Improving Self-interacting Proteins Prediction Accuracy Using Protein Evolutionary Information and Weighed-Extreme Learning Machine

DeepAFP: An effective computational framework for identifying antifungal peptides based on deep learning

Fungal infections have become a significant global health issue, affecting millions worldwide. Antifungal peptides (AFPs) have emerged as a promising alternative to conventional antifungal drugs due to their low toxicity and low propensity for inducing resistance. In this study, we developed a deep learning-based framework called DeepAFP to efficiently identify AFPs. DeepAFP fully leverages and mines composition information, evolutionary information, and physicochemical properties of peptides by employing combined kernels from multiple branches of convolutional neural network with bi-directional long short-term memory layers. In addition, DeepAFP integrates a transfer learning strategy to obtain efficient representations of peptides for improving model performance. DeepAFP demonstrates strong predictive ability on carefully curated datasets, yielding an accuracy of 93.29% and an F1-score of 93.45% on the DeepAFP-Main dataset. The experimental results show that DeepAFP outperforms existing AFP prediction tools, achieving state-of-the-art performance. Finally, we provide a downloadable AFP prediction tool to meet the demands of large-scale prediction and facilitate the usage of our framework by the public or other researchers. Our framework can accurately identify AFPs in a short time without requiring significant human and material resources, and hence can accelerate the development of AFPs as well as contribute to the treatment of fungal infections. Furthermore, our method can provide new perspectives for other biological sequence analysis tasks.

TransDFL: Identification of Disordered Flexible Linkers in Proteins by Transfer Learning

Disordered flexible linkers (DFLs) are the functional disordered regions in proteins, which are the sub-regions of intrinsically disordered regions (IDRs) and play important roles in connecting domains and maintaining inter-domain interactions. Trained with the limited available DFLs, the existing DFL predictors based on the machine learning techniques tend to predict the ordered residues as DFLs, leading to a high falsepositive rate (FPR) and low prediction accuracy. Previous studies have shown that DFLs are extremely flexible disordered regions, which are usually predicted as disordered residues with high confidence [P(D) > 0.9] by an IDR predictor. Therefore, transferring an IDR predictor to an accurate DFL predictor is of great significance for understanding the functions of IDRs. In this study, we proposed a new predictor called TransDFL for identifying DFLs by transferring the RFPR-IDP predictor for IDR identification to the DFL prediction. The RFPR-IDP was pre-trained with IDR sequences to learn the general features between IDRs and DFLs, which is helpful to reduce the false positives in the ordered regions. RFPR-IDP was fine-tuned with the DFL sequences to capture the specific features of DFLs so as to be transferred into the TransDFL. Experimental results of two application scenarios (prediction of DFLs only in IDRs or prediction of DFLs in entire proteins) showed that TransDFL consistently outperformed other existing DFL predictors with higher accuracy. The corresponding web server of TransDFL can be freely accessed at http://bliulab.net/TransDFL/.

Accelerating the Discovery of Anticancer Peptides through Deep Forest Architecture with Deep Graphical Representation

Cancer is one of the leading diseases threatening human life and health worldwide. Peptide-based therapies have attracted much attention in recent years. Therefore, the precise prediction of anticancer peptides (ACPs) is crucial for discovering and designing novel cancer treatments. In this study, we proposed a novel machine learning framework (GRDF) that incorporates deep graphical representation and deep forest architecture for identifying ACPs. Specifically, GRDF extracts graphical features based on the physicochemical properties of peptides and integrates their evolutionary information along with binary profiles for constructing models. Moreover, we employ the deep forest algorithm, which adopts a layer-by-layer cascade architecture similar to deep neural networks, enabling excellent performance on small datasets but without complicated tuning of hyperparameters. The experiment shows GRDF exhibits state-of-the-art performance on two elaborate datasets (Set 1 and Set 2), achieving 77.12% accuracy and 77.54% F1-score on Set 1, as well as 94.10% accuracy and 94.15% F1-score on Set 2, exceeding existing ACP prediction methods. Our models exhibit greater robustness than the baseline algorithms commonly used for other sequence analysis tasks. In addition, GRDF is well-interpretable, enabling researchers to better understand the features of peptide sequences. The promising results demonstrate that GRDF is remarkably effective in identifying ACPs. Therefore, the framework presented in this study could assist researchers in facilitating the discovery of anticancer peptides and contribute to developing novel cancer treatments.

The Development of Machine Learning Methods in discriminating Secretory Proteins of Malaria Parasite

Malaria caused by Plasmodium falciparum is one of the major infectious diseases in the world. It is essential to exploit an effective method to predict secretory proteins of malaria parasites to develop effective cures and treatment. Biochemical assays can provide details for accurate identification of the secretory proteins, but these methods are expensive and time-consuming. In this paper, we summarized the machine learning-based identification algorithms and compared the construction strategies between different computational methods. Also, we discussed the use of machine learning to improve the ability of algorithms to identify proteins secreted by malaria parasites.

ProtRe-CN: Protein Remote Homology Detection by Combining Classification Methods and Network Methods via Learning to Rank

Protein remote homology detection is one of fundamental research tasks for downstream analysis (i.e., protein structure and function prediction). Many advanced methods are proposed from different views with complementary detection ability, such as the classification method, the network method, and the ranking method. A framework integrating these heterogeneous methods is urgently desired to reduce the false positive rate and predictive bias. We propose a novel ranking method called ProtRe-CN by fusing the classification methods and network methods via Learning to Rank. Experimental results on the benchmark dataset and the independent dataset show that ProtRe-CN outperforms other existing state-of-the-art predictors. ProtRe-CN improves the detective performance via correcting the false positives in the ranking list by combining the heterogeneous methods. The web server of ProtRe-CN can be accessed at http://bliulab.net/ProtRe-CN.

Deep-BSC: Predicting Raw DNA Binding Pattern in Arabidopsis Thaliana

Background:

With the rapid development of the sequencing methods in recent years, binding sites have been systematically identified in such projects as Nested-MICA and MEME. Prediction of DNA motifs with higher accuracy and precision has been a very important task for bioinformaticians. Nevertheless, experimental approaches are still time-consuming for big data set, making computational identification of binding sites indispensable.

Objective:

To facilitate the identification of the binding site, we proposed a deep learning architecture, named Deep-BSC (Deep-Learning Binary Search Classification), to predict binding sites in a raw DNA sequence with more precision and accuracy.

Methods:

Our proposed architecture purely relies on the raw DNA sequence to predict the binding sites for protein by using a convolutional neural network (CNN). We trained our deep learning model on binding sites at the nucleotide level. DNA sequence of A. thaliana is used in this study because it is a model plant.

Results:

The results demonstrate the effectiveness and efficiency of our method in the classification of binding sites against random sequences, using deep learning. We construct a CNN with different layers and filters to show the usefulness of max-pooling technique in the proposed method. To gain the interpretability of our approach, we further visualized binding sites in the saliency map and successfully identified similar motifs in the raw sequence. The proposed computational framework is time and resource efficient.

Conclusion:

Deep-BSC enables the identification of binding sites in the DNA sequences via a highly accurate CNN. The proposed computational framework can also be applied to problems such as operator, repeats in the genome, DNA markers, and recognition sites for enzymes, thereby promoting the use of Deep-BSC method in life sciences.

ATSE: a peptide toxicity predictor by exploiting structural and evolutionary information based on graph neural network and attention mechanism

MOTIVATION

Peptides have recently emerged as promising therapeutic agents against various diseases. For both research and safety regulation purposes, it is of high importance to develop computational methods to accurately predict the potential toxicity of peptides within the vast number of candidate peptides.

RESULTS

In this study, we proposed ATSE, a peptide toxicity predictor by exploiting structural and evolutionary information based on graph neural networks and attention mechanism. More specifically, it consists of four modules: (i) a sequence processing module for converting peptide sequences to molecular graphs and evolutionary profiles, (ii) a feature extraction module designed to learn discriminative features from graph structural information and evolutionary information, (iii) an attention module employed to optimize the features and (iv) an output module determining a peptide as toxic or non-toxic, using optimized features from the attention module.

CONCLUSION

Comparative studies demonstrate that the proposed ATSE significantly outperforms all other competing methods. We found that structural information is complementary to the evolutionary information, effectively improving the predictive performance. Importantly, the data-driven features learned by ATSE can be interpreted and visualized, providing additional information for further analysis. Moreover, we present a user-friendly online computational platform that implements the proposed ATSE, which is now available at http://server.malab.cn/ATSE. We expect that it can be a powerful and useful tool for researchers of interest.

prPred: A Predictor to Identify Plant Resistance Proteins by Incorporating k-Spaced Amino Acid (Group) Pairs

To infect plants successfully, pathogens adopt various strategies to overcome their physical and chemical barriers and interfere with the plant immune system. Plants deploy a large number of resistance (R) proteins to detect invading pathogens. The R proteins are encoded by resistance genes that contain cell surface-localized receptors and intracellular receptors. In this study, a new plant R protein predictor called prPred was developed based on a support vector machine (SVM), which can accurately distinguish plant R proteins from other proteins. Experimental results showed that the accuracy, precision, sensitivity, specificity, F1-score, MCC, and AUC of prPred were 0.935, 1.000, 0.806, 1.000, 0.893, 0.857, and 0.948, respectively, on an independent test set. Moreover, the predictor integrated the HMMscan search tool and Phobius to identify protein domain families and transmembrane protein regions to differentiate subclasses of R proteins. prPred is available at https://github.com/Wangys-prog/prPred. The tool requires a valid Python installation and is run from the command line.

A Deep Convolutional Neural Network to Improve the Prediction of Protein Secondary Structure

Background:

Protein secondary structure prediction (PSSP) is a fundamental task in bioinformatics that is helpful for understanding the three-dimensional structure and biological function of proteins. Many neural network-based prediction methods have been developed for protein secondary structures. Deep learning and multiple features are two obvious means to improve prediction accuracy.

Objective:

To promote the development of PSSP, a deep convolutional neural network-based method is proposed to predict both the eight-state and three-state of protein secondary structure.

Methods:

In this model, sequence and evolutionary information of proteins are combined as multiple input features after preprocessing. A deep convolutional neural network with no pooling layer and connection layer is then constructed to predict the secondary structure of proteins. L2 regularization, batch normalization, and dropout techniques are employed to avoid over-fitting and obtain better prediction performance, and an improved cross-entropy is used as the loss function.

Results:

Our proposed model can obtain Q3 prediction results of 86.2%, 84.5%, 87.8%, and 84.7%, respectively, on CullPDB, CB513, CASP10 and CASP11 datasets, with corresponding Q8 prediction results of 74.1%, 70.5%, 74.9%, and 71.3%.

Conclusion:

We have proposed the DCNN-SS deep convolutional-network-based PSSP method, and experimental results show that DCNN-SS performs competitively with other methods.

An Efficient Feature Extraction Technique Based on Local Coding PSSM and Multifeatures Fusion for Predicting Protein-Protein Interactions

Background

Increasing evidence has indicated that protein-protein interactions (PPIs) play important roles in various aspects of the structural and functional organization of a cell. Thus, continuing to uncover potential PPIs is an important topic in the biomedical domain. Although various feature extraction methods with machine learning approaches have enhanced the prediction of PPIs. There remains room for improvement by developing novel and effective feature extraction methods and classifier approaches to identify PPIs.

Method

In this study, we proposed a sequence-based feature extraction method called LCPSSMMF, which combined local coding position-specific scoring matrix (PSSM) with multifeatures fusion. First, we used a novel local coding method based on PSSM to build a new PSSM (CPSSM); the advantage of this method is that it incorporated global and local feature extraction, which can account for the interactions between residues in both continuous and discontinuous regions of amino acid sequences. Second, we adopted 2 different feature extraction methods (Local Average Group [LAG] and Bigram Probability [BP]) to capture multiple key feature information by employing the evolutionary information embedded in the CPSSM matrix. Finally, feature vectors were acquired by using multifeatures fusion method.

Result

To evaluate the performance of the proposed feature extraction approach, we employed support vector machine (SVM) as a prediction classifier and applied this method to yeast and human PPI datasets. The prediction accuracies of LCPSSMMF were 93.43% and 90.41% on the yeast and human datasets, respectively. Moreover, we also compared the proposed method with the previous sequence-based approaches on the yeast datasets by using the same SVM classifier. The experimental results indicated that the performance of LCPSSMMF significantly exceeded that of several other state-of-the-art methods. It is proven that the LCPSSMMF approach can capture more local and global discriminatory information than almost all previous methods and can function remarkably well in identifying PPIs. To facilitate extensive research in future proteomics studies, we developed a LCPSSMMFSVM server, which is freely available for academic use at http://219.219.62.123:8888/LCPSSMMFSVM.