Experience of undergraduate nursing students participating in artificial intelligence + project task driven learning at different stages: a qualitative study.

BACKGROUND: Artificial intelligence is a growing phenomenon that will soon facilitate wide-scale changes in many professions, and is expected to play an important role in the field of medical education. This study explored the realistic feelings and experiences of nursing undergraduates participating in different stages of artificial intelligence + project task driven learning, and provide a basis for artificial intelligence participation in nursing teaching. METHODS: We conducted face-to-face semi-structured interviews with nursing undergraduates participating in Nursing Research Course which adopts artificial intelligence + project task driven learning from a medical university in Ningxia from September to November 2023, to understand their experience of using artificial intelligence for learning and the emotional changes at different stages. The interview guide included items about their personal experience and feelings of completing project tasks through dialogue with artificial intelligence, and suggestions for course content. Thematic analysis was used to analyze interview data. This study followed the COREQ checklist. RESULTS: According to the interview data, three themes were summarized. Undergraduate nursing students have different experiences in participating in artificial intelligence + project task driven learning at different stages, mainly manifested as diverse emotional experiences under initial knowledge deficiency, the individual growth supported by external forces during the adaptation period, and the expectations and suggestions after the birth of the results in the end period. CONCLUSIONS: Nursing undergraduates can actively adapt to the integration of artificial intelligence into nursing teaching, dynamically observe students' learning experience, strengthen positive guidance, and provide support for personalized teaching models, better leveraging the advantages of artificial intelligence participation in teaching.

BTCRSleep: a boundary temporal context refinement-based fully convolutional network for sleep staging with single-channel EEG.

Objective. Sleep staging studies on single-channel EEG mainly exploit deep learning methods that combine convolutional neural networks (CNNs) and recurrent neural networks. However, when typical brain waves (such as K-complexes or sleep spindles) that identify sleep stages span two epochs, the abstract process of a CNN extracting features from each sleep stage may cause the loss of boundary context information. This study attempts to capture the boundary context, which contains the characteristics of brain waves during sleep stage transition, to improve the performance of sleep staging.Approach. In this paper we propose a fully convolutional network with boundary temporal context refinement, called BTCRSleep (Boundary Temporal Context Refinement Sleep). The boundary temporal context refinement module refines the boundary information on sleep stages on the basis of extracting multi-scale temporal dependences between epochs and enhances the abstract capability of the boundary temporal context. In addition, we design a class-aware data augmentation method to effectively learn the boundary temporal context between the minority class and other sleep stages.Main results. We evaluate the performance of our proposed network using four public datasets: the 2013 version of Sleep-EDF Expanded (SEDF), the 2018 version of Sleep-EDF Expanded (SEDFX), the Sleep Heart Health Study (SHHS) and CAP Sleep Database (CAP). The evaluation results on the four datasets showed that our model obtains the best total accuracy and kappa score compared with state-of-the-art methods. On average, accuracies of 84.9% in SEDF, 82.9% in SEDFX, 85.2% in SHHS and 76.9% in CAP are obtained under subject-independent cross-validation. We demonstrate that the boundary temporal context contributes to the improvement in capturing the temporal dependences across different epochs.

Effects of Oat ß-Glucan on the Textural and Sensory Properties of Low-Fat Set Type Pea Protein Yogurt.

This study investigated the effect of oat ß-glucan as a fat substitute on the structure formation, texture, and sensory properties of pea protein yogurt. The results showed that the incorporation of 0.5% ß-glucan significantly accelerated the lactic acid bacteria-induced fermentation, with the time for reaching the target pH of 4.6 shortened from 3.5 h to 3 h (p < 0.05); increased the plastic module (G') from 693 Pa to 764 Pa when fermenting 3 h (p < 0.05); and enhanced the water-holding capacity from 77.29% to 82.15% (p < 0.05). The identification of volatile organic compounds (VOCs) in low-fat pea protein yogurt by GC-IMS revealed a significant decrease in aldehydes and a significant increase in alcohols, ketones and acids in the pea yogurt after fermentation (p < 0.05). Among them, the levels of acetic acid, acetone, 2,3-butanedione, 3-hydroxy-2-butanone, and ethyl acetate all significantly increased with the addition of oat ß-glucan (p < 0.05), thereby providing prominent fruity, sweet, and creamy flavors, respectively. Combined with the results of sensory analysis, the quality characteristics of pea protein yogurt with 1% oil by adding 1% oat ß-glucan were comparable to the control sample with 3% oil. Therefore, oat ß-glucan has a good potential for fat replacement in pea protein yogurt.

Supplement and Suppression: Both Boundary and Nonboundary Are Helpful for Salient Object Detection.

Current methods aggregate multilevel features from the backbone and introduce edge information to get more refined saliency maps. However, little attention is paid to how to suppress the regions with similar saliency appearances in the background. These regions usually exist in the vicinity of salient objects and have high contrast with the background, which is easy to be misclassified as foreground. To solve this problem, we propose a gated feature interaction network (GFINet) to integrate multiple saliency features, which can utilize nonboundary features with background information to suppress pseudosalient objects and simultaneously apply boundary features to supplement edge details. Different from previous methods that only consider the complementarity between saliency and boundary, the proposed network introduces nonboundary features into the decoder to filter the pseudosalient objects. Specifically, GFINet consists of global features aggregation branch (GFAB), boundary and nonboundary features' perception branch (B&NFPB), and gated feature interaction module (GFIM). According to the global features generated by GFAB, boundary and nonboundary features produced by B&NFPB and GFIM employ a gate structure to adaptively optimize the saliency information interchange between abovementioned features and, thus, predict the final saliency maps. Besides, due to the imbalanced distribution between the boundary pixels and nonboundary ones, the binary cross-entropy (BCE) loss is difficult to predict the pixels near the boundary. Therefore, we design a border region aware (BRA) loss to further boost the quality of boundary and nonboundary, which can guide the network to focus more on the hard pixels near the boundary by assigning different weights to different positions. Compared with 12 counterparts, experimental results on five benchmark datasets show that our method has better generalization and improves the state-of-the-art approach by 4.85% averagely in terms of the regional and boundary evaluation measures. In addition, our model is more efficient with an inference speed of 50.3 FPS when processing a 320 ×320 image. Code has been made available at https://github.com/lesonly/GFINet.

Dietary diversity, diet quality, and oxidative stress in older adults.

This cross-sectional study aimed to explore the relationship between diet quality, dietary diversity, and oxidative stress levels in older adults in parts of the Ningxia Hui Autonomous Region in northwestern China. A total of 335 participants voluntarily participated in the study from April to July 2021. Laboratory tests and questionnaires were used to obtain general characteristics, dietary conditions, and indicators of oxidative stress. The dietary diversity scores of the participants were 5.20±1.39, the diet quality indices were 56.91±11.14, and most had poor diet quality. The levels of the oxidative stress markers malondialdehyde, superoxide dismutase, 8-iso-prostaglandin F2α, and total antioxidant capacity were 4.77±1.77, 40.33±9.85, 763.10±245.41, and 1.02±0.14, respectively. The results showed that the diet quality scores of participants were related to dietary diversity and whether the scores passed or not was significantly related to total antioxidant capacity.

SleepContextNet: A temporal context network for automatic sleep staging based single-channel EEG.

BACKGROUND AND OBJECTIVE: Single-channel EEG is the most popular choice of sensing modality in sleep staging studies, because it widely conforms to the sleep staging guidelines. The current deep learning method using single-channel EEG signals for sleep staging mainly extracts the features of its surrounding epochs to obtain the short-term temporal context information of EEG epochs, and ignore the influence of the long-term temporal context information on sleep staging. However, the long-term context information includes sleep stage transition rules in a sleep cycle, which can further improve the performance of sleep staging. The aim of this research is to develop a temporal context network to capture the long-term context between EEG sleep stages. METHODS: In this paper, we design a sleep staging network named SleepContextNet for sleep stage sequence. SleepContextNet can extract and utilize the long-term temporal context between consecutive EEG epochs, and combine it with the short-term context. we utilize Convolutional Neural Network(CNN) layers for learning representative features from each sleep stage and the representation features sequence learned are fed into a Recurrent Neural Network(RNN) layer for learning long-term and short-term context information among sleep stage in chronological order. In addition, we design a data augmentation algorithm for EEG to retain the long-term context information without changing the number of samples. RESULTS: We evaluate the performance of our proposed network using four public datasets, the 2013 version of Sleep-EDF (SEDF), the 2018 version of Sleep-EDF Expanded (SEDFX), Sleep Heart Health Study (SHHS) and the CAP Sleep Database. The experimental results demonstrate that SleepContextNet outperforms state-of-the-art techniques in terms of different evaluation metrics by capturing long-term and short-term temporal context information. On average, accuracy of 84.8% in SEDF, 82.7% in SEDFX, 86.4% in SHHS and 78.8% in CAP are obtained under subject-independent cross validation. CONCLUSIONS: The network extracts the long-term and short-term temporal context information of sleep stages from the sequence features, which utilizes the temporal dependencies among the EEG epochs effectively and improves the accuracy of sleep stages. The sleep staging method based on forward temporal context information is suitable for real-time family sleep monitoring system.

Micro-polyethylene particles reduce the toxicity of nano zinc oxide in marine microalgae by adsorption.

Contaminant adsorption by microplastics (MPs) allows them to act as contaminant vehicles or vectors, complicating eco-toxicological study of MPs. The contaminants adsorbed are mainly organic contaminants, especially hydrophobic organic contaminants (HOCs), although heavy-metal adsorption has also been reported. Compared to the mechanisms of HOC adsorption, those for metals are not fully understood. In the present study, combined-exposure assays revealed that polyethylene microplastics (PEMPs, 150 µm) alleviate the toxic effect of nano zinc oxide (nZnO, 20-30 nm) on marine microalgal growth by 14.4%. Thus, we hypothesized that nZnO adsorption onto PEMP surfaces ameliorates its toxicity to microorganisms. To test this hypothesis, PEMP samples isolated from nZnO suspensions were characterized. Their surfaces were observed by SEM, their Zn levels were measured by ICP-MS, and the compound form of Zn on the PEMP surface was determined by XRD analysis. The results indicated that 5.53%-7.16% of the Zn in the suspension is adsorbed during the first 24 h of exposure and that the Zn remains as the ZnO form upon adsorption. The findings in the present study provide important information on the role of MPs as metal oxide vehicles.

The first mitochondrial genome of the genus Exhippolysmata (Decapoda: Caridea: Lysmatidae), with gene rearrangements and phylogenetic associations in Caridea.

The complete mitochondrial genome (mitogenome) of animals can provide useful information for evolutionary and phylogenetic analyses. The mitogenome of the genus Exhippolysmata (i.e., Exhippolysmata ensirostris) was sequenced and annotated for the first time, its phylogenetic relationship with selected members from the infraorder Caridea was investigated. The 16,350 bp mitogenome contains the entire set of 37 common genes. The mitogenome composition was highly A + T biased at 64.43% with positive AT skew (0.009) and negative GC skew (- 0.199). All tRNA genes in the E. ensirostris mitogenome had a typical cloverleaf secondary structure, except for trnS1 (AGN), which appeared to lack the dihydrouridine arm. The gene order in the E. ensirostris mitogenome was rearranged compared with those of ancestral decapod taxa, the gene order of trnL2-cox2 changed to cox2-trnL2. The tandem duplication-random loss model is the most likely mechanism for the observed gene rearrangement of E. ensirostris. The ML and BI phylogenetic analyses place all Caridea species into one group with strong bootstrap support. The family Lysmatidae is most closely related to Alpheidae and Palaemonidae. These results will help to better understand the gene rearrangements and evolutionary position of E. ensirostris and lay a foundation for further phylogenetic studies of Caridea.

Novel gene rearrangement in the mitochondrial genome of Siliqua minima (Bivalvia, Adapedonta) and phylogenetic implications for Imparidentia.

Siliqua minima (Gmelin, 1791) is an important economic shellfish species belonging to the family Pharidae. To date, the complete mitochondrial genome of only one species in this family (Sinonovacula constricta) has been sequenced. Research on the Pharidae family is very limited; to improve the evolution of this bivalve family, we sequenced the complete mitochondrial genome of S. minima by next-generation sequencing. The genome is 17,064 bp in length, consisting of 12 protein-coding genes (PCGs), 22 transfer RNA genes (tRNA), and two ribosomal RNA genes (rRNA). From the rearrangement analysis of bivalves, we found that the gene sequences of bivalves greatly variable among species, and with closer genetic relationship, the more consistent of the gene arrangement is higher among the species. Moreover, according to the gene arrangement of seven species from Adapedonta, we found that gene rearrangement among families is particularly obvious, while the gene order within families is relatively conservative. The phylogenetic analysis between species of the superorder Imparidentia using 12 conserved PCGs. The S. minima mitogenome was provided and will improve the phylogenetic resolution of Pharidae species.

Can Swimming Microalgal Cells be Vehicles for ZnO Nanoparticle Transportation and Thus Lead to Zn Diffusion?

The concentration of eco-toxic zinc oxide nanoparticles (nZnO) in aquatic ecosystems is increasing, and an effective method for their removal is needed. We hypothesize that microalgal cells may act as nZnO vehicles-if the nZnO concentration does not affect their swimming ability-enabling Zn diffusion and sedimentation. We conducted experiments using flasks connected via a U-type vessel; the first flask contained nZnO suspensions and second flask contained artificial seawater, respectively. We added microalgae to the first flask and illuminated the second. The microalgae appeared to promote sedimentation. However, only a few microalgal cells passed via phototaxis into the second flask, so the detection of nZnO or Zn ions in the second flask was not possible. Therefore, to confirm whether the microalgae affect Zn transportation, a more accurate method to detect nZnO or Zn ions at very low concentrations is needed.

Characterization of the complete mitochondrial genomes of two species of Penaeidae (Decapoda: Dendrobranchiata) and the phylogenetic implications for Penaeoidea.

In the present study, mitogenomes of the species Trachypenaeus curvirostris and Parapenaeus fissuroides (Decapoda: Dendrobranchiata: Penaeidae) were sequenced. The total lengths of the two species were 15,956 bp and 15,937 bp in length with A + T biases of 67.08% and 67.69%, respectively. Both two species showed positive AT skews (0.016, 0.058) and negative GC skews (-0.254, -0.310). Both mitogenomes contained 13 protein-coding genes, 22 transfer RNA genes, and 2 ribosomal RNA genes. Results of phylogenetic analyses support close relationships among Aristeidae, Benthesicymidae and Solenoceridae. The family Sicyoniidae was observed to be deeply nested within Penaeidae. Within Penaeidae, T. curvirostris and P. fissuroides were most closely related to the genus Parapenaeopsis and Metapenaeopsis, respectively, indicated that these two species belong to Penaeidae. These results will help to better understand the evolutionary position of Penaeidae and provide reference for further phylogenetic research on Penaeoidea species.

The complete mitochondrial genome and phylogenetic analysis of Littorina brevicula (Gastropoda, Littorinidea).

The complete mitochondrial genome of the Littorina brevicula was determined in this study. The complete mitogenome (mtDNA) is 16,356 base pairs (bp) in length and contains 13 protein-coding genes, two rRNA genes, and 22 tRNA genes. The control region was divided into two parts. The overall base composition of the genome in descending order was 35.33%-T, 28.41%-A, 20.43%-C, and 15.81%-G. In 13 protein-coding genes, 12 genes start with ATG, except nad5 starts with ATT. For the stop codon, seven genes end with TAA, atp6, nad41, and nad3 end with TAG. Phylogenetic analysis indicated that L. brevicula is close to Naticidea family. This study first determined the complete mitochondrial genome of L. brevicula. It would be a supplement for the genetic analysis of L. brevicula and promote the phylogenetic of Littorinidea.

The complete mitochondrial genome and phylogenetic analysis of Omphalius rusticus (Gastropoda, Tegulidae).

The complete mitochondrial genome of the Omphalius rusticus has been determined. The complete genome is 18,067 bp and contained 13 protein-coding genes, two rRNA genes and 22 tRNA genes. The overall base composition is 33.48% (A), 33.52% (T), 15.58% (G) and 17.42% (C). The all start codon for 13 protein-coding genes is ATG and the most common termination codon is TAA. The phylogenetic tree showed that O. rusticus is most closely related to the Tectus pyramis. We suggest that this result will further supplement the genome information in mitochondria of the family Tegulidae and facilitate the study on population genetics.

Comparative analysis of the complete mitochondrial genomes in two limpets from Lottiidae (Gastropoda: Patellogastropoda): rare irregular gene rearrangement within Gastropoda.

To improve the systematics and taxonomy of Patellogastropoda within the evolution of gastropods, we determined the complete mitochondrial genome sequences of Lottia goshimai and Nipponacmea fuscoviridis in the family Lottiidae, which presented sizes of 18,192 bp and 18,720 bp, respectively. In addition to 37 common genes among metazoa, we observed duplication of the trnM gene in L. goshimai and the trnM and trnW genes in N. fuscoviridis. The highest A + T contents of the two species were found within protein-coding genes (59.95% and 54.55%), followed by rRNAs (56.50% and 52.44%) and tRNAs (56.42% and 52.41%). trnS1 and trnS2 could not form the canonical cloverleaf secondary structure due to the lack of a dihydrouracil arm in both species. The gene arrangements in all Patellogastropoda compared with those of ancestral gastropods showed different levels of gene rearrangement, including the shuffling, translocation and inversion of single genes or gene fragments. This kind of irregular rearrangement is particularly obvious in the Lottiidae family. The results of phylogenetic and gene rearrangement analyses showed that L. goshimai and Lottia digitalis clustered into one group, which in turn clustered with N. fuscoviridis in Patellogastropoda. This study demonstrates the significance of complete mitogenomes for phylogenetic analysis and enhances our understanding of the evolution of Patellogastropoda.

Sequence comparison of the mitochondrial genomes in two species of the genus Nerita (Gastropoda: Neritimorpha: Neritidae): phylogenetic implications and divergence time estimation for Neritimorpha.

Many Nerita species live in warm-water environments, and they are some of the few organisms from the intertidal zone that can live in both freshwater and seawater. Previous comparative studies of the mitogenomes of Nerita species suggest that the genome rearrangements are very conservative. Generally, the species possess a set of similar mitochondrial gene arrangements, but nucleotide sequences can be used to elucidate phylogenetic relationships at various levels of divergence. Here, the mitogenomes of Nerita undata and Nerita balteata were sequenced and found to be 15,583 bp and 15,571 bp, respectively. The mitogenomes of both species contain 13 protein-coding genes, 22 tRNA genes, and two rRNA genes. The nucleotides of the two mitogenomes are highly similar, with the same gene composition and genomic organization as those present in other Nerita species. The tRNA secondary structures were different from those of other gastropods: trnS2 is not folded into typical secondary structures, and the dihydrouridine (DHU) arm simply forms a loop. The phylogenetic analysis showed that Neritimorpha is a sister group of Vetigastropoda and Caenogastropoda. Nerita balteata is a sister group of Nerita versicolor and Nerita undata, and all three species belong to Neritimorpha. This study contributes towards the comparative mitogenomic analysis of Neritidae and phylogenetic considerations among Neritimorpha species. The estimation of divergence time revealed that the two Nerita species were differentiated in the late Paleogene of the Cenozoic Era, and their evolution may be related to environmental changes.

Effects of microplastics on growth, phenanthrene stress, and lipid accumulation in a diatom, Phaeodactylum tricornutum.

Most laboratory studies have focused on the effects of nanoplastics instead of plastics at the micrometer scale, which are the major microplastics (MPs) discarded in marine environments. Knowledge on the potential effects of micrometer scale plastics on marine microalgae remains limited. It remains unknown whether the micrometer scale plastics also affect microalgal growth, lipid accumulation and resistance to organic contaminants? In addition, the role of polymer-size on the potential hazardous effects of MPs on microalgae is unknown. In the present study, cell populations of a marine diatom, Phaeodactylum tricornutum, were treated with micrometer scale polyethylene (PEMP, 150 µm) and unplasticized polyvinyl chloride (uPVCMP, 250 µm) powders in the laboratory. Growth was assessed using a hemacytometer and neutral lipid concentrations were evaluated using the Nile Red staining method under short-term (four days) and long-term (nine days) exposure. The effects of combined PEMP and phenanthrene (Phe), and uPVCMP and Phe exposures over four days on growth were investigated. Importance scores and SHapley Additive exPlanations (SHAP) values were calculated to assess the contributions of seven factors in exposure systems to the hazardous effects of MPs on microalgae using a machine-learning prediction based on 165 data sets. Both MP types did not influence algal growth and lipid accumulation but minimized algal inhibition by the action of Phe at four days. In addition, lipid accumulation was induced at nine days. Both importance scores and SHAP values indicated that MP polymer-size was the key factor influencing MP toxicity in microalgae. In conclusion, MPs had adverse effects only in chronic tests and the potential adsorption of MPs could have led to the lower levels of toxicity in a combined MP-Phe exposure system. Compared to nanoplastics, MPs in the hundred-micrometer range do not significantly affect growth and their adsorption would not be influenced by size. Therefore, MP size is the most critical factor that should be considered in future laboratory tests and eco-toxicological risk assessments for microalgae.

Strong genetic differentiation of the clam Meretrix lamarckii in the China Sea revealed by mitochondrial DNA marker.

The hard clam Meretrix lamarckii is ecologically and economically important in the coastal regions of China. In this study, we evaluated the genetic diversity and population structure of six M. lamarckii populations in the East China Sea and the South China Sea using mitochondrial cytochrome c oxidase subunit 1 (COI) and cytochrome b (Cytb) genes. We obtained 582 bp of partly sequences comprising 28 novel haplotypes of COI gene from 138 specimens and 1168 bp of partly sequences comprising 22 novel haplotypes of Cytb gene from 125 specimens. The haplotype diversity of COI and Cytb genes ranged from 0.606 to 0.862 and 0.562 to 0.863, respectively. The nucleotide diversity ranged from 0.0015 to 0.0038 in COI gene and ranged from 0.0007 to 0.0032 in Cytb gene. Thus, there is moderate-level genetic diversity in M. lamarckii in the China Sea. The F-statistics showed that the Zhoushan (ZS) and Xiangshan (XS) populations were significantly (P < 0.01) differed from the populations of Wenzhou (WZ), Zhangpu (ZP), Shantou (ST), and Zhanjiang (ZJ) in both COI and Cytb genes. Both haplotypes network and plot of STRUCTURE analysis suggested obviously genetic divergence between East China Sea and South China Sea regions. Knowledge on genetic variation and population structure of M. lamarckii populations along the Southeast China Sea obtained from this study will support the aquaculture management and conservation of M. lamarckii in China.

CNNDLP: A Method Based on Convolutional Autoencoder and Convolutional Neural Network with Adjacent Edge Attention for Predicting lncRNA-Disease Associations.

It is well known that the unusual expression of long non-coding RNAs (lncRNAs) is closely related to the physiological and pathological processes of diseases. Therefore, inferring the potential lncRNA-disease associations are helpful for understanding the molecular pathogenesis of diseases. Most previous methods have concentrated on the construction of shallow learning models in order to predict lncRNA-disease associations, while they have failed to deeply integrate heterogeneous multi-source data and to learn the low-dimensional feature representations from these data. We propose a method based on the convolutional neural network with the attention mechanism and convolutional autoencoder for predicting candidate disease-related lncRNAs, and refer to it as CNNDLP. CNNDLP integrates multiple kinds of data from heterogeneous sources, including the associations, interactions, and similarities related to the lncRNAs, diseases, and miRNAs. Two different embedding layers are established by combining the diverse biological premises about the cases that the lncRNAs are likely to associate with the diseases. We construct a novel prediction model based on the convolutional neural network with attention mechanism and convolutional autoencoder to learn the attention and the low-dimensional network representations of the lncRNA-disease pairs from the embedding layers. The different adjacent edges among the lncRNA, miRNA, and disease nodes have different contributions for association prediction. Hence, an attention mechanism at the adjacent edge level is established, and the left side of the model learns the attention representation of a pair of lncRNA and disease. A new type of lncRNA similarity and a new type of disease similarity are calculated by incorporating the topological structures of multiple bipartite networks. The low-dimensional network representation of the lncRNA-disease pairs is further learned by the autoencoder based convolutional neutral network on the right side of the model. The cross-validation experimental results confirm that CNNDLP has superior prediction performance compared to the state-of-the-art methods. Case studies on stomach cancer, breast cancer, and prostate cancer further show the ability of CNNDLP for discovering the potential disease lncRNAs.

Dual Convolutional Neural Network Based Method for Predicting Disease-Related miRNAs.

Identification of disease-related microRNAs (disease miRNAs) is helpful for understanding and exploring the etiology and pathogenesis of diseases. Most of recent methods predict disease miRNAs by integrating the similarities and associations of miRNAs and diseases. However, these methods fail to learn the deep features of the miRNA similarities, the disease similarities, and the miRNA⁻disease associations. We propose a dual convolutional neural network-based method for predicting candidate disease miRNAs and refer to it as CNNDMP. CNNDMP not only exploits the similarities and associations of miRNAs and diseases, but also captures the topology structures of the miRNA and disease networks. An embedding layer is constructed by combining the biological premises about the miRNA⁻disease associations. A new framework based on the dual convolutional neural network is presented for extracting the deep feature representation of associations. The left part of the framework focuses on integrating the original similarities and associations of miRNAs and diseases. The novel miRNA and disease similarities which contain the topology structures are obtained by random walks on the miRNA and disease networks, and their deep features are learned by the right part of the framework. CNNDMP achieves the superior prediction performance than several state-of-the-art methods during the cross-validation process. Case studies on breast cancer, colorectal cancer and lung cancer further demonstrate CNNDMP's powerful ability of discovering potential disease miRNAs.

Prediction of potential disease-associated microRNAs based on random walk.

MOTIVATION: Identifying microRNAs associated with diseases (disease miRNAs) is helpful for exploring the pathogenesis of diseases. Because miRNAs fulfill function via the regulation of their target genes and because the current number of experimentally validated targets is insufficient, some existing methods have inferred potential disease miRNAs based on the predicted targets. It is difficult for these methods to achieve excellent performance due to the high false-positive and false-negative rates for the target prediction results. Alternatively, several methods have constructed a network composed of miRNAs based on their associated diseases and have exploited the information within the network to predict the disease miRNAs. However, these methods have failed to take into account the prior information regarding the network nodes and the respective local topological structures of the different categories of nodes. Therefore, it is essential to develop a method that exploits the more useful information to predict reliable disease miRNA candidates. RESULTS: miRNAs with similar functions are normally associated with similar diseases and vice versa. Therefore, the functional similarity between a pair of miRNAs is calculated based on their associated diseases to construct a miRNA network. We present a new prediction method based on random walk on the network. For the diseases with some known related miRNAs, the network nodes are divided into labeled nodes and unlabeled nodes, and the transition matrices are established for the two categories of nodes. Furthermore, different categories of nodes have different transition weights. In this way, the prior information of nodes can be completely exploited. Simultaneously, the various ranges of topologies around the different categories of nodes are integrated. In addition, how far the walker can go away from the labeled nodes is controlled by restarting the walking. This is helpful for relieving the negative effect of noisy data. For the diseases without any known related miRNAs, we extend the walking on a miRNA-disease bilayer network. During the prediction process, the similarity between diseases, the similarity between miRNAs, the known miRNA-disease associations and the topology information of the bilayer network are exploited. Moreover, the importance of information from different layers of network is considered. Our method achieves superior performance for 18 human diseases with AUC values ranging from 0.786 to 0.945. Moreover, case studies on breast neoplasms, lung neoplasms, prostatic neoplasms and 32 diseases further confirm the ability of our method to discover potential disease miRNAs. AVAILABILITY AND IMPLEMENTATION: A web service for the prediction and analysis of disease miRNAs is available at http://bioinfolab.stx.hk/midp/.