Drug abuse has always been a severe issue, but the proportion of drug abuse and addiction is rising. According to research reports, youth are motivated to access drugs mainly due to curiosity and peer influence. Additionally, youth especially lack proper knowledge and education surrounding drug abuse. Analyzing whether potential addicts intend to access drugs is helpful in preventing drug abuse and addiction. We developed an Anti-drug Chatbot for young people on a popular online social platform. We can detect potential risks, obtain warnings from the user-entered query and provide these to professional consultants for help. In this article, we present a hierarchical system with bidirectional encoder representation from transformers (BERT) to efficiently recognize and classify a user’s intent. We use the Chinese BERT-based model to utilize contextual information to perform classification and recognition. We evaluate our proposed system on our conversational dataset.

In the past decade, convolutional neural networks (CNNs) have been used as powerful tools by scientists to solve visual data tasks. However, many efforts of convolutional neural networks in solving protein function prediction and extracting useful information from protein sequences have certain limitations. In this research, we propose a new method to improve the weaknesses of the previous method. mCNN-ETC is a deep learning model which can transform the protein evolutionary information into image-like data composed of 20 channels, which correspond to the 20 amino acids in the protein sequence. We constructed CNN layers with different scanning windows in parallel to enhance the useful pattern detection ability of the proposed model. Then we filtered specific patterns through the 1-max pooling layer before inputting them into the prediction layer. This research attempts to solve a basic problem in biology in terms of application: predicting electron transporters and classifying their corresponding complexes. The performance result reached an accuracy of 97.41%, which was nearly 6% higher than its predecessor. We have also published a web server on http://bio219.bioinfo.yzu.edu.tw, which can be used for research purposes free of charge.

Living organisms receive necessary energy substances directly from cellular respiration. The completion of electron storage and transportation requires the process of cellular respiration with the aid of electron transport chains. Therefore, the work of deciphering electron transport proteins is inevitably needed. The identification of these proteins with high performance has a prompt dependence on the choice of methods for feature extraction and machine learning algorithm. In this study, protein sequences served as natural language sentences comprising words. The nominated word embedding-based feature sets, hinged on the word embedding modulation and protein motif frequencies, were useful for feature choosing. Five word embedding types and a variety of conjoint features were examined for such feature selection. The support vector machine algorithm consequentially was employed to perform classification. The performance statistics within the 5-fold cross-validation including average accuracy, specificity, sensitivity, as well as MCC rates surpass 0.95. Such metrics in the independent test are 96.82, 97.16, 95.76 percent, and 0.9, respectively. Compared to state-of-the-art predictors, the proposed method can generate more preferable performance above all metrics indicating the effectiveness of the proposed method in determining electron transport proteins. Furthermore, this study reveals insights about the applicability of various word embeddings for understanding surveyed sequences.

It is well-known that the major reason for the rapid proliferation of cancer cells are the hypomethylation of the whole cancer genome and the hypermethylation of the promoter of particular tumor suppressor genes. Locating 5-methylcytosine (5mC) sites in promoters is therefore a crucial step in further understanding of the relationship between promoter methylation and the regulation of mRNA gene expression. High throughput identification of DNA 5mC in wet lab is still time-consuming and labor-extensive. Thus, finding the 5mC site of genome-wide DNA promoters is still an important task. We compared the effectiveness of the most popular and strong machine learning techniques namely XGBoost, Random Forest, Deep Forest, and Deep Feedforward Neural Network in predicting the 5mC sites of genome-wide DNA promoters. A feature extraction method based on k-mers embeddings learned from a language model were also applied. Overall, the performance of all the surveyed models surpassed deep learning models of the latest studies on the same dataset employing other encoding scheme. Furthermore, the best model achieved AUC scores of 0.962 on both cross-validation and independent test data. We concluded that our approach was efficient for identifying 5mC sites of promoters with high performance.

Recently, language representation models have drawn a lot of attention in the natural language processing field due to their remarkable results. Among them, bidirectional encoder representations from transformers (BERT) has proven to be a simple, yet powerful language model that achieved novel state-of-the-art performance. BERT adopted the concept of contextualized word embedding to capture the semantics and context of the words in which they appeared. In this study, we present a novel technique by incorporating BERT-based multilingual model in bioinformatics to represent the information of DNA sequences. We treated DNA sequences as natural sentences and then used BERT models to transform them into fixed-length numerical matrices. As a case study, we applied our method to DNA enhancer prediction, which is a well-known and challenging problem in this field. We then observed that our BERT-based features improved more than 5-10% in terms of sensitivity, specificity, accuracy and Matthews correlation coefficient compared to the current state-of-the-art features in bioinformatics. Moreover, advanced experiments show that deep learning (as represented by 2D convolutional neural networks; CNN) holds potential in learning BERT features better than other traditional machine learning techniques. In conclusion, we suggest that BERT and 2D CNNs could open a new avenue in biological modeling using sequence information.

Since 2015, a fast growing number of deep learning-based methods have been proposed for protein-ligand binding site prediction and many have achieved promising performance. These methods, however, neglect the imbalanced nature of binding site prediction problems. Traditional data-based approaches for handling data imbalance employ linear interpolation of minority class samples. Such approaches may not be fully exploited by deep neural networks on downstream tasks. We present a novel technique for balancing input classes by developing a deep neural network-based variational autoencoder (VAE) that aims to learn important attributes of the minority classes concerning nonlinear combinations. After learning, the trained VAE was used to generate new minority class samples that were later added to the original data to create a balanced dataset. Finally, a convolutional neural network was used for classification, for which we assumed that the nonlinearity could be fully integrated. As a case study, we applied our method to the identification of FAD- and FMN-binding sites of electron transport proteins. Compared with the best classifiers that use traditional machine learning algorithms, our models obtained a great improvement on sensitivity while maintaining similar or higher levels of accuracy and specificity. We also demonstrate that our method is better than other data imbalance handling techniques, such as SMOTE, ADASYN, and class weight adjustment. Additionally, our models also outperform existing predictors in predicting the same binding types. Our method is general and can be applied to other data types for prediction problems with moderate-to-heavy data imbalances.

Motivation: Efflux protein plays a key role in pumping xenobiotics out of the cells. The prediction of efflux family proteins involved in transport process of compounds is crucial for understanding family structures, functions and energy dependencies. Many methods have been proposed to classify efflux pump transporters without considerations of any pump specific of efflux protein families. In other words, efflux proteins protect cells from extrusion of foreign chemicals. Moreover, almost all efflux protein families have the same structure based on the analysis of significant motifs. The motif sequences consisting of the same amount of residues will have high degrees of residue similarity and thus will affect the classification process. Consequently, it is challenging but vital to recognize the structures and determine energy dependencies of efflux protein families. In order to efficiently identify efflux protein families with considering about pump specific, we developed a 2 D convolutional neural network (2 D CNN) model called DeepEfflux. DeepEfflux tried to capture the motifs of sequences around hidden target residues to use as hidden features of families. In addition, the 2 D CNN model uses a position-specific scoring matrix (PSSM) as an input. Three different datasets, each for one family of efflux protein, was fed into DeepEfflux, and then a 5-fold cross validation approach was used to evaluate the training performance. Results: The model evaluation results show that DeepEfflux outperforms traditional machine learning algorithms. Furthermore, the accuracy of 96.02%, 94.89% and 90.34% for classes A, B and C, respectively, in the independent test results show that our model can perform well and can be used as a reliable tool for identifying families of efflux proteins in transporters. Availability and implementation: The online version of deepefflux is available at http://deepefflux.irit.fr. The source code of deepefflux is available both on the deepefflux website and at http://140.138.155.216/deepefflux/. Supplementary information: Supplementary data are available at Bioinformatics online.

In several years, deep learning is a new area of machine learning field, which is the motivation of developing machine learning near to artificial intelligent. The neural networks belongs to deep learning are progressively important ideas in a variety of fields with great performance. Accordingly, utilization of deep learning in bioinformatics to enhance performance is very important. Convolutional neural networks is a network of deep learning which is claimed to be the best model to solve the problem of object recognition and detection utilizing GPU computing. In this study, we try to use CNN to identify efflux proteins in membrane and transport proteins, which is a famous problem in bioinformatics field. We construct the CNN from PSSM profiles with CUDA and Keras package based on Theano backend. Finally this approach achieved a significant improvement after we compare with the previous paper on efflux proteins. The proposed method can serve as an effective tool for identifying efflux proteins and can help biologists understand the functions of the efflux proteins. Moreover this study provides a basis for further research that can enrich a field of applying deep learning in bioinformatics.

BACKGROUND: Cellular respiration is a catabolic pathway for producing adenosine triphosphate (ATP) and is the most efficient process through which cells harvest energy from consumed food. When cells undergo cellular respiration, they require a pathway to keep and transfer electrons (i.e., the electron transport chain). Due to oxidation-reduction reactions, the electron transport chain produces a transmembrane proton electrochemical gradient. In case protons flow back through this membrane, this mechanical energy is converted into chemical energy by ATP synthase. The convert process is involved in producing ATP which provides energy in a lot of cellular processes. In the electron transport chain process, flavin adenine dinucleotide (FAD) is one of the most vital molecules for carrying and transferring electrons. Therefore, predicting FAD binding sites in the electron transport chain is vital for helping biologists understand the electron transport chain process and energy production in cells. RESULTS: We used an independent data set to evaluate the performance of the proposed method, which had an accuracy of 69.84 %. We compared the performance of the proposed method in analyzing two newly discovered electron transport protein sequences with that of the general FAD binding predictor presented by Mishra and Raghava and determined that the accuracy of the proposed method improved by 9-45 % and its Matthew's correlation coefficient was 0.14-0.5. Furthermore, the proposed method enabled reducing the number of false positives significantly and can provide useful information for biologists. CONCLUSIONS: We developed a method that is based on PSSM profiles and SAAPs for identifying FAD binding sites in newly discovered electron transport protein sequences. This approach achieved a significant improvement after we added SAAPs to PSSM features to analyze FAD binding proteins in the electron transport chain. The proposed method can serve as an effective tool for predicting FAD binding sites in electron transport proteins and can help biologists understand the functions of the electron transport chain, particularly those of FAD binding sites. We also developed a web server which identifies FAD binding sites in electron transporters available for academics.

BACKGROUND: Guanonine-protein (G-protein) is known as molecular switches inside cells, and is very important in signals transmission from outside to inside cell. Especially in transport protein, most of G-proteins play an important role in membrane trafficking; necessary for transferring proteins and other molecules to a variety of destinations outside and inside of the cell. The function of membrane trafficking is controlled by G-proteins via Guanosine triphosphate (GTP) binding sites. The GTP binding sites active G-proteins initiated to membrane vesicles by interacting with specific effector proteins. Without the interaction from GTP binding sites, G-proteins could not be active in membrane trafficking and consequently cause many diseases, i.e., cancer, Parkinson… Thus it is very important to identify GTP binding sites in membrane trafficking, in particular, and in transport protein, in general. RESULTS: We developed the proposed model with a cross-validation and examined with an independent dataset. We achieved an accuracy of 95.6% for evaluating with cross-validation and 98.7% for examining the performance with the independent data set. For newly discovered transport protein sequences, our approach performed remarkably better than similar methods such as GTPBinder, NsitePred and TargetSOS. Moreover, a friendly web server was developed for identifying GTP binding sites in transport proteins available for all users. CONCLUSIONS: We approached a computational technique using PSSM profiles and SAAPs for identifying GTP binding residues in transport proteins. When we included SAAPs into PSSM profiles, the predictive performance achieved a significant improvement in all measurement metrics. Furthermore, the proposed method could be a power tool for determining new proteins that belongs into GTP binding sites in transport proteins and can provide useful information for biologists.

Membrane proteins perform diverse functions in living organisms such as transporters, receptors and channels. The functions of membrane proteins have been investigated with several computational approaches, such as developing databases, analyzing the structure-function relationship and establishing algorithms to discriminate different type of membrane proteins. However, compilation of bioinformatics resources for the functions of membrane proteins is not well documented compared with their structural aspects. In this comprehensive review, we elaborately focus on three aspects of membrane protein functions: (i) databases for different types of membrane proteins based on their functions including transporters, receptors and ion channels, annotated functional data for genomes, as well as functionally important amino acid residues in membrane proteins obtained from experimental data, (ii) analysis of membrane protein functions based on their structures, motifs, amino acid properties and other features and (iii) algorithms for discriminating different types of membrane proteins and annotating them in genomic sequences. In addition, we provide a list of online resources for the databases and web servers for functional annotation of membrane proteins.

SUMMARY: Transporters are proteins that are involved in the movement of ions or molecules across biological membranes. Currently, our knowledge about the functions of transporters is limited due to the paucity of their 3D structures. Hence, computational techniques are necessary to annotate the functions of transporters. In this work, we focused on an important functional aspect of transporters, namely annotation of targets for transport proteins. We have systematically analyzed four major classes of transporters with different transporter targets: (i) electron, (ii) protein/mRNA, (iii) ion and (iv) others, using amino acid properties. We have developed a radial basis function network-based method for predicting transport targets with amino acid properties and position specific scoring matrix profiles. Our method showed a 10-fold cross-validation accuracy of 90.1, 80.1, 70.3 and 82.3% for electron transporters, protein/mRNA transporters, ion transporters and others, respectively, in a dataset of 543 transporters. We have also evaluated the performance of the method with an independent dataset of 108 proteins and we obtained similar accuracy. We suggest that our method could be an effective tool for functional annotation of transport proteins. AVAILABILITY: http://rbf.bioinfo.tw/~sachen/ttrbf.html

BACKGROUND: While occurring enzymatically in biological systems, O-linked glycosylation affects protein folding, localization and trafficking, protein solubility, antigenicity, biological activity, as well as cell-cell interactions on membrane proteins. Catalytic enzymes involve glycotransferases, sugar-transferring enzymes and glycosidases which trim specific monosaccharides from precursors to form intermediate structures. Due to the difficulty of experimental identification, several works have used computational methods to identify glycosylation sites. RESULTS: By investigating glycosylated sites that contain various motifs between Transmembrane (TM) and non-Transmembrane (non-TM) proteins, this work presents a novel method, GlycoRBF, that implements radial basis function (RBF) networks with significant amino acid pairs (SAAPs) for identifying O-linked glycosylated serine and threonine on TM proteins and non-TM proteins. Additionally, a membrane topology is considered for reducing the false positives on glycosylated TM proteins. Based on an evaluation using five-fold cross-validation, the consideration of a membrane topology can reduce 31.4% of the false positives when identifying O-linked glycosylation sites on TM proteins. Via an independent test, GlycoRBF outperforms previous O-linked glycosylation site prediction schemes. CONCLUSION: A case study of Cyclic AMP-dependent transcription factor ATF-6 alpha was presented to demonstrate the effectiveness of GlycoRBF. Web-based GlycoRBF, which can be accessed at http://GlycoRBF.bioinfo.tw, can identify O-linked glycosylated serine and threonine effectively and efficiently. Moreover, the structural topology of Transmembrane (TM) proteins with glycosylation sites is provided to users. The stand-alone version of GlycoRBF is also available for high throughput data analysis.

How to handle unbalanced datasets and how to handle high-dimensional datasets are two of the most challenging issues faced by the latest machine learning research. This article reports a study aimed at providing effective solutions to these two challenges. For handling unbalanced datasets, we proposed that a different value of the cost parameter in Support Vector Machine (SVM) is employed for each class of samples. For handling high-dimensional datasets, we resorted to Independent Components Analysis (ICA), which is a multivariate analysis algorithm, along with the conventional univariate analysis. Experimental results confirmed that the proposed approaches all together significantly improved the prediction accuracy delivered by SVM.

BACKGROUND: More and more disordered regions have been discovered in protein sequences, and many of them are found to be functionally significant. Previous studies reveal that disordered regions of a protein can be predicted by its primary structure, the amino acid sequence. One observation that has been widely accepted is that ordered regions usually have compositional bias toward hydrophobic amino acids, and disordered regions are toward charged amino acids. Recent studies further show that employing evolutionary information such as position specific scoring matrices (PSSMs) improves the prediction accuracy of protein disorder. As more and more machine learning techniques have been introduced to protein disorder detection, extracting more useful features with biological insights attracts more attention. RESULTS: This paper first studies the effect of a condensed position specific scoring matrix with respect to physicochemical properties (PSSMP) on the prediction accuracy, where the PSSMP is derived by merging several amino acid columns of a PSSM belonging to a certain property into a single column. Next, we decompose each conventional physicochemical property of amino acids into two disjoint groups which have a propensity for order and disorder respectively, and show by experiments that some of the new properties perform better than their parent properties in predicting protein disorder. In order to get an effective and compact feature set on this problem, we propose a hybrid feature selection method that inherits the efficiency of uni-variant analysis and the effectiveness of the stepwise feature selection that explores combinations of multiple features. The experimental results show that the selected feature set improves the performance of a classifier built with Radial Basis Function Networks (RBFN) in comparison with the feature set constructed with PSSMs or PSSMPs that adopt simply the conventional physicochemical properties. CONCLUSION: Distinguishing disordered regions from ordered regions in protein sequences facilitates the exploration of protein structures and functions. Results based on independent testing data reveal that the proposed predicting model DisPSSMP performs the best among several of the existing packages doing similar tasks, without either under-predicting or over-predicting the disordered regions. Furthermore, the selected properties are demonstrated to be useful in finding discriminating patterns for order/disorder classification.

This paper proposes a novel method to construct a radial basis function network (RBFN) classifier. Our contribution consists of two parts. The first one is an incremental hierarchical clustering algorithm for constructing the hidden layer, and the second one is to improve the least mean square error method that calculates the weights between the hidden and the output layers of an RBFN. This paper discusses the effects of incorporating an incremental hierarchical clustering algorithm for constructing an RBFN optimized for data classification applications. The formation of clusters is controlled by the class labels of training samples and therefore the clusters identified are well adapted to the local distributions of training instances. In addition, the incremental framework largely reduces the requirement of memory space when the training data set is large. In regard to the calculation of weights, we employ the regularization theory to solve the singular matrix problem that might happen in determining the optimal weights. Experimental results show that the data classifier constructed is capable of delivering comparable classification accuracy as the support vector machine (SVM) and the kernel density estimation based classifier that we have recently proposed, while enjoying significant execution efficiency in handling data sets that contains a high percentage of redundant training instances.

In recent years, kernel density estimation has been exploited by computer scientists to model several important problems in machine learning, bioinformatics, and computer vision. However, in case the dimension of the data set is high, then the conventional kernel density estimators suffer poor convergence rates of the pointwise mean square error (MSE) and the integrated mean square error (IMSE). Therefore, design of a novel kernel density estimator that overcomes this problem has been a great challenge for many years. This paper proposes a relaxed model of the variable kernel density estimation and analyzes its performance in data classification applications. It is proved in this paper that, in terms of pointwise MSE, the convergence rate of the relaxed variable kernel density estimator can approach O(n/sup -1/) regardless of the dimension of the data set, where n is the number of sampling instances. Experiments with the data classification applications have shown that the improved convergence rate of the pointwise MSE leads to higher prediction accuracy. In fact, the experimental results have also shown that the data classifier constructed based on the relaxed variable kernel density estimator is capable of delivering the same level of prediction accuracy as the SVM with the Gaussian kernel.

This paper presents a novel learning algorithm for efficient construction of the radial basis function (RBF) networks that can deliver the same level of accuracy as the support vector machines (SVMs) in data classification applications. The proposed learning algorithm works by constructing one RBF subnetwork to approximate the probability density function of each class of objects in the training data set. With respect to algorithm design, the main distinction of the proposed learning algorithm is the novel kernel density estimation algorithm that features an average time complexity of O(n log n), where n is the number of samples in the training data set. One important advantage of the proposed learning algorithm, in comparison with the SVM, is that the proposed learning algorithm generally takes far less time to construct a data classifier with an optimized parameter setting. This feature is of significance for many contemporary applications, in particular, for those applications in which new objects are continuously added into an already large database. Another desirable feature of the proposed learning algorithm is that the RBF networks constructed are capable of carrying out data classification with more than two classes of objects in one single run. In other words, unlike with the SVM, there is no need to resort to mechanisms such as one-against-one or one-against-all for handling datasets with more than two classes of objects. The comparison with SVM is of particular interest, because it has been shown in a number of recent studies that SVM generally are able to deliver higher classification accuracy than the other existing data classification algorithms. As the proposed learning algorithm is instance-based, the data reduction issue is also addressed in this paper. One interesting observation in this regard is that, for all three data sets used in data reduction experiments, the number of training samples remaining after a naive data reduction mechanism is applied is quite close to the number of support vectors identified by the SVM software. This paper also compares the performance of the RBF networks constructed with the proposed learning algorithm and those constructed with a conventional cluster-based learning algorithm. The most interesting observation learned is that, with respect to data classification, the distributions of training samples near the boundaries between different classes of objects carry more crucial information than the distributions of samples in the inner parts of the clusters.

In recent years, Support Vector Machines (SVM) have been extensively applied to deal with various data classification problems. However, in some cases, the application of SVM is limited due to the time taken to conduct model selection for SVM. This issue is of particular significant for some modern applications, such as web mining, in which the large-scale database is frequently updated. This paper proposes a data reduction based mechanism aimed at expediting the model selection process in SVM. Experimental results show that the proposed mechanism is able to greatly reduce the time taken to carry out model selection at minimum cost.