IEEE Transactions on NanoBioscience最新文献

This paper presents a terahertz metasurface based sensor design incorporating graphene and other plasmonic materials for highly sensitive detection of different chemicals. The proposed sensor employs the combination of multiple resonator designs - including circular and square ring resonators - to attain enhanced sensitivity among other performance parameters. Machine learning techniques like Random Forest regression, are employed to enhance the sensor design and predict its performance. The optimized sensor demonstrates excellent sensitivity of 417 GHzRIU^-1 and a low detection limit of 0.264 RIU for ethanol and benzene detection. Furthermore, the integration of machine learning cuts down the simulation time and computational requirements by approximately 90% without compromising accuracy. The sensor's unique design and performance characteristics, including its high-quality factor of 14.476, position it as a promising candidate for environmental monitoring and chemical sensing applications. Moreover, it also demonstrates potential for 2-bit encoding applications through strategic modulation of graphene chemical potential values. On the other hand, it also shows prospects of 2-bit encoding applications via the modulation of graphene chemical. This work provides a major advancement to the terahertz sensing application by proposing new materials, structures, and methods in computation in order to develop a high-performance chemical sensor.

Circular RNAs (circRNAs) play a crucial role in gene regulation and association with diseases because of their unique closed continuous loop structure, which is more stable and conserved than ordinary linear RNAs. As fundamental work to clarify their functions, a large number of computational approaches for identifying circRNA formation have been proposed. However, these methods fail to fully utilize the important characteristics of back-splicing events, i.e., the positional information of the splice sites and the interaction features of its flanking sequences, for predicting circRNAs. To this end, we hereby propose a novel approach called SIDE for predicting circRNA back-splicing events using only raw RNA sequences. Technically, SIDE employs a dual encoder to capture global and interactive features of the RNA sequence, and then a decoder designed by the contrastive learning to fuse out discriminative features improving the prediction of circRNAs formation. Empirical results on three real-world datasets show the effectiveness of SIDE. Further analysis also reveals that the effectiveness of SIDE.

Computational synthetic lethality (SL) method has become a promising strategy to identify SL gene pairs for targeted cancer therapy and cancer medicine development. Feature representation for integrating various biological networks is crutial to improve the identification performance. However, previous feature representation, such as matrix factorization and graph neural network, projects gene features onto latent variables by keeping a specific geometric metric. There is a lack of models of gene representational latent space with considerating multiple dimentionalities correlation and preserving latent geometric structures in both sample and feature spaces. Therefore, we propose a novel method to model gene Latent Space using matrix Tri-Factorization (LSTF) to obtain gene representation with embedding variables resulting from the potential interpretation of synthetic lethality. Meanwhile, manifold subspace regularization is applied to the tri-factorization to capture the geometrical manifold structure in the latent space with gene PPI functional and GO semantic embeddings. Then, SL gene pairs are identified by the reconstruction of the associations with gene representations in the latent space. The experimental results illustrate that LSTF is superior to other state-of-the-art methods. Case study demonstrate the effectiveness of the predicted SL associations.

Detecting side effects of drugs is a fundamental task in drug development. With the expansion of publicly available biomedical data, researchers have proposed many computational methods for predicting drug-side effect associations (DSAs), among which network-based methods attract wide attention in the biomedical field. However, the problem of data scarcity poses a great challenge for existing DSAs prediction models. Although several data augmentation methods have been proposed to address this issue, most of existing methods employ a random way to manipulate the original networks, which ignores the causality of existence of DSAs, leading to the poor performance on the task of DSAs prediction. In this paper, we propose a counterfactual inference-based data augmentation method for improving the performance of the task. First, we construct a heterogeneous information network (HIN) by integrating multiple biomedical data. Based on the community detection on the HIN, a counterfactual inference-based method is designed to derive augmented links, and an augmented HIN is obtained accordingly. Then, a meta-path-based graph neural network is applied to learn high-quality representations of drugs and side effects, on which the predicted DSAs are obtained. Finally, comprehensive experiments are conducted, and the results demonstrate the effectiveness of the proposed counterfactual inference-based data augmentation for the task of DSAs prediction.

Given the persistent global challenge presented by rapidly spreading diseases, as evidenced notably by the widespread impact of the COVID-19 pandemic on both human health and economies worldwide, the necessity of developing effective infectious disease prediction models has become of utmost importance. In this context, the utilization of online social media platforms as valuable tools in healthcare settings has gained prominence, offering direct avenues for disseminating critical health information to the public in a timely and accessible manner. Propelled by the ubiquitous accessibility of the internet through computers and mobile devices, these platforms promise to revolutionize traditional detection methods, providing more immediate and reliable epidemiological insights. Leveraging this paradigm shift, our proposed framework harnesses Twitter data associated with infectious disease symptoms, employing ontology to identify and curate relevant tweets. Central to our methodology is a hybrid model that integrates XGBoost and Bidirectional Long Short-Term Memory (BiLSTM) architectures. The integration of XGBoost addresses the challenge of handling small dataset sizes, inherent during outbreaks due to limited time series data. XGBoost serves as a cornerstone for minimizing the loss function and identifying optimal features from our multivariate time series data. Subsequently, the combined dataset, comprising original features and predicted values by XGBoost, is channeled into the BiLSTM for further processing. Through extensive experimentation with a dataset spanning multiple infectious disease outbreaks, our hybrid model demonstrates superior predictive performance compared to state-of-the-art and baseline models. By enhancing forecasting accuracy and outbreak tracking capabilities, our model offers promising prospects for assisting health authorities in mitigating fatalities and proactively preparing for potential outbreaks.

Bioinformatics is a rapidly growing field involving the application of computational methods to the analysis and interpretation of biological data. An important task in bioinformatics is the identification of novel drug-target interactions (DTIs), which is also an important part of the drug discovery process. Most computational methods for predicting DTI consider it as a binary classification task to predict whether drug target pairs interact with each other. With the increasing amount of drug-target binding affinity data in recent years, this binary classification task can be transformed into a regression task of drug-target affinity (DTA), which reflects the degree of drug-target binding and can provide more detailed and specific information than DTI, making it an important tool in drug discovery through virtual screening. Effectively predicting how compounds interact with targets can help speed up the drug discovery process. In this study, we propose a deep learning model called TC-DTA for the prediction of the DTA, which makes use of the convolutional neural networks (CNN) and encoder module of the transformer architecture. First, the raw drug SMILES strings and protein amino acid sequences are extracted from the dataset. These are then represented using different encoding methods. We then use CNN and the Transformer's encoder module to extract feature information from drug SMILES strings and protein amino acid sequences, respectively. Finally, the feature information obtained is concatenated and fed into a multi-layer perceptron for prediction of the binding affinity score. We evaluated our model on two benchmark DTA datasets, Davis and KIBA, against methods including KronRLS, SimBoost and DeepDTA. On evaluation metrics such as Mean Squared Error, Concordance Index and r²_m index, TC-DTA outperforms these baseline methods. These results demonstrate the effectiveness of the Transformer's encoder and CNN in the extraction of meaningful representations from sequences, thereby improving the accuracy of DTA prediction. The deep learning model for DTA prediction can accelerate drug discovery by identifying drug candidates with high binding affinity to specific targets. Compared to traditional methods, the use of machine learning technology allows for a more effective and efficient drug discovery process.

Pancreatic cancer is one of the most malignant cancers with rapid progression and poor prognosis. The use of transcriptional data can be effective in finding new biomarkers for pancreatic cancer. Many network-based methods used to identify cancer biomarkers are proposed, among which the combination of network controllability appears. However, most of the existing methods do not study RNA, rely on priori and mutations information, or can only achieve classification tasks. In this study, we propose a method combined Relational Graph Convolutional Network and Deep Q-Network called RDDriver to identify pancreatic cancer biomarkers based on multi-layer heterogeneous transcriptional regulation network. Firstly, we construct a regulation network containing long non-coding RNA, microRNA, and messenger RNA. Secondly, Relational Graph Convolutional Network is used to learn the node representation. Finally, we use the idea of Deep Q-Network to build a model, which score and prioritize each RNA with the Popov-Belevitch-Hautus criterion. We train RDDriver on three small simulated networks, and calculate the average score after applying the model parameters to the regulation networks separately. To demonstrate the effectiveness of the method, we perform experiments for comparison between RDDriver and other eight methods based on the approximate benchmark of three types cancer drivers RNAs.