Search

Determination of protein structural class using a fast and suitable computational method has become a hot issue in protein science. Prediction of protein structural class for low-similarity sequences remains a challenge problem. In this study, a 111-dimensional feature vector is constructed to predict protein structural classes. Among the 111 features, 100 features based on pseudo-position specific scoring matrix (PsePSSM) are selected to reflect the evolutionary information and the sequence-order information, and the other 11 rational features based on predicted protein secondary structure sequences (PSSS) are designed in the previous works. To evaluate the performance of the proposed method (named by PSSS–PsePSSM), jackknife cross-validation tests are performed on three widely used benchmark datasets: 1189, 25PDB and 640. Our method achieves competitive performance on prediction accuracies, especially for the overall prediction accuracies for datasets 1189, 25PDB and 640, which reach 86.6%, 89.5% and 81.0%, respectively. The PSSS–PsePSSM algorithm also outperforms other existing methods, indicating that our proposed method is a cost-effective computational tool for protein structural class prediction. [ABSTRACT FROM AUTHOR]

Therapeutic peptides, as active substances involved in a variety of cell functions in the organism, are essential participants to complete complex physiological activities of the body. Therefo r e, the prediction of therapeutic peptides is essential for researching on peptide-based therapies. The method of using biological experiments is considered to be time-consuming and labor-intensive. As a fast and accurate method, deep learning can process massive amounts of data on therapeutic peptides. In this research, we raise a deep learning model called Pep-CNN to accurately predict therapeutic peptides. Firstly, we represent the features of the peptide sequence based on the sequence position, the physicochemical property, and the evolutionary-derived feature and use the vectors to represent the sequence. After fusing the features, we use the improved classifier of Convolutional Neural Network (imCNN) to classify and predict eight kinds of peptides. The results show that, compared with other models, Pep-CNN can identify peptides more accurately, which is more conductive to the further research of therapeutic peptides by biomedical scientists. The codes and benchmark datasets are accessible at https://github.com/alivelxj/Pep-CNN. • A new model called Pep-CNN was proposed to predict therapeutic peptides. • The different methods are applied to extract features from the dataset. • An improved convolutional neural network is used to classify the model. [ABSTRACT FROM AUTHOR]

Pseudouridine (Ψ) is one of the most abundant RNA modifications existing in ubiquitous organisms and participates a bunch of biological processes. Identification of pseudouridine sites has significant meanings for the study of biological process and drug developments related to RNA. Wet experiments for pseudouridine sites are expensive, time-consuming and easily influenced by the environment although they can detect Ψ sites in whole transcriptome. While some computational methods have been presented, their performances are still unsatisfactory and their computation costs are usually large. In this article, we propose an incremental identification method called PA-PseU which is based on Passive-Aggressive algorithm as Ψ sites classifier. The combination of chi-square test and logistic regression is used to select optimal feature subsets. Its effectiveness has been demonstrated via 10-fold cross-validation, jack-knife test and independent test. The average leave-one-out accuracies of PA-PseU among different species are respectively 87.80 % in cross validation sets and 86.70 % in independent testing sets, which improve 16.40 % and 12.00 % respectively, contrasting with 71.40 % and 74.70 % of the best previous predictor RF-PseU. All results demonstrate that our model outperforms the state-of-art methods by a big margin and improves the previous models substantially. Moreover, PA-PseU is an incremental method which takes low computation cost with rapid computation speed. PA-PseU can be a useful tool for identifying the Ψ sites in transcriptome analysis. All the data, code and related materials are accessible at https://github.com/Jensen-Wang/PA-PseU. • A predictor called PA-PseU is provided based on multiple species with incremental Passive-Aggressive classifier. • Sequence order correlated factors and k-mer are applied to extract the sequence and structure information of RNA. • Combination of chi-square test and logistic regression is applied to reduce the dimension of features. [ABSTRACT FROM AUTHOR]