human cells without cytotoxic effects, indicating these sequences are anti-melanogenic. Our study revealed that human tyrosinase siRNAs could be efficiently designed using multiple prediction algorithms.

In Silico Design and Experimental Validation of siRNAs Targeting Conserved Regions of Multiple Hepatitis C Virus Genotypes

RNA interference (RNAi) is a post-transcriptional gene silencing mechanism that mediates the sequence-specific degradation of targeted RNA and thus provides a tremendous opportunity for development of oligonucleotide-based drugs. Here, we report on the design and validation of small interfering RNAs (siRNAs) targeting highly conserved regions of the hepatitis C virus (HCV) genome. To aim for therapeutic applications by optimizing the RNAi efficacy and reducing potential side effects, we considered different factors such as target RNA variations, thermodynamics and accessibility of the siRNA and target RNA, and off-target effects. This aim was achieved using an in silico design and selection protocol complemented by an automated MysiRNA-Designer pipeline. The protocol included the design and filtration of siRNAs targeting highly conserved and accessible regions within the HCV internal ribosome entry site, and adjacent core sequences of the viral genome with high-ranking efficacy scores. Off-target analysis excluded siRNAs with potential binding to human mRNAs. Under this strict selection process, two siRNAs (HCV353 and HCV258) were selected based on their predicted high specificity and potency. These siRNAs were tested for antiviral efficacy in HCV genotype 1 and 2 replicon cell lines. Both in silico-designed siRNAs efficiently inhibited HCV RNA replication, even at low concentrations and for short exposure times (24h); they also exceeded the antiviral potencies of reference siRNAs targeting HCV. Furthermore, HCV353 and HCV258 siRNAs also inhibited replication of patient-derived HCV genotype 4 isolates in infected Huh-7 cells. Prolonged treatment of HCV replicon cells with HCV353 did not result in the appearance of escape mutant viruses. Taken together, these results reveal the accuracy and strength of our integrated siRNA design and selection protocols. These protocols could be used to design highly potent and specific RNAi-based therapeutic oligonucleotide interventions.

Soft computing model for optimized siRNA design by identifying off target possibilities using artificial neural network model

The ability of small interfering RNA (siRNA) to do posttranscriptional gene regulation by knocking down targeted genes is an important research topic in functional genomics, biomedical research and in cancer therapeutics. Many tools had been developed to design exogenous siRNA with high experimental inhibition. Even though considerable amount of work has been done in designing exogenous siRNA, design of effective siRNA sequences is still a challenging work because the target mRNAs must be selected such that their corresponding siRNAs are likely to be efficient against that target and unlikely to accidentally silence other transcripts due to sequence similarity. In some cases, siRNAs may tolerate mismatches with the target mRNA, but knockdown of genes other than the intended target could make serious consequences. Hence to design siRNAs, two important concepts must be considered: the ability in knocking down target genes and the off target possibility on any nontarget genes. So before doing gene silencing by siRNAs, it is essential to analyze their off target effects in addition to their inhibition efficacy against a particular target. Only a few methods have been developed by considering both efficacy and off target possibility of siRNA against a gene. In this paper we present a new design of neural network model with whole stacking energy (ΔG) that enables to identify the efficacy and off target effect of siRNAs against target genes. The tool lists all siRNAs against a particular target with their inhibition efficacy and number of matches or sequence similarity with other genes in the database. We could achieve an excellent performance of Pearson Correlation Coefficient (R=0. 74) and Area Under Curve (AUC=0.906) when the threshold of whole stacking energy is ≥-34.6 kcal/mol. To the best of the author's knowledge, this is one of the best score while considering the "combined efficacy and off target possibility" of siRNA for silencing a gene. The proposed model shall be useful for designing exogenous siRNA for therapeutic applications and gene silencing techniques in the area of bioinformatics. The software is developed as a desktop application and available at http://opsid.in/opsid/.

Improved asymmetry prediction for short interfering RNAs

In the development of RNA interference therapeutics, merely selecting short interfering RNA (siRNA) sequences that are complementary to the mRNA target does not guarantee target silencing. Current algorithms for selecting siRNAs rely on many parameters, one of which is asymmetry, often predicted through calculation of the relative thermodynamic stabilities of the two ends of the siRNA. However, we have previously shown that highly active siRNA sequences are likely to have particular nucleotides at each 5'-end, independently of their thermodynamic asymmetry. Here, we describe an algorithm for predicting highly active siRNA sequences based only on these two asymmetry parameters. The algorithm uses end-sequence nucleotide preferences and predicted thermodynamic stabilities, each weighted on the basis of training data from the literature, to rank the probability that an siRNA sequence will have high or low activity. The algorithm successfully predicts weakly and highly active sequences for enhanced green fluorescent protein and protein kinase R. Use of these two parameters in combination improves the prediction of siRNA activity over current approaches for predicting asymmetry. Going forward, we anticipate that this approach to siRNA asymmetry prediction will be incorporated into the next generation of siRNA selection algorithms.

Design of siRNA Therapeutics from the Molecular Scale

While protein-based therapeutics is well-established in the market, development of nucleic acid therapeutics has lagged. Short interfering RNAs (siRNAs) represent an exciting new direction for the pharmaceutical industry. These small, chemically synthesized RNAs can knock down the expression of target genes through the use of a native eukaryotic pathway called RNA interference (RNAi). Though siRNAs are routinely used in research studies of eukaryotic biological processes, transitioning the technology to the clinic has proven challenging. Early efforts to design an siRNA therapeutic have demonstrated the difficulties in generating a highly-active siRNA with good specificity and a delivery vehicle that can protect the siRNA as it is transported to a specific tissue. In this review article, we discuss design considerations for siRNA therapeutics, identifying criteria for choosing therapeutic targets, producing highly-active siRNA sequences, and designing an optimized delivery vehicle. Taken together, these design considerations provide logical guidelines for generating novel siRNA therapeutics.

The effect of regions flanking target site on siRNA potency

For a successful RNA interference (RNAi) experiment, selecting the small interference RNA (siRNA) candidates which maximize the knock down effect of the given gene is the critical step. Although various computational approaches have been attempted, the design of efficient siRNA candidates is far from satisfactory yet. In this study, we proposed a novel feature selection algorithm of combined random forest and support vector machine to predict active siRNAs. Using a publically available dataset, we demonstrated that the predictive accuracy would be markedly improved when the context sequence features outside the target site were included. The Pearson correlation coefficient for regression is as high as 0.721, compared to 0.671, 0.668, 0.680, and 0.645, for Biopredsi, i-score, ThermoComposition21 and DSIR, respectively. It revealed that siRNA-target interaction requires appropriate sequence context not only in the target site but also in a broad region flanking the target site.

What parameters to consider and which software tools to use for target selection and molecular design of small interfering RNAs

The design of small gene silencing RNAs with a high probability of being efficient still has some elements of an art, especially when the lowest concentration of small molecules needs to be utilized. The design of highly target-specific small interfering RNAs or short hairpin RNAs is even a greater challenging task. Some logical schemes and software tools that can be used for simplifying both tasks are presented here. In addition, sequence motifs and sequence composition biases of small interfering RNAs that have to be avoided because of specificity concerns are also detailed.

Methods for selecting effective siRNA target sequences using a variety of statistical and analytical techniques

Short interfering RNA (siRNA) has been widely used for studying gene function in mammalian cells but varies markedly in its gene silencing efficacy. Although many design rules/guidelines for effective siRNAs based on various criteria have been reported recently, there are only a few consistencies among them. This makes it difficult to select effective siRNA sequences in mammalian genes. This chapter first reviews the recently reported siRNA design guidelines and then proposes new methods for selecting effective siRNA sequences from many possible candidates by using decision tree learning, Bayes' theorem, and average silencing probability on the basis of a large number of known effective siRNAs. These methods differ from the previous score-based siRNA design techniques and can predict the probability that a candidate siRNA sequence will be effective. Evaluation of these methods by applying them to recently reported effective and ineffective siRNA sequences for a number of genes indicates that they would be useful for many other genes. They should, therefore, be of general utility for selecting effective siRNA sequences for mammalian genes. The chapter also describes another method using a hidden Markov model to select the optimal functional siRNAs and discusses the frequencies of combinations of two successive nucleotides as an important characteristic of effective siRNA sequences.

Engineering small interfering RNAs by strategic chemical modification

Synthetic small interfering RNAs (siRNAs) have revolutionized functional genomics in mammalian cell cultures due to their reliability, efficiency, and ease of use. This success, however, has not fully translated into siRNA applications in vivo and in siRNA therapeutics where initial optimism has been dampened by a lack of efficient delivery strategies and reports of siRNA off-target effects and immunogenicity. Encouragingly, most aspects of siRNA behavior can be addressed by careful engineering of siRNAs incorporating beneficial chemical modifications into discrete nucleotide positions during siRNA synthesis. Here, we review the literature (Subheadings 1 -3) and provide a quick guide (Subheading 4) to how the performance of siRNA can be improved by chemical modification to suit specific applications in vitro and in vivo.

Designing functional siRNA with reduced off-target effects

RNA interference (RNAi) mediated by small interfering RNA (siRNA) is now widely used to knock down gene expression in a sequence-specific manner, making it a powerful tool not only for studying gene functions but also for therapeutic applications. siRNA decreases the expression level of the intended target gene with complete complementarity by cleaving its mRNA. However, the efficacy of each siRNA widely varies depending on its sequence in mammalian cells; only a limited fraction of randomly designed siRNAs is functional. Moreover, off-target silencing effects arise when the siRNA has partial complementarity in the seed region with unintended genes. Here, we describe the rational designing of functional, off-target effect-reduced siRNAs using siDirect 2.0 Web server (http://siDirect2.RNAi.jp/). By using the default parameters, siDirect 2.0 can design at least one qualified siRNA for >94% of human mRNA sequences in the RefSeq database.

Reconsideration of in silico siRNA design from a perspective of heterogeneous data integration: problems and solutions

The success of RNA interference (RNAi) depends on the interaction between short interference RNAs (siRNAs) and mRNAs. Design of highly efficient and specific siRNAs has become a challenging issue in applications of RNAi. Here, we present a detailed survey on the state-of-the-art siRNAs design, focusing on several key issues with the current in silico RNAi studies, including: (i) inconsistencies among the proposed guidelines for siRNAs design and the incomplete list of siRNAs features, (ii) improper integration of the heterogeneous cross-platform siRNAs data, (iii) inadequate consideration of the binding specificity of the target mRNAs and (iv) reduction in the 'off-target' effect in siRNAs design. With these considerations, the popular in silico siRNAs design rules are reexamined and several inconsistent viewpoints toward siRNAs feature identifications are clarified. In addition, novel computational models for siRNAs design using state-of-art machine learning techniques are discussed, which focus on heterogeneous data integration, joint feature selection and customized siRNAs screening toward highly specific targets. We believe that addressing such issues in siRNA study will provide new clues for further improved design of more efficient and specific siRNAs in RNAi.

Mini-clusters with mean probabilities for identifying effective siRNAs

Background

The distinction between the effective siRNAs and the ineffective ones is in high demand for gene knockout technology. To design effective siRNAs, many approaches have been proposed. Those approaches attempt to classify the siRNAs into effective and ineffective classes but they are difficult to decide the boundary between these two classes.

Findings

Here, we try to split effective and ineffective siRNAs into many smaller subclasses by RMP-MiC(the relative mean probabilities of siRNAs with the mini-clusters algorithm). The relative mean probabilities of siRNAs are the modified arithmetic mean value of three probabilities, which come from three Markov chain of effective siRNAs. The mini-clusters algorithm is a modified version of micro-cluster algorithm.

Conclusions

When the RMP-MiC was applied to the experimental siRNAs, the result shows that all effective siRNAs can be identified correctly, and no more than 9% ineffective siRNAs are misidentified as effective ones. We observed that the efficiency of those misidentified ineffective siRNAs exceed 70%, which is very closed to the used efficiency threshold. From the analysis of the siRNAs data, we suggest that the mini-clusters algorithm with relative mean probabilities can provide new insights to the applications for distinguishing effective siRNAs from ineffective ones.

Optimized models for design of efficient miR30-based shRNAs

Small hairpin RNAs (shRNAs) became an important research tool in cell biology. Reliable design of these molecules is essential for the needs of large functional genomics projects. To optimize the design of efficient shRNAs, we performed comparative, thermodynamic, and correlation analyses of ~18,000 miR30-based shRNAs with known functional efficiencies, derived from the Sensor Assay project (Fellmann et al., 2011). We identified features of the shRNA guide strand that significantly correlate with the silencing efficiency and performed multiple regression analysis, using 4/5 of the data for training purposes and 1/5 for cross validation. A model that included the position-dependent nucleotide preferences was predictive in the cross-validation data subset (R = 0.39). However, a model, which in addition to the nucleotide preferences included thermodynamic shRNA features such as a thermodynamic duplex stability and position-dependent thermodynamic profile (dinucleotide free energy) was performing better (R = 0.53). Software "miR_Scan" was developed based upon the optimized models. Calculated mRNA target secondary structure stability showed correlation with shRNA silencing efficiency but failed to improve the model. Correlation analysis demonstrates that our algorithm for identification of efficient miR30-based shRNA molecules performs better than approaches that were developed for design of chemically synthesized siRNAs (R(max) = 0.36).

Development of Therapeutic-Grade Small Interfering RNAs by Chemical Engineering

Recent successes in clinical trials have provided important proof of concept that small interfering RNAs (siRNAs) indeed constitute a new promising class of therapeutics. Although great efforts are still needed to ensure efficient means of delivery in vivo, the siRNA molecule itself has been successfully engineered by chemical modification to meet initial challenges regarding specificity, stability, and immunogenicity. To date, a great wealth of siRNA architectures and types of chemical modification are available for promoting safe siRNA-mediated gene silencing in vivo and, consequently, the choice of design and modification types can be challenging to individual experimenters. Here we review the literature and devise how to improve siRNA performance by structural design and specific chemical modification to ensure potent and specific gene silencing without unwarranted side-effects and hereby complement the ongoing efforts to improve cell targeting and delivery by other carrier molecules.

siRNA Design Software for a Target Gene-Specific RNA Interference

RNA interference (RNAi) is a mechanism through which small interfering RNA (siRNA) induces sequence-specific posttranscriptional gene silencing. RNAi is commonly recognized as a powerful tool not only for functional genomics but also for therapeutic applications. Twenty-one-nucleotide-long siRNA suppresses the expression of the intended gene whose transcript possesses perfect complementarity to the siRNA guide strand. Hence, its silencing effect has been assumed to be extremely specific. However, accumulated evidences revealed that siRNA could downregulate unintended genes with partial complementarities mainly to the seven-nucleotide seed region of siRNA. This phenomenon is referred to as off-target effect. We have revealed that the capability to induce off-target effect is strongly correlated to the thermodynamic stability in siRNA seed-target duplex. For understanding accurate target gene function and successful therapeutic application, it may be critical to select a target gene-specific siRNA with minimized off-target effect. Here we present our siRNA design software for a target-specific RNAi. In addition, we also introduce the software programs open to the public for designing functional siRNAs.

MysiRNA-designer: a workflow for efficient siRNA design

The design of small interfering RNA (siRNA) is a multi factorial problem that has gained the attention of many researchers in the area of therapeutic and functional genomics. MysiRNA score was previously introduced that improves the correlation of siRNA activity prediction considering state of the art algorithms. In this paper, a new program, MysiRNA-Designer, is described which integrates several factors in an automated work-flow considering mRNA transcripts variations, siRNA and mRNA target accessibility, and both near-perfect and partial off-target matches. It also features the MysiRNA score, a highly ranked correlated siRNA efficacy prediction score for ranking the designed siRNAs, in addition to top scoring models Biopredsi, DISR, Thermocomposition21 and i-Score, and integrates them in a unique siRNA score-filtration technique. This multi-score filtration layer filters siRNA that passes the 90% thresholds calculated from experimental dataset features. MysiRNA-Designer takes an accession, finds conserved regions among its transcript space, finds accessible regions within the mRNA, designs all possible siRNAs for these regions, filters them based on multi-scores thresholds, and then performs SNP and off-target filtration. These strict selection criteria were tested against human genes in which at least one active siRNA was designed from 95.7% of total genes. In addition, when tested against an experimental dataset, MysiRNA-Designer was found capable of rejecting 98% of the false positive siRNAs, showing superiority over three state of the art siRNA design programs. MysiRNA is a freely accessible (Microsoft Windows based) desktop application that can be used to design siRNA with a high accuracy and specificity. We believe that MysiRNA-Designer has the potential to play an important role in this area.

Chemical modification of small interfering RNA

Chemically synthesized siRNAs are widely used for gene silencing. For in vitro applications, stability, delivery, and immunological issues are rarely problematic, but for in vivo applications the situation is different. Limited stability, undesirable pharmacokinetic behaviour, and unanticipated side effects from the immune system call for more careful structural siRNA design and inclusion of chemical modifications at selected positions. Also the notion that siRNA induces significant off-target silencing of many non-related genes has promted new effective measures to enhance specificity. The scope of this review is to provide a simple guide to successful chemical and structural modification of siRNAs with improved activity, stability, specificity, and low toxicity.

Predicting siRNA potency with random forests and support vector machines

Background

Short interfering RNAs (siRNAs) can be used to knockdown gene expression in functional genomics. For a target gene of interest, many siRNA molecules may be designed, whereas their efficiency of expression inhibition often varies.

Results

To facilitate gene functional studies, we have developed a new machine learning method to predict siRNA potency based on random forests and support vector machines. Since there were many potential sequence features, random forests were used to select the most relevant features affecting gene expression inhibition. Support vector machine classifiers were then constructed using the selected sequence features for predicting siRNA potency. Interestingly, gene expression inhibition is significantly affected by nucleotide dimer and trimer compositions of siRNA sequence.

Conclusions

The findings in this study should help design potent siRNAs for functional genomics, and might also provide further insights into the molecular mechanism of RNA interference.

Efficient prediction methods for selecting effective siRNA sequences

Although short interfering RNA (siRNA) has been widely used for studying gene functions in mammalian cells, its gene silencing efficacy varies markedly and there are only a few consistencies among the recently reported design rules/guidelines for selecting siRNA sequences effective for mammalian genes. Another shortcoming of the previously reported methods is that they cannot estimate the probability that a candidate sequence will silence the target gene. This paper first reviewed the recently reported siRNA design guidelines and clarified the problems concerning the guidelines. It then proposed two prediction methods-Radial Basis Function (RBF) network and decision tree learning-and their combined method for selecting effective siRNA target sequences from many possible candidate sequences. They are quite different from the previous score-based siRNA design techniques and can predict the probability that a candidate siRNA sequence will be effective. The methods imply high estimation accuracy for selecting candidate siRNA sequences.

Optimization of duplex stability and terminal asymmetry for shRNA design

Prediction of efficient oligonucleotides for RNA interference presents a serious challenge, especially for the development of genome-wide RNAi libraries which encounter difficulties and limitations due to ambiguities in the results and the requirement for significant computational resources. Here we present a fast and practical algorithm for shRNA design based on the thermodynamic parameters. In order to identify shRNA and siRNA features universally associated with high silencing efficiency, we analyzed structure-activity relationships in thousands of individual RNAi experiments from publicly available databases (ftp://ftp.ncbi.nlm.nih.gov/pub/shabalin/siRNA/si_shRNA_selector/). Using this statistical analysis, we found free energy ranges for the terminal duplex asymmetry and for fully paired duplex stability, such that shRNAs or siRNAs falling in both ranges have a high probability of being efficient. When combined, these two parameters yield a approximately 72% success rate on shRNAs from the siRecords database, with the target RNA levels reduced to below 20% of the control. Two other parameters correlate well with silencing efficiency: the stability of target RNA and the antisense strand secondary structure. Both parameters also correlate with the short RNA duplex stability; as a consequence, adding these parameters to our prediction scheme did not substantially improve classification accuracy. To test the validity of our predictions, we designed 83 shRNAs with optimal terminal asymmetry, and experimentally verified that small shifts in duplex stability strongly affected silencing efficiency. We showed that shRNAs with short fully paired stems could be successfully selected by optimizing only two parameters: terminal duplex asymmetry and duplex stability of the hypothetical cleavage product, which also relates to the specificity of mRNA target recognition. Our approach performs at the level of the best currently utilized algorithms that take into account prediction of the secondary structure of the target and antisense RNAs, but at significantly lower computational costs. Based on this study, we created the si-shRNA Selector program that predicts both highly efficient shRNAs and functional siRNAs (ftp://ftp.ncbi.nlm.nih.gov/pub/shabalin/siRNA/si_shRNA_selector/).

Selection of hyperfunctional siRNAs with improved potency and specificity

One critical step in RNA interference (RNAi) experiments is to design small interfering RNAs (siRNAs) that can greatly reduce the expression of the target transcripts, but not of other unintended targets. Although various statistical and computational approaches have been attempted, this remains a challenge facing RNAi researchers. Here, we present a new experimentally validated method for siRNA design. By analyzing public siRNA data and focusing on hyperfunctional siRNAs, we identified a set of sequence features as potency selection criteria to build an siRNA design algorithm with support vector machines. Additional bioinformatics filters were also included in the algorithm to increase RNAi specificity by reducing potential sequence cross-hybridization or microRNA-like effects. Independent validation experiments were performed, which indicated that the newly designed siRNAs have significantly improved performance, and worked effectively even at low concentrations. Furthermore, our cell-based studies demonstrated that the siRNA off-target effects were significantly reduced when the siRNAs were delivered into cells at the 3 nM concentration compared to 30 nM. Thus, the capability of our new design program to select highly potent siRNAs also renders increased RNAi specificity because these siRNAs can be used at a much lower concentration. The siRNA design web server is available at http://www5.appliedbiosystems.com/tools/siDesign/.

Designing optimal siRNA based on target site accessibility

RNA interference, mediated by small interfering RNAs (siRNAs), is a powerful tool for investigation of gene functions and it is increasingly being used as a therapeutic agent. However, not all siRNAs are equally potent - although simple rules for the selection of good siRNAs were proposed early on, siRNAs are still plagued with widely fluctuating efficiency. Recently, new design tools that incorporate both the structural features of the targeted RNAs and the sequence features of the siRNAs have substantially improved the efficacy of siRNAs. In this chapter, we present the algorithms behind these accessibility-aided tools and show how to design efficient siRNAs with their help.

A potential therapeutic for pandemic influenza using RNA interference

RNA interference (RNAi) involves sequence-specific downregulation of target genes, leading to gene silencing in vitro and in vivo. Synthetic small interfering RNAs (siRNAs), formulated with appropriate delivery agents, can serve as effective tools for RNAi-based therapeutics. The potential of siRNA to provide antiviral activity has been studied extensively in many respiratory viruses, including influenza virus, wherein specific siRNAs target highly-conserved regions of influenza viral genome, leading to potent inhibition of viral RNA replication. Despite various delivery strategies, such as polycations and liposomes that have been employed to formulate siRNAs, effective delivery modalities are still needed. Although current strategies can provide significant biodistribution and delivery into lungs allowing gene silencing, complete protection and prolonged survival rates against multiple strains of influenza virus still remains a key challenge. Here, we describe methods and procedures pertaining to screening and selection of highly effective influenza-specific siRNAs in cell culture, in mice, and in the ferret model. This will be potentially useful to evaluate RNAi as a therapeutic modality for future clinical application.

RNA interference inhibits hepatitis E virus mRNA accumulation and protein synthesis in vitro

Hepatitis E virus (HEV) is a zoonotic pathogen to which several species, including human beings, pigs and rodents, are reported to be susceptible. To date, vaccines developed against HEV still need to be improved and a structural gene (ORF2), which encodes a capsid protein with high sequence conservation found across HEV genotypes, is a potential candidate. To exploit the possibility of using RNA interference (RNAi) as a strategy against HEV infection, four small interference RNA (siRNA) duplex targeting ORF2 gene were constructed. A challenge against HEV infection by RNAi was performed in A549 cells. Real-Time quantitative polymerase chain reaction (Real-Time qPCR) and Western blot assay demonstrated that four HEV specific siRNAs (si-ORF2-1, si-ORF2-2, si-ORF2-3 and si-ORF2-4) were capable of protecting cells against HEV infection with very high specificity and efficiency. The results suggest that RNAi is a potent anti-HEV infection prophylaxis strategy.

Comparing artificial neural networks, general linear models and support vector machines in building predictive models for small interfering RNAs

Background

Exogenous short interfering RNAs (siRNAs) induce a gene knockdown effect in cells by interacting with naturally occurring RNA processing machinery. However not all siRNAs induce this effect equally. Several heterogeneous kinds of machine learning techniques and feature sets have been applied to modeling siRNAs and their abilities to induce knockdown. There is some growing agreement to which techniques produce maximally predictive models and yet there is little consensus for methods to compare among predictive models. Also, there are few comparative studies that address what the effect of choosing learning technique, feature set or cross validation approach has on finding and discriminating among predictive models.

Principal Findings

Three learning techniques were used to develop predictive models for effective siRNA sequences including Artificial Neural Networks (ANNs), General Linear Models (GLMs) and Support Vector Machines (SVMs). Five feature mapping methods were also used to generate models of siRNA activities. The 2 factors of learning technique and feature mapping were evaluated by complete 3×5 factorial ANOVA. Overall, both learning techniques and feature mapping contributed significantly to the observed variance in predictive models, but to differing degrees for precision and accuracy as well as across different kinds and levels of model cross-validation.

Conclusions

The methods presented here provide a robust statistical framework to compare among models developed under distinct learning techniques and feature sets for siRNAs. Further comparisons among current or future modeling approaches should apply these or other suitable statistically equivalent methods to critically evaluate the performance of proposed models. ANN and GLM techniques tend to be more sensitive to the inclusion of noisy features, but the SVM technique is more robust under large numbers of features for measures of model precision and accuracy. Features found to result in maximally predictive models are not consistent across learning techniques, suggesting care should be taken in the interpretation of feature relevance. In the models developed here, there are statistically differentiable combinations of learning techniques and feature mapping methods where the SVM technique under a specific combination of features significantly outperforms all the best combinations of features within the ANN and GLM techniques.

Energy profile and secondary structure impact shRNA efficacy

Background

RNA interference (RNAi) is a cellular mechanism in which a short/small double stranded RNA induces the degradation of its sequence specific target mRNA, leading to specific gene silencing. Since its discovery, RNAi has become a powerful biological technique for gene function studies and drug discovery. The very first requirement of applying RNAi is to design functional small interfering RNA (siRNA) that can uniquely induce the degradation of the targeted mRNA. It has been shown that many functional synthetic siRNAs share some common characteristics, such as GC content limitation and free energy preferences at both terminals, etc.

Results

Our three-phase algorithm was developed to design siRNA on a whole-genome scale based on those identified characteristics of functional siRNA. When this algorithm was applied to design short hairpin RNA (shRNA), the validated success rate of shRNAs was over 70%, which was almost double the rate reported for TRC library. This indicates that the designs of siRNA and shRNA may share the same concerns. Further analysis of the shRNA dataset of 444 designs reveals that the high free energy states of the two terminals have the largest positive impact on the shRNA efficacy. Enforcing these energy characteristics of both terminals can further improve the shRNA design success rate to 83.1%. We also found that functional shRNAs have less probability for their 3' terminals to be involved in mRNA secondary structure formation.

Conclusion

Functional shRNAs prefer high free energy states at both terminals. High free energy states of the two terminals were found to be the largest positive impact factor on shRNA efficacy. In addition, the accessibility of the 3' terminal is another key factor to shRNA efficacy.

Approximate Bayesian feature selection on a large meta-dataset offers novel insights on factors that effect siRNA potency

MOTIVATION

Short interfering RNA (siRNA)-induced RNA interference is an endogenous pathway in sequence-specific gene silencing. The potency of different siRNAs to inhibit a common target varies greatly and features affecting inhibition are of high current interest. The limited success in predicting siRNA potency being reported so far could originate in the small number and the heterogeneity of available datasets in addition to the knowledge-driven, empirical basis on which features thought to be affecting siRNA potency are often chosen. We attempt to overcome these problems by first constructing a meta-dataset of 6483 publicly available siRNAs (targeting mammalian mRNA), the largest to date, and then applying a Bayesian analysis which accommodates feature set uncertainty. A stochastic logistic regression-based algorithm is designed to explore a vast model space of 497 compositional, structural and thermodynamic features, identifying associations with siRNA potency.

RESULTS

Our algorithm reveals a number of features associated with siRNA potency that are, to the best of our knowledge, either under reported in literature, such as anti-sense 5' -3' motif 'UCU', or not reported at all, such as the anti-sense 5' -3' motif 'ACGA'. These findings should aid in improving future siRNA potency predictions and might offer further insights into the working of the RNA-induced silencing complex (RISC).

Methods for selecting effective siRNA sequences by using statistical and clustering techniques

Short interfering RNAs (siRNAs) have been widely used for studying gene functions in mammalian cells but vary markedly in their gene-silencing efficacy. Although many design rules/guidelines for effective siRNAs based on various criteria have been reported recently, there are only a few consistencies among them. This makes it difficult to select effective siRNA sequences targeting mammalian genes. This chapter first reviews the reported siRNA design guidelines and clarifies the problems concerning the current guidelines. It then describes the recently reported new scoring methods for selecting effective siRNA sequences by using statistics and clustering techniques such as the self-organizing map (SOM) and the radial basis function (RBF) network. In the proposed three methods, individual scores are defined as a gene degradation measure based on position-specific statistical significances. The effectiveness of the methods was confirmed by evaluating effective and ineffective siRNAs for recently reported genes and comparison with other reported scoring methods. The sizes (values) of these scores are closely correlated with the degree of gene degradation, and the scores can easily be used for selecting high-potential siRNA candidates. The evaluation results indicate that the proposed new methods are useful for selecting siRNA sequences targeting mammalian mRNA sequences.

OligoWalk: an online siRNA design tool utilizing hybridization thermodynamics

Given an mRNA sequence as input, the OligoWalk web server generates a list of small interfering RNA (siRNA) candidate sequences, ranked by the probability of being efficient siRNA (silencing efficacy greater than 70%). To accomplish this, the server predicts the free energy changes of the hybridization of an siRNA to a target mRNA, considering both siRNA and mRNA self-structure. The free energy changes of the structures are rigorously calculated using a partition function calculation. By changing advanced options, the free energy changes can also be calculated using less rigorous lowest free energy structure or suboptimal structure prediction methods for the purpose of comparison. Considering the predicted free energy changes and local siRNA sequence features, the server selects efficient siRNA with high accuracy using a support vector machine. On average, the fraction of efficient siRNAs selected by the server that will be efficient at silencing is 78.6%. The OligoWalk web server is freely accessible through internet at http://rna.urmc.rochester.edu/servers/oligowalk.

Fundamental differences in the equilibrium considerations for siRNA and antisense oligodeoxynucleotide design

Both siRNA and antisense oligodeoxynucleotides (ODNs) inhibit the expression of a complementary gene. In this study, fundamental differences in the considerations for RNA interference and antisense ODNs are reported. In siRNA and antisense ODN databases, positive correlations are observed between the cost to open the mRNA target self-structure and the stability of the duplex to be formed, meaning the sites along the mRNA target with highest potential to form strong duplexes with antisense strands also have the greatest tendency to be involved in pre-existing structure. Efficient siRNA have less stable siRNA-target duplex stability than inefficient siRNA, but the opposite is true for antisense ODNs. It is, therefore, more difficult to avoid target self-structure in antisense ODN design. Self-structure stabilities of oligonucleotide and target correlate to the silencing efficacy of siRNA. Oligonucleotide self-structure correlations to efficacy of antisense ODNs, conversely, are insignificant. Furthermore, self-structure in the target appears to correlate with antisense ODN efficacy, but such that more effective antisense ODNs appear to target mRNA regions with greater self-structure. Therefore, different criteria are suggested for the design of efficient siRNA and antisense ODNs and the design of antisense ODNs is more challenging.

The impact of target site accessibility on the design of effective siRNAs

Small-interfering RNAs (siRNAs) assemble into RISC, the RNA-induced silencing complex, which cleaves complementary mRNAs. Despite their fluctuating efficacy, siRNAs are widely used to assess gene function. Although this limitation could be ascribed, in part, to variations in the assembly and activation of RISC, downstream events in the RNA interference (RNAi) pathway, such as target site accessibility, have so far not been investigated extensively. In this study we present a comprehensive analysis of target RNA structure effects on RNAi by computing the accessibility of the target site for interaction with the siRNA. Based on our observations, we developed a novel siRNA design tool, RNAxs, by combining known siRNA functionality criteria with target site accessibility. We calibrated our method on two data sets comprising 573 siRNAs for 38 genes, and tested it on an independent set of 360 siRNAs targeting four additional genes. Overall, RNAxs proves to be a robust siRNA selection tool that substantially improves the prediction of highly efficient siRNAs.

Efficient siRNA selection using hybridization thermodynamics

Small interfering RNA (siRNA) are widely used to infer gene function. Here, insights in the equilibrium of siRNA-target hybridization are used for selection of efficient siRNA. The accessibilities of siRNA and target mRNA for hybridization, as measured by folding free energy change, are shown to be significantly correlated with efficacy. For this study, a partition function calculation that considers all possible secondary structures is used to predict target site accessibility; a significant improvement over calculations that consider only the predicted lowest free energy structure or a set of low free energy structures. The predicted thermodynamic features, in addition to siRNA sequence features, are used as input for a support vector machine that selects functional siRNA. The method works well for predicting efficient siRNA (efficacy >70%) in a large siRNA data set from Novartis. The positive predictive value (percentage of sites predicted to be efficient for silencing that are) is as high as 87.6%. The sensitivity and specificity are 22.7 and 96.5%, respectively. When tested on data from different sources, the positive predictive value increased 8.1% by adding equilibrium terms to 25 local sequence features. Prediction of hybridization affinity using partition functions is now available in the RNAstructure software package.

A protocol for designing siRNAs with high functionality and specificity

Effective gene silencing by the RNA interference (RNAi) pathway requires a comprehensive understanding of the elements that influence small interfering RNA (siRNA) functionality and specificity. These include (i) sequence space restrictions that define the boundaries of siRNA targeting, (ii) structural and sequence features required for efficient siRNA performance, (iii) mechanisms that underlie nonspecific gene modulation and (iv) additional features specific to the intended use (i.e., inclusion of native sugar or base chemical modifications for increased stability or specificity, vector design, etc.). Attention to each of these factors enhances siRNA performance and heightens overall confidence in the output of RNAi-mediated functional genomic studies. Here, we provide a detailed protocol explaining the methodologies used for manual and web-based design of siRNAs.

Thermodynamic instability of siRNA duplex is a prerequisite for dependable prediction of siRNA activities

We developed a simple algorithm, i-Score (inhibitory-Score), to predict active siRNAs by applying a linear regression model to 2431 siRNAs. Our algorithm is exclusively comprised of nucleotide (nt) preferences at each position, and no other parameters are taken into account. Using a validation dataset comprised of 419 siRNAs, we found that the prediction accuracy of i-Score is as good as those of s-Biopredsi, ThermoComposition21 and DSIR, which employ a neural network model or more parameters in a linear regression model. Reynolds and Katoh also predict active siRNAs efficiently, but the numbers of siRNAs predicted to be active are less than one-eighth of that of i-Score. We additionally found that exclusion of thermostable siRNAs, whose whole stacking energy (ΔG) is less than −34.6 kcal/mol, improves the prediction accuracy in i-Score, s-Biopredsi, ThermoComposition21 and DSIR. We also developed a universal target vector, pSELL, with which we can assay an siRNA activity of any sequence in either the sense or antisense direction. We assayed 86 siRNAs in HEK293 cells using pSELL, and validated applicability of i-Score and the whole ΔG value in designing siRNAs.

Platination of the siRNA sense-strand modulates both efficacy and selectivity in vitro

The use of short interfering RNAs (siRNA) for selective suppression of protein production has rapidly become a commonly used technique for transient modulation of protein levels. In the present paper, we investigate whether introduction of platinated bases in the sense strand can be used to modulate the efficacy of siRNAs. Four different siRNAs were studied, all targeting the initial AU-rich 3' UTR of Wnt-5a mRNA. The siRNAs were characterized with respect to melting properties and translational inhibitory effect in vitro using luciferase as a reporter gene. The translation inhibition studies reveal that all platinated siRNA remain efficient. For an siRNA with partial complementarity to the luciferase gene, platination was shown to reduce the off-target effects. All siRNAs were found to be active in cellular in vitro translation systems, reaching suppression levels well above 80% for the majority of siRNAs investigated.

RFRCDB-siRNA: improved design of siRNAs by random forest regression model coupled with database searching

Although the observations concerning the factors which influence the siRNA efficacy give clues to the mechanism of RNAi, the quantitative prediction of the siRNA efficacy is still a challenge task. In this paper, we introduced a novel non-linear regression method: random forest regression (RFR), to quantitatively estimate siRNAs efficacy values. Compared with an alternative machine learning regression algorithm, support vector machine regression (SVR) and four other score-based algorithms [A. Reynolds, D. Leake, Q. Boese, S. Scaringe, W.S. Marshall, A. Khvorova, Rational siRNA design for RNA interference, Nat. Biotechnol. 22 (2004) 326-330; K. Ui-Tei, Y. Naito, F. Takahashi, T. Haraguchi, H. Ohki-Hamazaki, A. Juni, R. Ueda, K. Saigo, Guidelines for the selection of highly effective siRNA sequences for mammalian and chick RNA interference, Nucleic Acids Res. 32 (2004) 936-948; A.C. Hsieh, R. Bo, J. Manola, F. Vazquez, O. Bare, A. Khvorova, S. Scaringe, W.R. Sellers, A library of siRNA duplexes targeting the phosphoinositide 3-kinase pathway: determinants of gene silencing for use in cell-based screens, Nucleic Acids Res. 32 (2004) 893-901; M. Amarzguioui, H. Prydz, An algorithm for selection of functional siRNA sequences, Biochem. Biophys. Res. Commun. 316 (2004) 1050-1058) our RFR model achieved the best performance of all. A web-server, RFRCDB-siRNA (http://www.bioinf.seu.edu.cn/siRNA/index.htm), has been developed. RFRCDB-siRNA consists of two modules: a siRNA-centric database and a RFR prediction system. RFRCDB-siRNA works as follows: (1) Instead of directly predicting the gene silencing activity of siRNAs, the service takes these siRNAs as queries to search against the siRNA-centric database. The matched sequences with the exceeding the user defined functionality value threshold are kept. (2) The mismatched sequences are then processed into the RFR prediction system for further analysis.

Improving model predictions for RNA interference activities that use support vector machine regression by combining and filtering features

Background

RNA interference (RNAi) is a naturally occurring phenomenon that results in the suppression of a target RNA sequence utilizing a variety of possible methods and pathways. To dissect the factors that result in effective siRNA sequences a regression kernel Support Vector Machine (SVM) approach was used to quantitatively model RNA interference activities.

Results

Eight overall feature mapping methods were compared in their abilities to build SVM regression models that predict published siRNA activities. The primary factors in predictive SVM models are position specific nucleotide compositions. The secondary factors are position independent sequence motifs (N-grams) and guide strand to passenger strand sequence thermodynamics. Finally, the factors that are least contributory but are still predictive of efficacy are measures of intramolecular guide strand secondary structure and target strand secondary structure. Of these, the site of the 5' most base of the guide strand is the most informative.

Conclusion

The capacity of specific feature mapping methods and their ability to build predictive models of RNAi activity suggests a relative biological importance of these features. Some feature mapping methods are more informative in building predictive models and overall t-test filtering provides a method to remove some noisy features or make comparisons among datasets. Together, these features can yield predictive SVM regression models with increased predictive accuracy between predicted and observed activities both within datasets by cross validation, and between independently collected RNAi activity datasets. Feature filtering to remove features should be approached carefully in that it is possible to reduce feature set size without substantially reducing predictive models, but the features retained in the candidate models become increasingly distinct. Software to perform feature prediction and SVM training and testing on nucleic acid sequences can be found at the following site: .