Wenzheng Bao

Classification of Protein Structure Classes on Flexible Neutral Tree

Accurate classification on protein structural is playing an important role in Bioinformatics. An increase in evidence demonstrates that a variety of classification methods have been employed in such a field. In this research, the features of amino acids composition, secondary structures feature, and correlation coefficient of amino acid dimers and amino acid triplets have been used. Flexible neutral tree (FNT), a particular tree structure neutral network, has been employed as the classification model in the protein structures’ classification framework. Considering different feature groups owing diverse roles in the model, impact factors of different groups have been put forward in this research. In order to evaluate different impact factors, Impact Factors Scaling (IFS) algorithm, which aim at reducing redundant information of the selected features in some degree, have been put forward. To examine the performance of such framework, the 640, 1189, and ASTRAL datasets are employed as the low-homology protein structure benchmark datasets. Experimental results demonstrate that the performance of the proposed method is better than the other methods in the low-homology protein tertiary structures.

Novel human microbe-disease association prediction using network consistency projection

BackgroundAccumulating biological and clinical reports have indicated that imbalance of microbial community is closely associated with occurrence and development of various complex human diseases. Identifying potential microbe-disease associations, which could provide better understanding of disease pathology and further boost disease diagnostic and prognostic, has attracted more and more attention. However, hardly any computational models have been developed for large scale microbe-disease association prediction.ResultsIn this article, based on the assumption that microbes with similar functions tend to share similar association or non-association patterns with similar diseases and vice versa, we proposed the model of Network Consistency Projection for Human Microbe-Disease Association prediction (NCPHMDA) by integrating known microbe-disease associations and Gaussian interaction profile kernel similarity for microbes and diseases. NCPHMDA yielded outstanding AUCs of 0.9039, 0.7953 and average AUC of 0.8918 in global leave-one-out cross validation, local leave-one-out cross validation and 5-fold cross validation, respectively. Furthermore, colon cancer, asthma and type 2 diabetes were taken as independent case studies, where 9, 9 and 8 out of the top 10 predicted microbes were successfully confirmed by recent published clinical literature.ConclusionNCPHMDA is a non-parametric universal network-based method which can simultaneously predict associated microbes for investigated diseases but does not require negative samples. It is anticipated that NCPHMDA would become an effective biological resource for clinical experimental guidance.

A Novel Protein Structural Classes Prediction Method Based on Hierarchical Classification Model

In the post-genomic era prediction of protein structural classes is an important area in bioinformatics, it is beneficial to research protein function, regulation and interactions. In this paper, a novel hierarchical classification model based on flexible neural tree (FNT) was been built, different features were extracted based on the predicted secondary structure sequence and the corresponding E-H sequence for every classifiers. Three datasets with low homology were used to test the proposed method compared to existing methods. The overall accuracy of this method is all improved on three datasets.

Prediction of protein structure classes

Prediction of protein special structural plays a significant role to better recognize the protein folding patterns. Multiple prediction methods may be used to predict the structures based on the information of sequences and biostatistics. The accuracy, nevertheless, is strongly affected by the efficiency of classification, the robustness of model and other factors. In our research, flexible neutral tree (FNT), a novel classification model, is employed as the base classifiers. The alterable structural tree take advantage of the selection of available features, aims to improve the efficiency. To examine the performance and efficiency of such algorithm combination, an ASTRAL dataset is selected as the test dataset. Our results show that a higher prediction accuracy could be achieved compared with other methods, the structure of the classification model for prediction of protein structural may make incremental improvements possible.

Prediction of Protein Structure Classes with Ensemble Classifiers

Protein structure prediction is an important area of research in bioinformatics. In this research, a novel method to predict the structure of the protein is introduced. The amino acid frequencies, generalization dipeptide composition and typical hydrophobic composition of protein structure are treated as candidate feature. Flexible neural tree and neural network are employed as classification model. To evaluate the efficiency of the proposed method, a classical protein sequence dataset (1189) is selected as the test dataset. The results show that the method is efficient for protein structure prediction.

Recurrent Neural Network for Predicting Transcription Factor Binding Sites

It is well known that DNA sequence contains a certain amount of transcription factors (TF) binding sites, and only part of them are identified through biological experiments. However, these experiments are expensive and time-consuming. To overcome these problems, some computational methods, based on k-mer features or convolutional neural networks, have been proposed to identify TF binding sites from DNA sequences. Although these methods have good performance, the context information that relates to TF binding sites is still lacking. Research indicates that standard recurrent neural networks (RNN) and its variants have better performance in time-series data compared with other models. In this study, we propose a model, named KEGRU, to identify TF binding sites by combining Bidirectional Gated Recurrent Unit (GRU) network with k-mer embedding. Firstly, DNA sequences are divided into k-mer sequences with a specified length and stride window. And then, we treat each k-mer as a word and pre-trained word representation model though word2vec algorithm. Thirdly, we construct a deep bidirectional GRU model for feature learning and classification. Experimental results have shown that our method has better performance compared with some state-of-the-art methods. Additional experiments about embedding strategy show that k-mer embedding will be helpful to enhance model performance. The robustness of KEGRU is proved by experiments with different k-mer length, stride window and embedding vector dimension.

HSCVFNT: Inference of Time-Delayed Gene Regulatory Network Based on Complex-Valued Flexible Neural Tree Model

Gene regulatory network (GRN) inference can understand the growth and development of animals and plants, and reveal the mystery of biology. Many computational approaches have been proposed to infer GRN. However, these inference approaches have hardly met the need of modeling, and the reducing redundancy methods based on individual information theory method have bad universality and stability. To overcome the limitations and shortcomings, this thesis proposes a novel algorithm, named HSCVFNT, to infer gene regulatory network with time-delayed regulations by utilizing a hybrid scoring method and complex-valued flexible neural network (CVFNT). The regulations of each target gene can be obtained by iteratively performing HSCVFNT. For each target gene, the HSCVFNT algorithm utilizes a novel scoring method based on time-delayed mutual information (TDMI), time-delayed maximum information coefficient (TDMIC) and time-delayed correlation coefficient (TDCC), to reduce the redundancy of regulatory relationships and obtain the candidate regulatory factor set. Then, the TDCC method is utilized to create time-delayed gene expression time-series matrix. Finally, a complex-valued flexible neural tree model is proposed to infer the time-delayed regulations of each target gene with the time-delayed time-series matrix. Three real time-series expression datasets from (Save Our Soul) SOS DNA repair system in E. coli and Saccharomyces cerevisiae are utilized to evaluate the performance of the HSCVFNT algorithm. As a result, HSCVFNT obtains outstanding F-scores of 0.923, 0.8 and 0.625 for SOS network and (In vivo Reverse-Engineering and Modeling Assessment) IRMA network inference, respectively, which are 5.5%, 14.3% and 72.2% higher than the best performance of other state-of-the-art GRN inference methods and time-delayed methods.

Mutli-Features Prediction of Protein Translational Modification Sites

Post translational modification plays a significiant role in the biological processing. The potential post translational modification is composed of the center sites and the adjacent amino acid residues which are fundamental protein sequence residues. It can be helpful to perform their biological functions and contribute to understanding the molecular mechanisms that are the foundations of protein design and drug design. The existing algorithms of predicting modified sites often have some shortcomings, such as lower stability and accuracy. In this paper, a combination of physical, chemical, statistical, and biological properties of a protein have been ulitized as the features, and a novel framework is proposed to predict a proteins post translational modification sites. The multi-layer neural network and support vector machine are invoked to predict the potential modified sites with the selected features that include the compositions of amino acid residues, the E-H description of protein segments, and several properties from the AAIndex database. Being aware of the possible redundant information, the feature selection is proposed in the propocessing step in this research. The experimental results show that the proposed method has the ability to improve the accuracy in this classification issue.

Cross-validated smooth multi-instance learning

The problem of object localization in image appear ubiquitously in computer vision applications including image classification, object detection and visual tracking. Recently, it is shown that multiple-instance learning(MIL) which is regarded as the fourth machine learning framework compared with supervised learning, unsupervised learning and reinforce learning has been verified that will get good effect in object localization in images. In this paper, we propose a novel method to solve the classical MIL problem, named Cross-Validated Smooth Multi-Instance learning (CVS-MIL). We treat the positiveness of instance as a continuous variable. The softmax model is used to bring a bridge between instances and bags and jointly optimize the bag label and instance label in a unified framework. The extensive experiments demonstrate that CVS-MIL consistently achieves superior performance on various MIL benchmarks. Moreover, we simply applied CVS-MIL to a challenging vision task, common object discovery. The state-of-the-art results of object discovery on Pascal VOC datasets further confirm the advantages of the proposed method.

Convex local sensitive low rank matrix approximation

The problem of matrix approximation appears ubiquitously in recommendation systems, computer vision and text mining. The prevailing assumption is that the partially observed matrix has a low-rank or can be well approximated by a low-rank matrix. However, this assumption is strictly that the partially observed matrix is globally low rank. In this paper, we propose a local sensitive formulation of matrix approximation which relaxes the global low-rank assumption, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We solve the problem by an efficient way based on the alternating direction method of multipliers (ADMM). Our experiments show improvements in prediction accuracy over classical approaches for recommendation tasks.