Search

Dipeptidyl peptidase-IV (DPP-IV) inhibitory peptides play a crucial role in drug development and the treatment of diabetes. Thus, it is an urgent task to fast and precise distinguishing DPP-IV inhibitory peptides from non-DPP-IV inhibitory peptides. This study developed a support vector machine (SVM) based model to accurately identify DPP-IV inhibitory peptides. Specifically, the peptide sequences were firstly encoded by fifty kinds of physicochemical properties, and dynamic time warping algorithm was introduced to capture the correlation information of distinct physicochemical properties of amino acids. To further remove the effect of noise, orthogonal minimum spanning tree algorithm was proposed. Finally, the features were inputted into SVM to discriminate DPP-IV from non-DPP-IV inhibitory peptides. In the jackknife test, our proposed method achieved 86.28% and 87.97% classification accuracies on benchmark and independent datasets, respectively. The experimental results showed that the proposed method achieved significant improvement in classification performance, as compared with the existing method. The datasets and code are publicly available at https://figshare.com/articles/online_resource/iDPPIV/14769174 .

As an essential post-transcriptional modification, N7-methylguanosine (m7G) regulates nearly every step of the life cycle of mRNA. Accurate identification of the m7G site in the transcriptome will provide insights into its biological functions and mechanisms. Although the m7G-methylated RNA immunoprecipitation sequencing (MeRIP-seq) method has been proposed in this regard, it is still cost-ineffective for detecting the m7G site. Therefore, it is urgent to develop new methods to identify the m7G site. In this work, we developed the first computational predictor called iRNA-m7G to identify m7G sites in the human transcriptome. The feature fusion strategy was used to integrate both sequence- and structure-based features. In the jackknife test, iRNA-m7G obtained an accuracy of 89.88%. The superiority of iRNA-m7G for identifying m7G sites was also demonstrated by comparing with other methods. We hope that iRNA-m7G can become a useful tool to identify m7G sites. A user-friendly web server for iRNA-m7G is freely accessible at http://lin-group.cn/server/iRNA-m7G/. Keywords: N7-methylguanosine, nucleotide chemical property, RNA secondary structure, pseudo nucleotide composition, feature fusion

Antimicrobial peptides are crucial components of the innate host defense system of most living organisms and promising candidates for antimicrobial agents. Accurate classification of antimicrobial peptides will be helpful to the discovery of new therapeutic targets. In this work, the Increment of Diversity with Quadratic Discriminant analysis (IDQD) was presented to classify antifungal and antibacterial peptides based on primary sequence information. In the jackknife test, the proposed IDQD model yields an accuracy of 86.02 percent with the sensitivity of 74.31 percent and specificity of 92.79 percent for identifying antimicrobial peptides, which is superior to other state-of-the-art methods. This result suggests that the proposed IDQD model can be efficiently used to antimicrobial peptide classification.

Due to the overuse of antibiotics, people are worried that existing antibiotics will become ineffective against pathogens with the rapid rise of antibiotic-resistant strains. The use of cell wall lytic enzymes to destroy bacteria has become a viable alternative to avoid the crisis of antimicrobial resistance. In this paper, an improved method for cell wall lytic enzymes prediction was proposed and the amino acid composition (AAC), the dipeptide composition (DC), the position-specific score matrix auto-covariance (PSSM-AC), and the auto-covariance average chemical shift (acACS) were selected to predict the cell wall lytic enzymes with support vector machine (SVM). In order to overcome the imbalanced data classification problems and remove redundant or irrelevant features, the synthetic minority over-sampling technique (SMOTE) was used to balance the dataset. The F-score was used to select features. The Sn, Sp, MCC, and Acc were 99.35%, 99.02%, 0.98, and 99.19% with jackknife test using the optimized combination feature AAC+DC+acACS+PSSM-AC. The Sn, Sp, MCC, and Acc of cell wall lytic enzymes in our predictive model were higher than those in existing methods. This improved method may be helpful for protein function prediction.

Antioxidant proteins play important roles in preventing free radical oxidation from damaging cells and DNA. They have become ideal candidates of disease prevention and treatment. Therefore, it is urgent to identify antioxidants from natural compounds. Since experimental methods are still cost ineffective, a series of computational methods have been proposed to identify antioxidant proteins. However, the performance of the current methods are still not satisfactory. In this study, a support vector machine based method, called Vote9, was proposed to identify antioxidants, in which the sequences were encoded by using the features generated from 9 optimal individual models. Results from jackknife test demonstrated that Vote9 is comparable with the best one of the existing predictors for this task. We hope that Vote9 will become a useful tool or at least can play a complementary role to the existing methods for identifying antioxidants.

MotivationDNA N6-methyladenine (6mA) is associated with a wide range of biological processes. Since the distribution of 6mA site in the genome is non-random, accurate identification of 6mA sites is crucial for understanding its biological functions. Although experimental methods have been proposed for this regard, they are still cost-ineffective for detecting 6mA site in genome-wide scope. Therefore, it is desirable to develop computational methods to facilitate the identification of 6mA site.ResultsIn this study, a computational method called i6mA-Pred was developed to identify 6mA sites in the rice genome, in which the optimal nucleotide chemical properties obtained by the using feature selection technique were used to encode the DNA sequences. It was observed that the i6mA-Pred yielded an accuracy of 83.13% in the jackknife test. Meanwhile, the performance of i6mA-Pred was also superior to other methods.Availability and implementationA user-friendly web-server, i6mA-Pred is freely accessible at http://lin-group.cn/server/i6mA-Pred.

The identification of bacteriophage proteins in the host subcellular localization could provide important clues for understanding the interaction between phage and host bacteria as well as antibacterial drug design. To date, computational methods have been reported to identify bacteriophage proteins located in the host cell. However, there is still space for improving the prediction accuracy. The existing methods considering the sequence order correlation and the physicochemical property of protein provide us insights to construct an integrated descriptor based on sequence for phage proteins. Meanwhile, we proposed a feature selection technique to obtain the optimal features. In the jackknife test, the prediction accuracies are 86.7% and 97.9%, respectively for discrimination between PH proteins and non-PH proteins as well as PHM proteins and PHC proteins. Based on our model, we updated the web server PHPred to version 2.0 which can be freely accessed from http://lin-group.cn/server/PHPred2.0 .

Mitochondrion is important organelle of most eukaryotes and play an important role in participating in various life activities of cells. However, some functions of mitochondria can only be achieved in specific submitochondrial location, the study of submitochondrial locations will help to further understand the biological function of protein, which is a hotspot in proteomics research. In this paper, we propose a new method for protein submitochondrial locations prediction. Firstly, the features of protein sequence are extracted by combining Chou's pseudo-amino acid composition (PseAAC) and pseudo-position specific scoring matrix (PsePSSM). Then the extracted feature information is denoised by two-dimensional (2-D) wavelet denoising. Finally, the optimal feature vectors are input to the SVM classifier to predict the protein submitochondrial locations. We obtained the ideal prediction results by jackknife test and compared with other prediction methods. The results indicate that the proposed method is significantly better than the existing research results, which can provide a new method to predict protein locations in other organelles. The source code and all datasets are available at https://github.com/QUST-BSBRC/PseAAC-PsePSSM-WD/ for academic use.

Presynaptic neurotoxins and postsynaptic neurotoxins are two important neurotoxins isolated from venoms of venomous animals and have been proven to be potential effective in neurosciences and pharmacology. With the number of toxin sequences appeared in the public databases, there was a need for developing a computational method for fast and accurate identification and classification of the novel presynaptic neurotoxins and postsynaptic neurotoxins in the large databases. In this study, the Multinomial Naive Bayes Classifier (MNBC) had been developed to discriminate the presynaptic neurotoxins and postsynaptic neurotoxins based on the different kinds of features. The Minimum Redundancy Maximum Relevance (MRMR) feature selection method was used for ranking 400 pseudo amino acid (PseAA) compositions and 50 top ranked PseAA compositions were selected for improving the prediction results. The motif features, 400 PseAA compositions and 50 PseAA compositions were combined together, and selected as the input parameters of MNBC. The best correlation coefficient (CC) value of 0.8213 was obtained when the prediction quality was evaluated by the jackknife test. It was anticipated that the algorithm presented in this study may become a useful tool for identification of presynaptic neurotoxin and postsynaptic neurotoxin sequences and may provide some useful help for in-depth investigation into the biological mechanism of presynaptic neurotoxins and postsynaptic neurotoxins.

Human immunodeficiency virus (HIV) is the retroviral agent that causes acquired immune deficiency syndrome (AIDS). The number of HIV caused deaths was about 4 million in 2016 alone; it was estimated that about 33 million to 46 million people worldwide living with HIV. The HIV disease is especially harmful because the progressive destruction of the immune system prevents the ability of forming specific antibodies and to maintain an efficacious killer T cell activity. Successful prediction of HIV protein has important significance for the biological and pharmacological functions. In this study, based on the concept of Chou’s pseudo amino acid (PseAA) composition and increment of diversity (ID), support vector machine (SVM), logisitic regression (LR), and multilayer perceptron (MP) were presented to predict HIV-1 proteins and HIV-2 proteins. The results of the jackknife test indicated that the highest prediction accuracy and CC values were obtained by the SVM and MP were 0.9909 and 0.9763, respectively, indicating that the classifiers presented in this study were suitable for predicting two groups of HIV proteins.

Aims and Objective: Chemical toxicity effect is one of the major reasons for declining candidate drugs. Detecting the toxicity effects of all chemicals can accelerate the procedures of drug discovery. However, it is time-consuming and expensive to identify the toxicity effects of a given chemical through traditional experiments. Designing quick, reliable and non-animal-involved computational methods is an alternative way. Method: In this study, a novel integrated multi-label classifier was proposed. First, based on five types of chemical-chemical interactions retrieved from STITCH, each of which is derived from one aspect of chemicals, five individual classifiers were built. Then, several integrated classifiers were built by integrating some or all individual classifiers. Result and Conclusion: By testing the integrated classifiers on a dataset with chemicals and their toxicity effects in Accelrys Toxicity database and non-toxic chemicals with their performance evaluated by jackknife test, an optimal integrated classifier was selected as the proposed classifier, which provided quite high prediction accuracies and wide applications.

As one of the most important and common histones post-translational modifications, crotonylation plays a key role in regulating various biological processes. The accurate identification of crotonylation sites is crucial to elucidate the underlying molecular mechanisms of crotonylation. In this study, a novel bioinformatics tool named CKSAAP_CrotSite is developed to predict crotonylation sites. The highlight of CKSAAP_CrotSite is to adopt the composition of k-spaced amino acid pairs as input encoding, and the support vector machine is employed as the classifier. As illustrated by jackknife test, CKSAAP_CrotSite achieves a promising performance with a Sensitivity of 92.45%, a Specificity of 99.17%, an Accuracy of 98.11% and a Matthew's correlation coefficient of 0.9283, which is much better than those of the existing prediction methods. Feature analysis shows that some amino acid pairs such as 'KxG', 'KG' and 'PxP' may play an important role in the prediction of crotonylation sites. The results of analysis and prediction could offer useful information for elucidating the molecular mechanisms of crotonylation and related experimental validations. A user-friendly web-server for CKSAAP_CrotSite is available at 123.206.31.171/CKSAAP_CrotSite/.

DNase I Hypersensitive sites (DHS) are the regions that are sensitive to cleavage by the DNase I enzyme. Knowledge regarding these sites is helpful for decryption of the functions of non-coding genomic regions. Various biological processes need its intervention. Traditional techniques are laborious and time-consuming to predict DHS sites. Particularly, with the avalanche of DNA sequences generated in the post-genomic era, the development of computational approaches is highly essential to precisely and timely predict DHS sites in DNA sequences. The existing feature encoding schemes such as pseudo dinucleotide composition, pseudo trinucleotide composition etc. cannot effectively express features from DHS sequences. In the current study, we proposed a new computational technique to predict DHS sites which uses Un-biased Pseudo Trinucleotide Composition (Unb-PseTNC) strategy to extract nominal descriptors from the DHS benchmark dataset and avoid biasness among the classes during the classification phase. Several classification algorithm including Support vector machine (SVM), probabilistic neural network and k-nearest neighbor are employed to classify extracted features. It was observed that SVM in conjunction with Unb-PseTNC outperforms other techniques. By comparing with other existing predictors, it was perceived that our proposed method achieved higher prediction rates by applying rigorous jackknife test. This indicates that the proposed model will become a useful tool to predict DHS sites and can also be utilized for in-depth study of DNA and genome analysis.

Proteins are the central constitute of a cell or biological system. Proteins execute their functions by interacting with other molecules such as RNA, DNA and other proteins. The major functionality of protein-protein interactions (PPIs) is the execution of biochemical activities in living species. Therefore, an accurate identification of PPIs becomes a challenging and demanding task for investigators from last few decades. Various traditional and computational methods have been applied but they have not achieved quite encouraging results. In order to extend the concept of computational model by incorporating intelligent, contemporary machine learning algorithms have been utilized for identification of PPIs. In this prediction model, protein sequences are expressed by using two distinct feature extraction methods namely: physiochemical properties of amino acids and evolutionary profiles method position specific scoring matrix (PSSM). Jackknife test and numerous performance parameters namely: specificity, recall, accuracy, MCC, precision, and F-measure were employed to compute the predictive quality of proposed model. After empirical analysis, it is determined that the proposed prediction model yielded encouraging predictive outcomes compared to existing state-of-the-art models. This achievement is ascribed with PSSM because it has clearly discerned a motif of PPIs. It is realized that the proposed prediction model will lead to be a practical and very useful tool for research community.

Apoptosis is a fundamental process controlling normal tissue homeostasis by regulating a balance between cell proliferation and death. Predicting subcellular location of apoptosis proteins is very helpful for understanding its mechanism of programmed cell death. Prediction of apoptosis protein subcellular location is still a challenging and complicated task, and existing methods mainly based on protein primary sequences. In this paper, we propose a new position-specific scoring matrix (PSSM)-based model by using Geary autocorrelation function and detrended cross-correlation coefficient (DCCA coefficient). Then a 270-dimensional (270D) feature vector is constructed on three widely used datasets: ZD98, ZW225 and CL317, and support vector machine is adopted as classifier. The overall prediction accuracies are significantly improved by rigorous jackknife test. The results show that our model offers a reliable and effective PSSM-based tool for prediction of apoptosis protein subcellular localization.

Noncoding RNAs (ncRNAs) are implicated in various biological processes. Recent findings have demonstrated that the function of ncRNAs correlates with their provenance. Therefore, the recognition of ncRNAs from different organelle genomes will be helpful to understand their molecular functions. However, the weakness of experimental techniques limits the progress toward studying organellar ncRNAs and their functional relevance. As a complement of experiments, computational method provides an important choice to identify ncRNA in different organelles. Thus, a computational model was developed to identify ncRNAs from kinetoplast and mitochondrion organelle genomes. In this model, RNA sequences are encoded by "pseudo dinucleotide composition." It was observed by the jackknife test that the overall success rate achieved by the proposed model was 90.08 %. We hope that the proposed method will be helpful in predicting ncRNA organellar locations.

N6-methyladenosine (m(6)A) is present ubiquitously in the RNA of living organisms from Escherichia coli to humans. Nonetheless, the exact molecular mechanism of this modification remains unclear. The experimental identification of m(6)A modification is time-consuming and expensive; therefore, bioinformatics tools with high accuracy represent desirable alternatives for the large-scale, rapid identification of N6-methyladenosine sites. In this study, RNA-MethylPred, a new bioinformatics model, was developed by incorporating bi-profile Bayes, dinucleotide composition, and k nearest neighbor (KNN) scores for three feature extractions. RNA-MethylPred yielded a Matthew's correlation coefficient (MCC) of 0.53 in a jackknife test, which was 0.24 higher than that of iRNA-Methyl and 0.13 higher than that of pRNAm-PC. The obvious improvements demonstrated that RNA-MethylPred might be a powerful and complementary tool for further experimental investigation of N6-methyladenosine modification.

N 6-Methyladenosine (m6A) plays important roles in many biological processes. The knowledge of the distribution of m6A is helpful for understanding its regulatory roles. Although the experimental methods have been proposed to detect m6A, the resolutions of these methods are still unsatisfying especially for Arabidopsis thaliana. Benefitting from the experimental data, in the current work, a support vector machine-based method was proposed to identify m6A sites in A. thaliana transcriptome. The proposed method was validated on a benchmark dataset using jackknife test and was also validated by identifying strain-specific m6A sites in A. thaliana. The obtained predictive results indicate that the proposed method is quite promising. For the convenience of experimental biologists, an online webserver for the proposed method was built, which is freely available at http://lin.uestc.edu.cn/server/M6ATH . These results indicate that the proposed method holds a potential to become an elegant tool in identifying m6A site in A. thaliana.

Apoptosis proteins are related to many diseases. Obtaining the subcellular localization information of apoptosis proteins is helpful to understand the mechanism of diseases and to develop new drugs. At present, the researchers mainly focus on the primary protein sequences, so there is still room for improvement in the prediction accuracy of the subcellular localization of apoptosis proteins. In this paper, a new method named ERT-ECT-PSSM-IS is proposed to predict apoptosis proteins based on the position-specific scoring matrix (PSSM). First, the local and global features of different directions are extracted by evolutionary row transformation (ERT) and cross-covariance of evolutionary column transformation (ECT) based on PSSM (ERT-ECT-PSSM). Second, an improved isometric mapping algorithm (I-SMA) is used to eliminate redundant features. Finally, we adopt a support vector machine (SVM) to classify our results, and the prediction accuracy is evaluated by jackknife cross-validation tests. The experimental results show that the proposed method not only extracts more abundant feature expression but also has better predictive performance and robustness for the subcellular localization of apoptosis proteins in ZD98, ZW225, and CL317 databases. Graphical abstract Framework of the proposed prediction model.

Antimicrobial peptides (AMPs) play an important role in the innate immune system that evolved in most living organisms. As a kind of natural antibiotics, it is promising for solving the problem of increasing antibiotic resistance. In view of this, it is highly desired to develop a fast and effective computational method for accurately predicting the functional types of AMPs, because the biological functions of AMPs are correlated with the type it belongs to. Although many efforts have been made in this area, to the best of our knowledge, most of the existing predictors only has the ability to deal with whether a peptide is an AMP or not, or a peptide belongs to which one type. However, there are many AMPs have two or more functional types, the phenomenon should worthy of our special notice, because they may have some unique biological functions for new drug design and disease treatment. In this study, in order to reflect the characteristic of multiplex AMPs, a new multi-label classifier based on sequence information and multi-label learning with label-specific features (LIFT) algorithm was developed. It was observed that, the absolute-true with jackknife test by the new predictor on a newly stringent benchmark dataset is 0.5040, and the success rates achieved by the new predictor are 5 % higher than this by iAMP-2L in the same dataset, indicating that our method is quite promising. We hope that the predictor may become a useful high-through tool in identifying the functional types of AMPs.

Conotoxins are a group of high specialized and functionally diverse peptides. Because the conotoxins are selectivity for membrane receptors and ion channels, so they can make valuable biological probes and drug targets. Successful prediction of the conotoxin superfamily peptide has important biological meaning in the pharmacology of the neurotoxins. In this work, based on the concept that the function of toxin protein is determined by its protein sequence, the Naive Bayes classifier and feature selection method are proposed to predict the conotoxin superfamilies. The obtained results of the jackknife test indicated that the overall prediction accuracy is 84.92% for the dataset with 305 conotoxins. This algorithm was also used to predict the dataset with 116 conotoxins and 60 non-conotoxins, the higher predictive rates than some previous studies are obtained in our study.

Membrane proteins were found to be involved in various cellular processes performing various important functions, which are mainly associated to their type. Given a membrane protein sequence, how can we identify its type(s)? Particularly, how can we deal with the multi-type problem since one membrane protein may simultaneously belong to two or more different types? To address these problems, which are obviously very important to both basic research and drug development, a new multi-label classifier was developed based on pseudo amino acid composition with multi-label k-nearest neighbor algorithm. The success rate achieved by the new predictor on the benchmark dataset by jackknife test is 73.94 %, indicating that the method is promising and the predictor may become a very useful high-throughput tool, or at least play a complementary role to the existing predictors in identifying functional types of membrane proteins.

Voltage-gated K+ channel (VKC) plays important roles in biology procession, especially in nervous system. Different subfamilies of VKCs have different biological functions. Thus, knowing VKCs’ subfamilies has become a meaningful job because it can guide the direction for the disease diagnosis and drug design. However, the traditional wet-experimental methods were costly and time-consuming. It is highly desirable to develop an effective and powerful computational tool for identifying different subfamilies of VKCs. In this study, a predictor, called iVKC-OTC, has been developed by incorporating the optimized tripeptide composition (OTC) generated by feature selection technique into the general form of pseudo-amino acid composition to identify six subfamilies of VKCs. One of the remarkable advantages of introducing the optimized tripeptide composition is being able to avoid the notorious dimension disaster or over fitting problems in statistical predictions. It was observed on a benchmark dataset, by using a jackknife test, that the overall accuracy achieved by iVKC-OTC reaches to 96.77% in identifying the six subfamilies of VKCs, indicating that the new predictor is promising or at least may become a complementary tool to the existing methods in this area. It has not escaped our notice that the optimized tripeptide composition can also be used to investigate other protein classification problems.

DNA N6-methyladenine (6mA) plays an important role in regulating the gene expression of eukaryotes. Accurate identification of 6mA sites may assist in understanding genomic 6mA distributions and biological functions. Various experimental methods have been applied to detect 6mA sites in a genome-wide scope, but they are too time-consuming and expensive. Developing computational methods to rapidly identify 6mA sites is needed. In this paper, a new machine learning-based method, i6mA-DNCP, was proposed for identifying 6mA sites in the rice genome. Dinucleotide composition and dinucleotide-based DNA properties were first employed to represent DNA sequences. After a specially designed DNA property selection process, a bagging classifier was used to build the prediction model. The jackknife test on a benchmark dataset demonstrated that i6mA-DNCP could obtain 84.43% sensitivity, 88.86% specificity, 86.65% accuracy, a 0.734 Matthew&rsquo, s correlation coefficient (MCC), and a 0.926 area under the receiver operating characteristic curve (AUC). Moreover, three independent datasets were established to assess the generalization ability of our method. Extensive experiments validated the effectiveness of i6mA-DNCP.

Membrane proteins play important roles in many biochemical processes and are also attractive targets of drug discovery for various diseases. The elucidation of membrane protein types provides clues for understanding the structure and function of proteins. Recently we developed a novel system for predicting protein subnuclear localizations. In this paper, we propose a simplified version of our system for predicting membrane protein types directly from primary protein structures, which incorporates amino acid classifications and physicochemical properties into a general form of pseudo-amino acid composition. In this simplified system, we will design a two-stage multi-class support vector machine combined with a two-step optimal feature selection process, which proves very effective in our experiments. The performance of the present method is evaluated on two benchmark datasets consisting of five types of membrane proteins. The overall accuracies of prediction for five types are 93.25% and 96.61% via the jackknife test and independent dataset test, respectively. These results indicate that our method is effective and valuable for predicting membrane protein types. A web server for the proposed method is available at http://www.juemengt.com/jcc/memty_page.php.

Drug-target interaction is a key research topic in drug discovery since correct identification of target proteins of drug candidates can help screen out those with unacceptable toxicities, thereby saving expense. In this study, we developed a novel computational approach to predict drug target groups that may reduce the number of candidate target proteins associated with a query drug. A benchmark dataset, consisting of 3028 drugs assigned within nine categories, was constructed by collecting data from KEGG. The nine categories are (1) G protein-coupled receptors, (2) cytokine receptors, (3) nuclear receptors, (4) ion channels, (5) transporters, (6) enzymes, (7) protein kinases, (8) cellular antigens and (9) pathogens. The proposed method combines the data gleaned from chemical-chemical similarities, chemical-chemical connections and chemical-protein connections to allocate drugs to each of the nine target groups. A jackknife test applied to the training dataset that was constructed from the benchmark dataset, provided an overall correct prediction rate of 87.45%, as compared to 87.79% for the test dataset that was constructed by randomly selecting 10% of samples from the benchmark dataset. These prediction rates are much higher than the 11.11% achieved by random guesswork. These promising results suggest that the proposed method can become a useful tool in identifying drug target groups. This article is part of a Special Issue entitled: Computational Proteomics, Systems Biology & Clinical Implications. Guest Editor: Yudong Cai.

Apoptosis proteins play an important role in the mechanism of programmed cell death. Predicting subcellular localization of apoptosis proteins is an essential step to understand their functions and identify drugs target. Many computational prediction methods have been developed for apoptosis protein subcellular localization. However, these existing works only focus on the proteins that have one location; proteins with multiple locations are either not considered or assumed as not existing when constructing prediction models, so that they cannot completely predict all the locations of the apoptosis proteins with multiple locations. To address this problem, this paper proposes a novel multilabel predictor named MultiP-Apo, which can predict not only apoptosis proteins with single subcellular location but also those with multiple subcellular locations. Specifically, given a query protein, GO-based feature extraction method is used to extract its feature vector. Subsequently, the GO feature vector is classified by a new multilabel classifier based on the label-specific features. It is the first multilabel predictor ever established for identifying subcellular locations of multilocation apoptosis proteins. As an initial study, MultiP-Apo achieves an overall accuracy of 58.49% by jackknife test, which indicates that our proposed predictor may become a very useful high-throughput tool in this area.

Knowledge of the mechanism of HIV protease cleavage specificity is critical to the design of specific and ef- fective HIV inhibitors. Searching for an accurate, robust, and rapid method to correctly predict the cleavage sites in pro- teins is crucial when searching for possible HIV inhibitors. In this article, HIV-1 protease specificity was studied using the correlation-based feature subset (CfsSubset) selection method combined with Genetic Algorithms method. Thirty impor- tant biochemical features were found based on a jackknife test from the original data set containing 4,248 features. By us- ing the AdaBoost method with the thirty selected features the prediction model yields an accuracy of 96.7% for the jack- knife test and 92.1% for an independent set test, with increased accuracy over the original dataset by 6.7% and 77.4%, re- spectively. Our feature selection scheme could be a useful technique for finding effective competitive inhibitors of HIV protease.

Apoptosis proteins are very important for regulating the balance between cell proliferation and death. Because the function of apoptosis protein is closely related to its subcellular location, it is desirable to explore their function by predicting the subcellular location of apoptosis protein. In this paper, based on evolutionary profiles and motifs information of protein sequences, an approach for predicting apoptosis proteins subcellular location is presented by using support vector machine (SVM). When the method is applied to three data sets (98 apoptosis proteins dataset, 225 apoptosis proteins dataset and 317 apoptosis proteins dataset), the overall accuracies of our method on the three data sets reach 95.9%, 86.7% and 91.8% in the jackknife test, respectively. The higher predictive success rates indicate that the proposed method is very useful for apoptosis proteins subcellular localization.

Motivation Previously constructed classifiers in predicting eukaryotic essential genes integrated a variety of features including experimental ones. If we can obtain satisfactory prediction using only nucleotide (sequence) information, it would be more promising. Three groups recently identified essential genes in human cancer cell lines using wet experiments and it provided wonderful opportunity to accomplish our idea. Here we improved the Z curve method into the λ-interval form to denote nucleotide composition and association information and used it to construct the SVM classifying model. Results Our model accurately predicted human gene essentiality with an AUC higher than 0.88 both for 5-fold cross-validation and jackknife tests. These results demonstrated that the essentiality of human genes could be reliably reflected by only sequence information. We re-predicted the negative dataset by our Pheg server and 118 genes were additionally predicted as essential. Among them, 20 were found to be homologues in mouse essential genes, indicating that some of the 118 genes were indeed essential, however previous experiments overlooked them. As the first available server, Pheg could predict essentiality for anonymous gene sequences of human. It is also hoped the λ-interval Z curve method could be effectively extended to classification issues of other DNA elements. Availability and Implementation http://cefg.uestc.edu.cn/Pheg Supplementary information Supplementary data are available at Bioinformatics online.

Voltage-gated ion channel is the molecular target for a broad range of toxins. Voltage-gated ion channel toxins are excellent pharmacological tools in toxicology and neuroscience. They have been used as molecular scaffolding agents, drugs, and insecticides. In this study, voltage-gated calcium, potassium and sodium channel toxins are predicted by the increment of diversity (ID) algorithm. Each protein is represented by 400 pseudo amino acid compositions and 9 MEME motif features. The Maximum Relevance Minimum Redundancy (MRMR) is applied for ranking 400 pseudo amino acid compositions. The results of jackknife test indicate that the best predictive results are obtained when using 50 higher ranked pseudo amino acid compositions and 9 MEME motif features. Based on the predictive results, our results suggest the usefulness and potential of ID algorithm for prediction of voltage-gated ion channel toxins using protein sequence derived information.

Cancer is one of the main causes of threats to human life. Identification of anticancer peptides is important for developing effective anticancer drugs. In this paper, we developed an improved predictor to identify the anticancer peptides. The amino acid composition (AAC), the average chemical shifts (acACS) and the reduced amino acid composition (RAAC) were selected to predict the anticancer peptides by using the support vector machine (SVM). The overall prediction accuracy reaches to 93.61% in jackknife test. The results indicated that the combined parameter was helpful to the prediction for anticancer peptides.

Machine learning is a kind of reliable technology for automated subcellular localization of viral proteins within a host cell or virus-infected cell. One challenge is that the viral protein samples are not only with multiple location sites, but also class-imbalanced. The imbalanced dataset often decreases the prediction performance. In order to accomplish this challenge, this paper proposes a novel approach named imbalance-weighted multi-label K-nearest neighbor to predict viral protein subcellular location with multiple sites. The experimental results by jackknife test indicate that the presented algorithm achieves a better performance than the existing methods and has great potentials in protein science.

Information of protein quaternary structure can help to understand the biological functions of proteins. Because wet-lab experiments are both time-consuming and costly, we adopt a novel computational approach to assign proteins into 10 kinds of quaternary structures. By coding each protein using its biochemical and physicochemical properties, feature selection was carried out using Incremental Feature Selection (IFS) method. The thus obtained optimal feature set consisted of 97 features, with which the prediction model was built. As a result, the overall prediction success rate is 74.90% evaluated by Jackknife test, much higher than the overall correct rate of a random guess 10% (1/10). The further feature analysis indicates that protein secondary structure is the most contributed feature in the prediction of protein quaternary structure.

Based on the concept of the pseudo amino acid composition (PseAAC), protein structural classes are predicted by using an approach of increment of diversity combined with support vector machine (ID-SVM), in which the dipeptide amino acid composition of proteins is used as the source of diversity. Jackknife test shows that total prediction accuracy is 96.6% and higher than that given by other approaches. Besides, the specificity (Sp) and the Matthew’s correlation coefficient (MCC) are also calculated for each protein structural class, the Sp is more than 88%, the MCC is higher than 92%, and the higher MCC and Sp imply that it is credible to use ID-SVM model predicting protein structural class. The results indicate that: ① the choice of the source of diversity is reasonable, ② the predictive performance of IDSVM is excellent, and ③ the amino acid sequences of proteins contain information of protein structural classes.

As an abundant post-transcriptional modification, dihydrouridine (D) has been found in transfer RNA (tRNA) from bacteria, eukaryotes, and archaea. Nonetheless, knowledge of the exact biochemical roles of dihydrouridine in mediating tRNA function is still limited. Accurate identification of the position of D sites is essential for understanding their functions. Therefore, it is desirable to develop novel methods to identify D sites. In this study, an ensemble classifier was proposed for the detection of D modification sites in the Saccharomyces cerevisiae transcriptome by using heterogeneous features. The jackknife test results demonstrate that the proposed predictor is promising for the identification of D modification sites. It is anticipated that the proposed method can be widely used for identifying D modification sites in tRNA.

Apoptosis proteins have a central role in the development and homeostasis of an organism, and their function is related to their types. In this paper, we constructed the character vectors of apoptosis proteins from their amino acid sequences by using the discrete wavelet transform, combined with support vector machine, to predict the type of given apoptosis proteins. For the widely used dataset z98, high success rates were obtained by Jackknife test, and the Matthews correlation coefficients were 0.92, 0.90, 0.81 and 0.80, respectively, which were higher than the other methods on average.

The location of a protein in a cell is closely correlated with its biological function. Based on the concept that the protein subcellular location is mainly determined by its amino acid and pseudo amino acid composition (PseAA), a new algorithm of increment of diversity combined with support vector machine is proposed to predict the protein subcellular location. The subcellular locations of plant and non-plant proteins are investigated by our method. The overall prediction accuracies in jackknife test are 88.3% for the eukaryotic plant proteins and 92.4% for the eukaryotic non-plant proteins, respectively. In order to estimate the effect of the sequence identity on predictive result, the proteins with sequence identity

Protein subcellular localization, which tells where a protein resides in a cell, is an important characteristic of a protein, and relates closely to the function of proteins. The prediction of their subcellular localization plays an important role in the prediction of protein function, genome annotation and drug design. Therefore, it is an important and challenging role to predict subcellular localization using bio-informatics approach. In this paper, a robust predictor, AdaBoost Learner is introduced to predict protein subcellular localization based on its amino acid composition. Jackknife cross-validation and independent dataset test were used to demonstrate that Adaboost is a robust and efficient model in predicting protein subcellular localization. As a result, the correct prediction rates were 74.98% and 80.12% for the Jackknife test and independent dataset test respectively, which are higher than using other existing predictors. An online server for predicting subcellular localization of proteins based on AdaBoost classifier was available on http://chemdata.shu. edu.cn/sl12.

Recently, several efforts have been made in predicting Golgi-resident proteins. However, it is still a challenging task to identify the type of a Golgi-resident protein. Precise prediction of the type of a Golgi-resident protein plays a key role in understanding its molecular functions in various biological processes. In this paper, we proposed to use a mutual information based feature selection scheme with the general form Chou's pseudo-amino acid compositions to predict the Golgi-resident protein types. The positional specific physicochemical properties were applied in the Chou's pseudo-amino acid compositions. We achieved 91.24% prediction accuracy in a jackknife test with 49 selected features. It has the best performance among all the present predictors. This result indicates that our computational model can be useful in identifying Golgi-resident protein types.

Tuberculosis is killing millions of lives every year and on the blacklist of the most appalling public health problems. Recent findings suggest that secretory protein ofMycobacterium tuberculosismay serve the purpose of developing specific vaccines and drugs due to their antigenicity. Responding to global infectious disease, we focused on the identification of secretory proteins inMycobacterium tuberculosis. A novel method calledMycoSecwas designed by incorporatingg-gap dipeptide compositions into pseudo amino acid composition. Analysis of variance-based technique was applied in the process of feature selection and a total of 374 optimal features were obtained and used for constructing the final predicting model. In the jackknife test,MycoSecyielded a good performance with the area under the receiver operating characteristic curve of 0.93, demonstrating that the proposed system is powerful and robust. For user’s convenience, the web server MycoSec was established and an obliging manual on how to use it was provided for getting around any trouble unnecessary.

Prediction of protein accessibility from sequence, as prediction of protein secondary structure is an intermediate step for predicting structures and consequently functions of proteins. Most of the currently used methods are based on single residue prediction, either by statistical means or evolutionary information, and accessibility state of central residue in a window predicted. By expansion of databases of proteins with known 3D structures, we extracted information of pairwise residue types and conformational states of pairs simultaneously. For solving the problem of ambiguity in state prediction by one residue window sliding, we used dynamic programming algorithm to find the path with maximum score. The three state overall per-residue accuracy, Q(3), of this method in a Jackknife test with dataset of known proteins is more than 65% which is an improvement on results of methods based on evolutionary information.

Nucleosome position plays a crucial role in many biological processes including repair, recombination, gene expression and disease development. With the rapidly increasing number of sequenced genomes in post genomic era, computational methods is starting to be used for timely predicting nucleosome position and linker. Correct identification of nucleosome position is the fundamental importance to understand the gene expression. In this article, we introduced a new computational approach called “PNP-DIPseAAC”, which incorporates the physical and chemistry properties of DNA sequences and statistical features based on the Fuzzy KNN (K-Nearest Neighbor) classifier for statistical prediction, for instance, nucleic acid composition, Ziv-Lempel Complexity Factor, the information entropy and related information redundancy and Low-Frequency Fourier Spectrum Components. According to the strict jackknife test that the result for the overall accuracy achieved 98.4% was encouraging, demonstrating that the novel approach can be a helpful supplementary tool for the existence approaches in identifying nucleosome position.

G-protein coupled receptors (GPCRs) are transmembrane proteins which via G-proteins initiate some of the important signaling pathways in a cell and are involved in various physiological processes. Thus, computational prediction and classification of GPCRs can supply significant information for the development of novel drugs in pharmaceutical industry. In this paper, a nearest neighbor method has been introduced to discriminate GPCRs from non-GPCRs and subsequently classify GPCRs at four levels on the basis of amino acid composition and dipeptide composition of proteins. Its performance is evaluated on a non-redundant dataset consisted of 1406 GPCRs for six families and 1406 globular proteins using the jackknife test. The present method based on amino acid composition achieved an overall accuracy of 96.4% and Matthew's correlation coefficient (MCC) of 0.930 for correctly picking out the GPCRs from globular proteins. The overall accuracy and MCC were further enhanced to 99.8% and 0.996 by dipeptide composition-based method. On the other hand, the present method has successfully classified 1406 GPCRs into six families with an overall accuracy of 89.6 and 98.8% using amino acid composition and dipeptide composition, respectively. For the subfamily prediction of 1181 GPCRs of rhodopsin-like family, the present method achieved an overall accuracy of 76.7 and 94.5% based on the amino acid composition and dipeptide composition, respectively. Finally, GPCRs belonging to the amine subfamily and olfactory subfamily of rhodopsin-like family were further analyzed at the type level. The overall accuracy of dipeptide composition-based method for the classification of amine type and olfactory type of GPCRs reached 94.5 and 86.9%, respectively, while the overall accuracy of amino acid composition-based method was very low for both subfamilies. In comparison with existing methods in the literature, the present method also displayed great competitiveness. These results demonstrate the effectiveness of our method on identifying and classifying GPCRs correctly. GPCRsIdentifier, a corresponding stand-alone executable program for GPCR identification and classification was also developed, which can be acquired freely on request from the authors for academic purposes.

Genes that are indispensable for survival are termed essential genes. The analysis and identification of essential genes are very important for understanding the minimal requirements of cellular survival and for practical purposes. Proteins do not exert their function in isolation of one another but rather interact together in PPI networks. A global analysis of protein interaction networks provides an effective way to elucidate the relationships between proteins. With the recent large-scale identifications of essential genes and the production of large amounts of PPIs in humans, we are able to investigate the topological properties and biological properties of essential genes. However, until recently, no one has ever investigated human essential genes using topological and biological properties. In this study, for the first time, 28 topological properties and 22 biological properties were used to investigate the characteristics of essential and non-essential genes in humans. Most of the properties were statistically discriminative between essential and non-essential genes. The F-score was used to estimate the essentiality of each property. The GO-enrichment analysis was performed to investigate the functions of the essential and non-essential genes. Finally, based on the topological features and the biological characteristics, a machine-learning classifier was constructed to predict the essential genes. The results of the jackknife test and 10-fold cross validation test are encouraging, indicating that our classifier is an effective human essential gene discovery method.

Protein signal sequences play a central role in the targeting and translocation of nearly all secreted proteins and many integral membrane proteins in both prokaryotes and eukaryotes. The knowledge of signal sequences has become a crucial tool for pharmaceutical scientists who genetically modify bacteria, plants, and animals to produce effective drugs. However, to effectively use such a tool, the first important thing is to find a fast and effective method to identify the “zipcode” entity; this is also evoked by both the huge amount of unprocessed data available and the industrial need to find more effective vehicles for the production of proteins in recombinant systems. In view of this, a sequence-encoded algorithm was developed to identify the signal sequences and predict their cleavage sites. The rate of correct prediction for 1,939 secretory proteins and 1,440 nonsecretory proteins by self-consistency test is 90.14% and that by jackknife test is 90.13%. The encouraging results indicate that the signal sequences share some common features although they lack similarity in sequence, length, and even composition and that they are predictable to a considerably accurate extent. Proteins 2001;42:136–139. © 2000 Wiley-Liss, Inc.