Secondary Structural Features Research Articles

Рt/SAPO-11 Catalytic Systems Differing in Acidity and Secondary Pore Structure in n-Hexadecane Hydroisomerization

SAPO-11 molecular sieve samples differing in the acid properties, crystal morphology and size, and secondary pore structure characteristics were obtained by crystallization of reaction gels with the SiO2/Al2O3 molar ratios of 0.1 and 0.3, prepared using aluminum isopropoxide or boehmite as an aluminum source. Platinum (0.5 wt %) was deposited onto the molecular sieves prepared, and the catalytic properties of the resulting samples in n-hexadecane hydroisomerization were studied. The hydrocarbon conversion on these samples varies from 76.8 to 87.7 wt %, and the isoparaffin formation selectivity, from 76.7 to 91.2 wt %.

SSDraw: Software for generating comparative protein secondary structure diagrams.

The program SSDraw generates publication-quality protein secondary structure diagrams from three-dimensional protein structures. To depict relationships between secondary structure and other protein features, diagrams can be colored by conservation score, B-factor, or custom scoring. Diagrams of homologous proteins can be registered according to an input multiple sequence alignment. Linear visualization allows the user to stack registered diagrams, facilitating comparison of secondary structure and other properties among homologous proteins. SSDraw can be used to compare secondary structures of homologous proteins with both conserved and divergent folds. It can also generate one secondary structure diagram from an input protein structure of interest. The source code can be downloaded (https://github.com/ncbi/SSDraw) and run locally for rapid structure generation, while a Google Colab notebook allows easy use.

DeepRaccess: high-speed RNA accessibility prediction using deep learning.

RNA accessibility is a useful RNA secondary structural feature for predicting RNA-RNA interactions and translation efficiency in prokaryotes. However, conventional accessibility calculation tools, such as Raccess, are computationally expensive and require considerable computational time to perform transcriptome-scale analysis. In this study, we developed DeepRaccess, which predicts RNA accessibility based on deep learning methods. DeepRaccess was trained to take artificial RNA sequences as input and to predict the accessibility of these sequences as calculated by Raccess. Simulation and empirical dataset analyses showed that the accessibility predicted by DeepRaccess was highly correlated with the accessibility calculated by Raccess. In addition, we confirmed that DeepRaccess could predict protein abundance in E.coli with moderate accuracy from the sequences around the start codon. We also demonstrated that DeepRaccess achieved tens to hundreds of times software speed-up in a GPU environment. The source codes and the trained models of DeepRaccess are freely available at https://github.com/hmdlab/DeepRaccess.

A fusion framework of deep learning and machine learning for predicting sgRNA cleavage efficiency

CRISPR/Cas9 system is a powerful tool for genome editing. Numerous studies have shown that sgRNAs can strongly affect the efficiency of editing. However, it is still not clear what rules should be followed for designing sgRNA with high cleavage efficiency. At present, several machine learning or deep learning methods have been developed to predict the cleavage efficiency of sgRNAs, however, the prediction accuracy of these tools is still not satisfactory. Here we propose a fusion framework of deep learning and machine learning, which first deals with the primary sequence and secondary structure features of the sgRNAs using both convolutional neural network (CNN) and recurrent neural network (RNN), and then uses the features extracted by the deep neural network to train a conventional machine learning model with LGBM. As a result, the new approach overwhelmed previous methods. The Spearman's correlation coefficient between predicted and measured sgRNA cleavage efficiency of our model (0.917) is improved by over 5% compared with the most advanced method (0.865), and the mean square error reduces from 7.89 × 10−3 to 4.75 × 10−3. Finally, we developed an online tool, CRISep (http://www.cuilab.cn/CRISep), to evaluate the availability of sgRNAs based on our models.

HDV RNA assays: Performance characteristics, clinical utility, and challenges.

Coinfection with HBV and HDV results in hepatitis D, the most severe form of chronic viral hepatitis, frequently leading to liver decompensation and HCC. Pegylated interferon alpha, the only treatment option for chronic hepatitis D for many years, has limited efficacy. New treatments are in advanced clinical development, with one recent approval. Diagnosis and antiviral treatment response monitoring are based on detection and quantification of HDV RNA. However, the development of reliable HDV RNA assays is challenged by viral heterogeneity (at least 8 different genotypes and several subgenotypes), intrahost viral diversity, rapid viral evolution, and distinct secondary structure features of HDV RNA. Different RNA extraction methodologies, primer/probe design for nucleic acid tests, lack of automation, and overall dearth of standardization across testing laboratories contribute to substantial variability in performance characteristics of research-based and commercial HDV RNA assays. A World Health Organization (WHO) standard for HDV RNA, available for about 10 years, has been used by many laboratories to determine the limit of detection of their assays and facilitates comparisons of RNA levels across study centers. Here we review challenges for robust pan genotype HDV RNA quantification, discuss particular clinical needs and the importance of reliable HDV RNA quantification in the context of drug development and patient monitoring. We summarize distinct technical features and performance characteristics of available HDV RNA assays. Finally, we provide considerations for the use of HDV RNA assays in the context of drug development and patient monitoring.

Protein structural insights into a rare PCSK9 gain-of-function variant (R496W) causing familial hypercholesterolemia in a Saudi family: whole exome sequencing and computational analysis.

Familial hypercholesterolemia (FH) is a globally underdiagnosed genetic condition associated with premature cardiovascular death. The genetic etiology data on Arab FH patients is scarce. Therefore, this study aimed to identify the genetic basis of FH in a Saudi family using whole exome sequencing (WES) and multidimensional bioinformatic analysis. Our WES findings revealed a rare heterozygous gain-of-function variant (R496W) in the exon 9 of the PCSK9 gene as a causal factor for FH in this family. This variant was absent in healthy relatives of the proband and 200 healthy normolipidemic controls from Saudi Arabia. Furthermore, this variant has not been previously reported in various regional and global population genomic variant databases. Interestingly, this variant is classified as "likely pathogenic" (PP5) based on the variant interpretation guidelines of the American College of Medical Genetics (ACMG). Computational functional characterization suggested that this variant could destabilize the native PCSK9 protein and alter its secondary and tertiary structural features. In addition, this variant was predicted to negatively influence its ligand-binding ability with LDLR and Alirocumab antibody molecules. This rare PCSK9 (R496W) variant is likely to expand our understanding of the genetic basis of FH in Saudi Arabia. This study also provides computational structural insights into the genotype-protein phenotype relationship of PCSK9 pathogenic variants and contributes to the development of personalized medicine for FH patients in the future.

Structural and functional relationships of plant allergenic proteins: gastrointestinal metabolism and structural heterogeneity

The prevalence of food allergies has increased in recent decades and affects up to 5% of the population, frequently resulting in severe symptoms. Specific structures of the corresponding proteins, the epitopes, are recognized by T‐cell receptors or IgE antibodies, resulting in an immune response. Many food allergens and immunogenic proteins for various plant‐based foods have been investigated thoroughly on different levels. However, research into the gastrointestinal metabolism of plant‐based food allergens on the molecular level is scarce, so far. The aim of this thesis was to characterize the gastrointestinal metabolism of food allergens and proteins involved in celiac disease, and to derive global properties that account for the immunogenic and allergenic potential of these proteins. Raw walnut, raw and cooked soybeans, and wheat‐based bread have been chosen as representatives for tree nuts, legumes and cereals, in order to adress the given research question. The project is divided into three sub‐steps, which are as follows: Adaption of the in vitro digestion model to different matrices and comprehensive analysis of proteins and peptides Development of a data‐processing strategy to enable visualisation of the resulting data Interpretation of the data regarding the gastrointestinal metabolism of relevant proteins in different matrices The in vitro digestion model by Minekus et al. was successfully adapted to simulate gastrointestinal digestion of a variety of food matrices. The analytical approach at protein level by SDS‐PAGE analysis and in‐gel digestion, combined with mass spectrometry at peptide level, enabled comprehensive identification of degradation products. As an example, up to 7279 peptides were assigned to soy proteins over the course of gastral digestion of raw soy, and a further 1027 peptides for samples of simulated intestinal digestion. In order to deal with these huge amounts of data, various data processing steps have been designed. Peptide identification by PEAKS® was extended using scripts in Python, to perform label‐free quantification of mass spectrometry data and to compare the identified degradation products with known epitopes. The multistage preprocessing approach developed in Python provided the basis for a set of data visualisations to further interpret the behaviour of proteins during gastrointestinal digestion. Analysis at protein and peptide levels revealed that quantities of most of the analysed proteins for all matrices exhibited stability against gastrointestinal digestion. Analysis of the resulting sequence coverage over the course of gastrointestinal digestion, combined with label‐free quantification in 3D surface plots, allowed conclusions on the behaviour of distinct protein regions over time. This approach provided detailed insights into the multistage formation and degradation mechanisms of differently sized peptides. ‐gliadin‐assigned peptides covering epitopes related to food allergy, as well as eight peptides covering celiac disease related epitopes, was characterized by label‐free quantification, outlining enhanced resistance against gastrointestinal digestion. Bioinformatic tools were applied to provide further characteristics of all proteins of interest related to the amino acid sequence. Sequence alignment with secondary structure features. The given results of this thesis provide insight into the stability of the proteins involved in adverse reactions to food, such as allergy or celiac disease, throughout the process of gastrointestinal digestion.

Improvement of functional characteristics of Hypophthalmichthys molitrix protein by modification with chitosan oligosaccharide.

In the food processing field, it is very often that fish proteins are denatured affecting the nutritional value of the product which is vital to be solved. By using appropriate sugar donors for glycosylation with protein, improving the stability and emulsification properties of fish proteins can be achieved. This research looks into the impacts of enzymatic chitosan oligosaccharide (CO) at various concentration (0.15%, 0.30%, 0.45%, 0.60%, w/v) upon the molecular makeup and function of silver carp myofibrillar protein (MP) in an attempt to comprehend the impact of electrostatic binding among MP as well as CO on protein conformation. Analysis was done on the impact of various CO concentrations upon MP's secondary structure, conformational changes, and functional characteristics. Twelve sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) assays were implemented to monitor MP; Fourier transform infrared spectroscopy, endogenous fluorescence spectroscopy, and UV absorption spectra were carried out to investigate the influence of CO on MP; Particle size distribution, emulsifying activity index (EAI), solubility, turbidity, sulfhydryl content, carbonyl content, foaming capacity, surface hydrophobicity, emulsifying stability index (ESI), and foam persistence were all investigated. In addition, we used dynamic light scattering, scanning electron microscope, and atomic force microscope to analyze myosin (MO) and 0.60% CO-MO complex. The results demonstrated that CO and MP form complexes through hydrogen bonding and electrostatic interactions. CO modification not only delayed the oxidation of MP but also promoted MP to show better solubility, foaming, and foaming stability. In addition, CO modified myosin particle size decreased, reducing myosin's roughness and making myosin's structure more compact. To sum up, molecular interaction could change functional characteristics, and products with special properties could be developed after modification with chitosan oligosaccharide.

Identifying Epitope Sites on the Receptor Binding Domain (RBD) of SARS-CoV-2

Abstract COVID-19 and SARS-CoV-2 variants continue to threaten human health and life worldwide. Thousands of structures related to SARS-CoV-2 have been rapidly determined, either by X-ray crystallography or CryoEM and deposited in the Protein Data Bank (PDB) since 2020. Here, we systematically investigated the structures of 302 antibodies and 78 nanobodies in complex with spike protein or the receptor binding domain (RBD) of SARS-CoV-2. We identified 23 common epitopic sites (ES) on the RBD surface and revealed the vital role of the complementarity-determining Region (CDR) loops in recognizing epitopes. The 23 ES are characterized according to the secondary structure feature and accessible surface area. About 75% of the total surface of RBD areas could access by the elicited antibodies, while the CDR3 loops occupied 50% of the contact surface. The analysis of paratope-epitope (antibody-antigen) interaction based on these epitope sites revealed the features of potent neutralizing the virus and the unique usage of amino acids of antibody. The clustering analysis of the antibody biophysical properties and surface area availabilities on the RBD found many binding motifs. Remarkably, most variants of concern (VOC) escape mutations, including Omicron occur within these 23 ES. This analysis not only explains the differential ability of antibodies to recognize epitope sites on the RBD surface but also offers predictive guidelines for understanding the role that accumulated SARS mutations might play in escape from antibodies elicited by immunogens. This structural characterization of the epitope-paratope interactions provides insights for structure-based vaccine design and therapeutic strategies and drugs against the future virus.

Lutein fortification of wheat bread with marigold powder: Impact on rheology, water dynamics, and structure.

Demands for foods conducive to eye health have been on the increase in the global healthcare sector. Marigold powder as a major source of lutein was utilized to produce lutein-fortified breads for ocular health. The physicochemical characteristics of the doughs and breads were investigated in terms of rheology, water mobility, and protein secondary structures. The incorporation of marigold powder decreased the water absorption of doughs without significantly altering thermo-mechanical properties. With a range of fortification levels (1 ~ 3%), marigold powder led to decreased storage and loss modulus of doughs by weakening their gluten network, which was supported by their T2 relaxation times. The resistances of the doughs weakened with increasing levels of marigold powder, while their extensibilities significantly incremented. FTIR spectral deconvolution revealed the changes of wheat protein structures upon marigold powder incorporation, in which the proportion of β-turn increased at the expense of β-sheet ratio. The breads with marigold powder displayed increased specific volume from 4.034 to 4.368 mL/g, accompanied by softer textures. The baking process led to heat-induced losses in lutein concentration of less than 10% within the crumbs and approximately 30% in the crusts. The use of marigold powder induced changes in protein secondary structure and extensional features of doughs, contributing to increased loaf volume and softer texture. Overall, this study provides fundamental information about the rheological and structural effects of marigold powder in a wheat bread system, consequently encouraging the food industry to utilize marigold power as a functional food ingredient. This article is protected by copyright. All rights reserved.

E-RNA: a collection of web-servers for the prediction and visualisation of RNA secondary structure and their functional features.

e-RNA is a collection of web-servers for the prediction and visualisation of RNA secondary structures and their functional features, including in particular RNA-RNA interactions. In this updated version, we have added novel tools for RNA secondary structure prediction and have significantly updated the visualisation functionality. The new method CoBold can identify transient RNA structure features and their potential functional effects on a known RNA structure during co-transcriptional structure formation. New tool ShapeSorter can predict evolutionarily conserved RNA secondary structure features while simultaneously taking experimental SHAPE probing evidence into account. The web-server R-Chie which visualises RNA secondary structure information in terms of arc diagrams, can now be used to also visualise and intuitively compare RNA-RNA, RNA-DNA and DNA-DNA interactions alongside multiple sequence alignments and quantitative information. The prediction generated by any method in e-RNA can be readily visualised on the web-server. For completed tasks, users can download their results and readily visualise them later on with R-Chie without having to re-run the predictions. e-RNA can be found at http://www.e-rna.org.

Expression, purification, and characterization of the histidine kinase CarS from Fusobacterium nucleatum

Fusobacterium nucleatum is an opportunistic pathogenic bacterium that can be enriched in colorectal cancer tissues, affecting multiple stages of colorectal cancer development. The two-component system plays an important role in the regulation and expression of genes related to pathogenic resistance and pathogenicity. In this paper, we focused on the CarRS two-component system of F. nucleatum, and the histidine kinase protein CarS was recombinantly expressed and characterized. Several online software such as SMART, CCTOP and AlphaFold2 were used to predict the secondary and tertiary structure of the CarS protein. The results showed that CarS is a membrane protein with two transmembrane helices and contains 9 α-helices and 12 β-folds. CarS protein is composed of two domains, one is the N-terminal transmembrane domain (amino acids 1-170), the other is the C-terminal intracellular domain. The latter is composed of a signal receiving domain (histidine kinases, adenylyl cyclases, methyl-accepting proteins, prokaryotic signaling proteins, HAMP), a phosphate receptor domain (histidine kinase domain, HisKA), and a histidine kinase catalytic domain (histidine kinase-like ATPase catalytic domain, HATPase_c). Since the full-length CarS protein could not be expressed in host cells, a fusion expression vector pET-28a(+)-MBP-TEV-CarScyto was constructed based on the characteristics of secondary and tertiary structures, and overexpressed in Escherichia coli BL21-Codonplus(DE3)RIL. CarScyto-MBP protein was purified by affinity chromatography, ion-exchange chromatography, and gel filtration chromatography with a final concentration of 20 mg/ml. CarScyto-MBP protein showed both protein kinase and phosphotransferase activities, and the MBP tag had no effect on the function of CarScyto protein. The above results provide a basis for in-depth analysis of the biological function of the CarRS two-component system in F. nucleatum.

Structure-Based GC Investigation Sheds New Light on ITS2 Evolution in Corydalis Species.

Guanine and cytosine (GC) content is a fundamental component of genetic diversity and essential for phylogenetic analyses. However, the GC content of the ribosomal internal transcribed spacer 2 (ITS2) remains unknown, despite the fact that ITS2 is a widely used phylogenetic marker. Here, the ITS2 was high-throughput sequenced from 29 Corydalis species, and their GC contents were comparatively investigated in the context of ITS2's characteristic secondary structure and concerted evolution. Our results showed that the GC contents of ITS2 were 131% higher than those of their adjacent 5.8S regions, suggesting that ITS2 underwent GC-biased evolution. These GCs were distributed in a heterogeneous manner in the ITS2 secondary structure, with the paired regions being 130% larger than the unpaired regions, indicating that GC is chosen for thermodynamic stability. In addition, species with homogeneous ITS2 sequences were always GC-rich, supporting GC-biased gene conversion (gBGC), which occurred with ITS2's concerted evolution. The RNA substitution model inferred also showed a GC preference among base pair transformations, which again supports gBGC. Overall, structurally based GC investigation reveals that ITS2 evolves under structural stability and gBGC selection, significantly increasing its GC content.

Characterizing nrDNA ITS1, 5.8S and ITS2 secondary structures and their phylogenetic utility in the legume tribe Hedysareae with special reference to Hedysarum.

This is the first study to systematically evaluate rRNA secondary structures of Hedysareae with an emphasis on Hedysarum. ITS2 and 5.8S regions of the genus shared a common secondary structure with a four-fingered central loop, whereas ITS1 possessed five distinct structures. The secondary structural features of the two regions provided advantageous data for clades, species groups, and closely related species. Hemi-CBCs were mostly observed in the reconstruction of species groups, and Nsts, mostly between closely related species. The investigations showed that ITS1 varied more than ITS2 in length, GC content, and most of the diversity indices within the tribe. Maximum likelihood analyses of the synchronized sequence-structure tree of ITS1 were performed. The accuracy and phylogenetic signals of ITS1 were higher than ITS2. The similar GC content, and no CBC, in both spacers, fortified the close relationship of CEGO and H. sections Stracheya and Hedysarum clades in the synchronized sequence-structure tree topology of ITS1. In both regions, no inter-generic CBCs were detected inside the CEGO clade and the inter-sectional level of Hedysarum. But, in the ITS2 region, a CBC was detected between H. section Multicaulia, and Taverniera versus H. sections Hedysarum, and Stracheya. The lowest inter-sectional genetic distance and structural features were found between H. sect. Hedysarum and H. sect. Stracheya clades in the ITS2 region.

Non-Perturbative Identification and Subtyping of Amyloidosis in Human Kidney Tissue with Raman Spectroscopy and Machine Learning.

Amyloids are proteins with characteristic beta-sheet secondary structures that display fibrillary ultrastructural configurations. They can result in pathologic lesions when deposited in human organs. Various types of amyloid protein can be routinely identified in human tissue specimens by special stains, immunolabeling, and electron microscopy, and, for certain forms of amyloidosis, mass spectrometry is required. In this study, we applied Raman spectroscopy to identify immunoglobulin light chain and amyloid A amyloidosis in human renal tissue biopsies and compared the results with a normal kidney biopsy as a control case. Raman spectra of amyloid fibrils within unstained, frozen, human kidney tissue demonstrated changes in conformation of protein secondary structures. By using t-distributed stochastic neighbor embedding (t-SNE) and density-based spatial clustering of applications with noise (DBSCAN), Raman spectroscopic data were accurately classified with respect to each amyloid type and deposition site. To the best of our knowledge, this is the first time Raman spectroscopy has been used for amyloid characterization of ex vivo human kidney tissue samples. Our approach, using Raman spectroscopy with machine learning algorithms, shows the potential for the identification of amyloid in pathologic lesions.

Evaluating propensity for fugitive gas migration from integrity compromised oil and gas wells in the Peace Region, British Columbia, Canada, with a petrophysical approach

Oil and gas wells can suffer integrity failure resulting in the release of fugitive gas, formed primarily of methane, into the subsurface and atmosphere. Factors affecting fugitive gas migration in the subsurface, and in particular within and between shallower units intersected by energy wells (i.e. <100 m depth), are usually poorly considered with little focus on specific lithological characteristics that control gas movement, determine environmental impacts and influence how best to detect, monitor and quantify fugitive gases. To better understand and identify properties that can control fugitive gas migration in shallow sedimentary rocks that form the Peace Region of British Columbia, Canada, we developed a workflow by which a series of rock cores from units present in the area were obtained and their flow properties characterised. Insights on propensity for fugitive gas migration within and between these rocks were achieved through laboratory techniques and analytical calculations that are usually applied to the study of deeper hydrocarbon migration and seal integrity. These included Mercury Intrusion Porosimetry, petrographic analysis, X-ray computed tomography, as well as estimations of threshold capillary pressure, gas-column heights, hydrostatic pore pressure and total displacement pressure. Results show a congruent correlation between the examined petrophysical rock characteristics and flow propensity. We find that high capillary pressure will act as an effective mechanism for limiting vertical gas migration in the region, even when potential barrier facies have porosities of more than 10% and permeability of 0.26 mD. Additionally, results suggest that the propensity of studied shallow sediments to transmit fugitive gases will be controlled by heterogeneity, permeability anisotropy and the presence of secondary structural features, such as microfractures. Our approach for evaluating the propensity of fugitive gas migration could be applied elsewhere to enhance understanding of potential migration pathways and aid in development of effective monitoring and detection strategies that can better anticipate gas behaviour in the subsurface and prognosticate reactive transport, impacts and fate.

Engineered Peptides Enable Biomimetic Route for Collagen Intrafibrillar Mineralization

Overcoming the short lifespan of current dental adhesives remains a significant clinical need. Adhesives rely on formation of the hybrid layer to adhere to dentin and penetrate within collagen fibrils. However, the ability of adhesives to achieve complete enclosure of demineralized collagen fibrils is recognized as currently unattainable. We developed a peptide-based approach enabling collagen intrafibrillar mineralization and tested our hypothesis on a type-I collagen-based platform. Peptide design incorporated collagen-binding and remineralization-mediating properties using the domain structure conservation approach. The structural changes from representative members of different peptide clusters were generated for each functional domain. Common signatures associated with secondary structure features and the related changes in the functional domain were investigated by attenuated total reflectance Fourier-transform infrared (ATR-FTIR) and circular dichroism (CD) spectroscopy, respectively. Assembly and remineralization properties of the peptides on the collagen platforms were studied using atomic force microscopy (AFM). Mechanical properties of the collagen fibrils remineralized by the peptide assemblies was studied using PeakForce-Quantitative Nanomechanics (PF-QNM)-AFM. The engineered peptide was demonstrated to offer a promising route for collagen intrafibrillar remineralization. This approach offers a collagen platform to develop multifunctional strategies that combine different bioactive peptides, polymerizable peptide monomers, and adhesive formulations as steps towards improving the long-term prospects of composite resins.

Fabrication of grains from bioprocessed nutri‐cereal flours: assessment of cooking profile, phytochemical composition, secondary protein fractions and structural characteristics

SummaryThe current study aimed at bioprocessing i.e. germination, fermentation and dual processing of different nutri‐cereals (red sorghum, white sorghum, foxtail millet, proso millet and barnyard millet) followed by their use in the development of fabricated grain in equal proportion with and without steaming. The fabricated grains were studied for cooking quality, bioactive composition and associated antioxidant activity, molecular interactions, structural characterisation and organoleptic evaluation. Bioprocessed fabricated grains had significantly lower cooking time, water absorption, and volume expansion but enhanced gruel solid loss while pre‐gelatinisation resulted in decreased losses. Bioprocessed fabricated grains had enhanced bioactive profiles when compared to pre‐gelatinised and control fabricated grains. Pre‐gelatinizaion modified the starch‐protein matrix preventing the leaching of compounds into the water. The results suggest that bioprocessed grains can be potential ingredients for the development of pre‐gelatinised fabricated grains.

A Novel Deep Learning Method for Predicting RNA-Protein Binding Sites

The cell cycle and biological processes rely on RNA and RNA-binding protein (RBP) interactions. It is crucial to identify the binding sites on RNA. Various deep-learning methods have been used for RNA-binding site prediction. However, they cannot extract the hierarchical features of the RNA secondary structure. Therefore, this paper proposes HPNet, which can automatically identify RNA-binding sites and -binding preferences. HPNet performs feature learning from the two perspectives of the RNA sequence and the RNA secondary structure. A convolutional neural network (CNN), a deep-learning method, is used to learn RNA sequence features in HPNet. To capture the hierarchical information for RNA, we introduced DiffPool into HPNet, a differentiable pooling graph neural network (GNN). A CNN and DiffPool were combined to improve the binding site prediction accuracy by leveraging both RNA sequence features and hierarchical features of the RNA secondary structure. Binding preferences can be extracted based on model outputs and parameters. Overall, the experimental results showed that HPNet achieved a mean area under the curve (AUC) of 94.5% for the benchmark dataset, which was more accurate than the state-of-the-art methods. Moreover, these results demonstrate that the hierarchical features of RNA secondary structure play an essential role in selecting RNA-binding sites.

NcDENSE: a novel computational method based on a deep learning framework for non-coding RNAs family prediction

BackgroundAlthough research on non-coding RNAs (ncRNAs) is a hot topic in life sciences, the functions of numerous ncRNAs remain unclear. In recent years, researchers have found that ncRNAs of the same family have similar functions, therefore, it is important to accurately predict ncRNAs families to identify their functions. There are several methods available to solve the prediction problem of ncRNAs family, whose main ideas can be divided into two categories, including prediction based on the secondary structure features of ncRNAs, and prediction according to sequence features of ncRNAs. The first type of prediction method requires a complicated process and has a low accuracy in obtaining the secondary structure of ncRNAs, while the second type of method has a simple prediction process and a high accuracy, but there is still room for improvement. The existing methods for ncRNAs family prediction are associated with problems such as complicated prediction processes and low accuracy, in this regard, it is necessary to propose a new method to predict the ncRNAs family more perfectly.ResultsA deep learning model-based method, ncDENSE, was proposed in this study, which predicted ncRNAs families by extracting ncRNAs sequence features. The bases in ncRNAs sequences were encoded by one-hot coding and later fed into an ensemble deep learning model, which contained the dynamic bi-directional gated recurrent unit (Bi-GRU), the dense convolutional network (DenseNet), and the Attention Mechanism (AM). To be specific, dynamic Bi-GRU was used to extract contextual feature information and capture long-term dependencies of ncRNAs sequences. AM was employed to assign different weights to features extracted by Bi-GRU and focused the attention on information with greater weights. Whereas DenseNet was adopted to extract local feature information of ncRNAs sequences and classify them by the full connection layer. According to our results, the ncDENSE method improved the Accuracy, Sensitivity, Precision, F-score, and MCC by 2.08%, 2.33%, 2.14%, 2.16%, and 2.39%, respectively, compared with the suboptimal method.ConclusionsOverall, the ncDENSE method proposed in this paper extracts sequence features of ncRNAs by dynamic Bi-GRU and DenseNet and improves the accuracy in predicting ncRNAs family and other data.