A deep learning ensemble approach to prioritize antiviral drugs against novel coronavirus SARS-CoV-2 for COVID-19 drug repurposing

The alarming pandemic situation of Coronavirus infectious disease COVID-19, caused by the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), has become a critical threat to public health. The unexpected outbreak and unrealistic progression of COVID-19 have generated an utmost need to realize promising therapeutic strategies to fight the pandemic. Drug repurposing-an efficient drug discovery technique from approved drugs is an emerging tactic to face the immediate global challenge. It​ offers a time-efficient and cost-effective way to find potential therapeutic agents for the disease. Artificial Intelligence-empowered deep learning models enable the rapid identification of potentially repurposable drug candidates against diseases. This study presents a deep learning ensemble model to prioritize clinically validated anti-viral drugs for their potential efficacy against SARS-CoV-2. The method integrates the similarities of drug chemical structures and virus genome sequences to generate feature vectors. The best combination of features is retrieved by the convolutional neural network in a deep learning manner. The extracted deep features are classified by the extreme gradient boosting classifier to infer potential virus–drug associations. The method could achieve an AUC of 0.8897 with 0.8571 prediction accuracy and 0.8394 sensitivity under the fivefold cross-validation. The experimental results and case studies demonstrate the suggested deep learning ensemble system yields competitive results compared with the state-of-the-art approaches. The top-ranked drugs are released for further wet-lab researches.

BERT-m7G: A Transformer Architecture Based on BERT and Stacking Ensemble to Identify RNA N7-Methylguanosine Sites from Sequence Information

As one of the most prevalent posttranscriptional modifications of RNA, N7-methylguanosine (m7G) plays an essential role in the regulation of gene expression. Accurate identification of m7G sites in the transcriptome is invaluable for better revealing their potential functional mechanisms. Although high-throughput experimental methods can locate m7G sites precisely, they are overpriced and time-consuming. Hence, it is imperative to design an efficient computational method that can accurately identify the m7G sites. In this study, we propose a novel method via incorporating BERT-based multilingual model in bioinformatics to represent the information of RNA sequences. Firstly, we treat RNA sequences as natural sentences and then employ bidirectional encoder representations from transformers (BERT) model to transform them into fixed-length numerical matrices. Secondly, a feature selection scheme based on the elastic net method is constructed to eliminate redundant features and retain important features. Finally, the selected feature subset is input into a stacking ensemble classifier to predict m7G sites, and the hyperparameters of the classifier are tuned with tree-structured Parzen estimator (TPE) approach. By 10-fold cross-validation, the performance of BERT-m7G is measured with an ACC of 95.48% and an MCC of 0.9100. The experimental results indicate that the proposed method significantly outperforms state-of-the-art prediction methods in the identification of m7G modifications.

iDeepSubMito: identification of protein submitochondrial localization with deep learning

Mitochondria are membrane-bound organelles containing over 1000 different proteins involved in mitochondrial function, gene expression and metabolic processes. Accurate localization of those proteins in the mitochondrial compartments is critical to their operation. A few computational methods have been developed for predicting submitochondrial localization from the protein sequences. Unfortunately, most of these computational methods focus on employing biological features or evolutionary information to extract sequence features, which greatly limits the performance of subsequent identification. Moreover, the efficiency of most computational models is still under explored, especially the deep learning feature, which is promising but requires improvement. To address these limitations, we propose a novel computational method called iDeepSubMito to predict the location of mitochondrial proteins to the submitochondrial compartments. First, we adopted a coding scheme using the ProteinELMo to model the probability distribution over the protein sequences and then represent the protein sequences as continuous vectors. Then, we proposed and implemented convolutional neural network architecture based on the bidirectional LSTM with self-attention mechanism, to effectively explore the contextual information and protein sequence semantic features. To demonstrate the effectiveness of our proposed iDeepSubMito, we performed cross-validation on two datasets containing 424 proteins and 570 proteins respectively, and consisting of four different mitochondrial compartments (matrix, inner membrane, outer membrane and intermembrane regions). Experimental results revealed that our method outperformed other computational methods. In addition, we tested iDeepSubMito on the M187, M983 and MitoCarta3.0 to further verify the efficiency of our method. Finally, the motif analysis and the interpretability analysis were conducted to reveal novel insights into subcellular biological functions of mitochondrial proteins. iDeepSubMito source code is available on GitHub at https://github.com/houzl3416/iDeepSubMito.

DNN-DTIs: Improved drug-target interactions prediction using XGBoost feature selection and deep neural network

Analysis and prediction of drug-target interactions (DTIs) play an important role in understanding drug mechanisms, as well as drug repositioning and design. Machine learning (ML)-based methods for DTIs prediction can mitigate the shortcomings of time-consuming and labor-intensive experimental approaches, while providing new ideas and insights for drug design. We propose a novel pipeline for predicting drug-target interactions, called DNN-DTIs. First, the target information is characterized by a number of features, namely, pseudo-amino acid composition, pseudo position-specific scoring matrix, conjoint triad composition, transition and distribution, Moreau-Broto autocorrelation, and structural features. The drug compounds are subsequently encoded using substructure fingerprints. Next, eXtreme gradient boosting (XGBoost) is used to determine the subset of non-redundant features of importance. The optimal balanced set of sample vectors is obtained by applying the synthetic minority oversampling technique (SMOTE). Finally, a DTIs predictor, DNN-DTIs, is developed based on a deep neural network (DNN) via a layer-by-layer learning scheme. Experimental results indicate that DNN-DTIs achieves better performance than other state-of-the-art predictors with ACC values of 98.78%, 98.60%, 97.98%, 98.24% and 98.00% on Enzyme, Ion Channels (IC), GPCR, Nuclear Receptors (NR) and Kuang's datasets. Therefore, the accurate prediction performance of DNN-DTIs makes it a favored choice for contributing to the study of DTIs, especially drug repositioning.

Prediction of protein ubiquitination sites via multi-view features based on eXtreme gradient boosting classifier

Ubiquitination is a common and reversible post-translational protein modification that regulates apoptosis and plays an important role in protein degradation and cell diseases. However, experimental identification of protein ubiquitination sites is usually time-consuming and labor-intensive, so it is necessary to establish effective predictors. In this study, we propose a ubiquitination sites prediction method based on multi-view features, namely UbiSite-XGBoost. Firstly, we use seven single-view features encoding methods to convert protein sequence fragments into digital information. Secondly, the least absolute shrinkage and selection operator (LASSO) is applied to remove the redundant information and get the optimal feature subsets. Finally, these features are inputted into the eXtreme gradient boosting (XGBoost) classifier to predict ubiquitination sites. Five-fold cross-validation shows that the AUC values of Set1-Set6 datasets are 0.8258, 0.7592, 0.7853, 0.8345, 0.8979 and 0.8901, respectively. The synthetic minority oversampling technique (SMOTE) is employed in Set4-Set6 unbalanced datasets, and the AUC values are 0.9777, 0.9782 and 0.9860, respectively. In addition, we have constructed three independent test datasets which the AUC values are 0.8007, 0.6897 and 0.7280, respectively. The results show that the proposed method UbiSite-XGBoost is superior to other ubiquitination prediction methods and it provides new guidance for the identification of ubiquitination sites. The source code and all datasets are available at https://github.com/QUST-AIBBDRC/UbiSite-XGBoost/.

iEnhancer-XG: interpretable sequence-based enhancers and their strength predictor

MOTIVATION

Enhancers are non-coding DNA fragments with high position variability and free scattering. They play an important role in controlling gene expression. As machine learning has become more widely used in identifying enhancers, a number of bioinformatic tools have been developed. Although several models for identifying enhancers and their strengths have been proposed, their accuracy and efficiency have yet to be improved.

RESULTS

We propose a two-layer predictor called 'iEnhancer-XG.' It comprises a one-layer predictor (for identifying enhancers) and a second classifier (for their strength) and uses 'XGBoost' as a base classifier and five feature extraction methods, namely, k-Spectrum Profile, Mismatch k-tuple, Subsequence Profile, Position-specific scoring matrix (PSSM) and Pseudo dinucleotide composition (PseDNC). Each method has an independent output. We place the feature vector matrix into the ensemble learning for fusion. This experiment involves the method of 'SHapley Additive explanations' to provide interpretability for the previous black box machine learning methods and improve their credibility. The accuracies of the ensemble learning method are 0.811 (first layer) and 0.657 (second layer). The rigorous 10-fold cross-validation confirms that the proposed method is significantly better than existing technologies.

AVAILABILITY AND IMPLEMENTATION

The source code and dataset for the enhancer predictions have been uploaded to https://github.com/jimmyrate/ienhancer-xg.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Prediction of protein-protein interaction sites through eXtreme gradient boosting with kernel principal component analysis

Predicting protein-protein interaction sites (PPI sites) can provide important clues for understanding biological activity. Using machine learning to predict PPI sites can mitigate the cost of running expensive and time-consuming biological experiments. Here we propose PPISP-XGBoost, a novel PPI sites prediction method based on eXtreme gradient boosting (XGBoost). First, the characteristic information of protein is extracted through the pseudo-position specific scoring matrix (PsePSSM), pseudo-amino acid composition (PseAAC), hydropathy index and solvent accessible surface area (ASA) under the sliding window. Next, these raw features are preprocessed to obtain more optimal representations in order to achieve better prediction. In particular, the synthetic minority oversampling technique (SMOTE) is used to circumvent class imbalance, and the kernel principal component analysis (KPCA) is applied to remove redundant characteristics. Finally, these optimal features are fed to the XGBoost classifier to identify PPI sites. Using PPISP-XGBoost, the prediction accuracy on the training dataset Dset186 reaches 85.4%, and the accuracy on the independent validation datasets Dtestset72, PDBtestset164, Dset_448 and Dset_355 reaches 85.3%, 83.9%, 85.8% and 85.4%, respectively, which all show an increase in accuracy against existing PPI sites prediction methods. These results demonstrate that the PPISP-XGBoost method can further enhance the prediction of PPI sites.

Regional Population Forecast and Analysis Based on Machine Learning Strategy

Regional population forecast and analysis is of essence to urban and regional planning, and a well-designed plan can effectively construct a sound national infrastructure and stabilize positive population growth. Traditionally, either urban or regional planning relies on the opinions of demographers in terms of how the population of a city or a region will grow. Multi-regional population forecast is currently possible, carried out mainly on the basis of the Interregional Cohort-Component model. While this model has its unique advantages, several demographic rates are determined based on the decisions made by primary planners. Hence, the only drawback for cohort-component type population forecasting is allowing the analyst to specify the demographic rates of the future, and it goes without saying that this tends to introduce a biased result in forecasting accuracy. To effectively avoid this problem, this work proposes a machine learning-based method to forecast multi-regional population growth objectively. Thus, this work, drawing upon the newly developed machine learning technology, attempts to analyze and forecast the population growth of major cities in Taiwan. By effectively using the advantage of the XGBoost algorithm, the evaluation of feature importance and the forecast of multi-regional population growth between the present and the near future can be observed objectively, and it can further provide an objective reference to the urban planning of regional population.

An interpretable machine learning prognostic system for locoregionally advanced nasopharyngeal carcinoma based on tumor burden features

OBJECTIVES

We aimed to build a survival system by combining a highly-accurate machine learning (ML) model with explainable artificial intelligence (AI) techniques to predict distant metastasis in locoregionally advanced nasopharyngeal carcinoma (NPC) patients using magnetic resonance imaging (MRI)-based tumor burden features.

MATERIALS AND METHODS

1643 patients from three hospitals were enrolled according to set criteria. We employed ML to develop a survival model based on tumor burden signatures and all clinical factors. Shapley Additive exPlanations (SHAP) was utilized to explain prediction results and interpret the complex non-linear relationship among features and distant metastasis. We also constructed other models based on routinely used cancer stages, Epstein-Barr virus (EBV) DNA, or other clinical features for comparison. Concordance index (C-index), receiver operating curve (ROC) analysis and decision curve analysis (DCA) were executed to assess the effectiveness of the models.

RESULTS

Our proposed system consistently demonstrated promising performance across independent cohorts. The concordance indexes were 0.773, 0.766 and 0.760 in the training, internal validation and external validation sets. SHAP provided personalized protective and risk factors for each NPC patient and uncovered some novel non-linear relationships between features and distant metastasis. Furthermore, high-risk patients who received induction chemotherapy (ICT) and concurrent chemoradiotherapy (CCRT) had better 5-year distant metastasis-free survival (DMFS) than those who only received CCRT, whereas ICT + CCRT and CCRT had similar DMFS in low-risk patients.

CONCLUSIONS

The interpretable machine learning system demonstrated superior performance in predicting metastasis in locoregionally advanced NPC. High-risk patients might benefit from ICT.

NeuroPred-FRL: an interpretable prediction model for identifying neuropeptide using feature representation learning

Neuropeptides (NPs) are the most versatile neurotransmitters in the immune systems that regulate various central anxious hormones. An efficient and effective bioinformatics tool for rapid and accurate large-scale identification of NPs is critical in immunoinformatics, which is indispensable for basic research and drug development. Although a few NP prediction tools have been developed, it is mandatory to improve their NPs' prediction performances. In this study, we have developed a machine learning-based meta-predictor called NeuroPred-FRL by employing the feature representation learning approach. First, we generated 66 optimal baseline models by employing 11 different encodings, six different classifiers and a two-step feature selection approach. The predicted probability scores of NPs based on the 66 baseline models were combined to be deemed as the input feature vector. Second, in order to enhance the feature representation ability, we applied the two-step feature selection approach to optimize the 66-D probability feature vector and then inputted the optimal one into a random forest classifier for the final meta-model (NeuroPred-FRL) construction. Benchmarking experiments based on both cross-validation and independent tests indicate that the NeuroPred-FRL achieves a superior prediction performance of NPs compared with the other state-of-the-art predictors. We believe that the proposed NeuroPred-FRL can serve as a powerful tool for large-scale identification of NPs, facilitating the characterization of their functional mechanisms and expediting their applications in clinical therapy. Moreover, we interpreted some model mechanisms of NeuroPred-FRL by leveraging the robust SHapley Additive exPlanation algorithm.

Accurate prediction of multi-label protein subcellular localization through multi-view feature learning with RBRL classifier

Multi-label proteins can participate in carrier transportation, enzyme catalysis, hormone regulation and other life activities. Meanwhile, they play a key role in the fields of biopharmaceuticals, gene and cell therapy. This article proposes a prediction method called Mps-mvRBRL to predict the subcellular localization (SCL) of multi-label protein. Firstly, pseudo position-specific scoring matrix, dipeptide composition, position specific scoring matrix-transition probability composition, gene ontology and pseudo amino acid composition algorithms are used to obtain numerical information from different views. Based on the contribution of five individual feature extraction methods, differential evolution is used for the first time to learn the weight of single feature, and then these original features use a weighted combination method to fuse multi-view information. Secondly, the fused high-dimensional features use a weighted linear discriminant analysis framework based on binary weight form to eliminate irrelevant information. Finally, the best feature vector is input into the joint ranking support vector machine and binary relevance with robust low-rank learning classifier to predict the SCL. After applying leave-one-out cross-validation, the overall actual accuracy (OAA) and overall location accuracy (OLA) of Mps-mvRBRL on the training set of Gram-positive bacteria are both 99.81%. The OAA on the test sets of plant, virus and Gram-negative bacteria datasets are 97.24%, 98.55% and 98.20%, respectively, and the OLA are 97.16%, 97.62% and 98.28%, respectively. The results show that the model achieves good prediction performance for predicting the SCL of multi-label protein.

GTB-PPI: Predict Protein-protein Interactions Based on L1-regularized Logistic Regression and Gradient Tree Boosting

Protein–protein interactions (PPIs) are of great importance to understand genetic mechanisms, delineate disease pathogenesis, and guide drug design. With the increase of PPI data and development of machine learning technologies, prediction and identification of PPIs have become a research hotspot in proteomics. In this study, we propose a new prediction pipeline for PPIs based on gradient tree boosting (GTB). First, the initial feature vector is extracted by fusing pseudo amino acid composition (PseAAC), pseudo position-specific scoring matrix (PsePSSM), reduced sequence and index-vectors (RSIV), and autocorrelation descriptor (AD). Second, to remove redundancy and noise, we employ L1-regularized logistic regression (L1-RLR) to select an optimal feature subset. Finally, GTB-PPI model is constructed. Five-fold cross-validation showed that GTB-PPI achieved the accuracies of 95.15% and 90.47% on Saccharomyces cerevisiae and Helicobacter pylori datasets, respectively. In addition, GTB-PPI could be applied to predict the independent test datasets for Caenorhabditis elegans, Escherichia coli, Homo sapiens, and Mus musculus, the one-core PPI network for CD9, and the crossover PPI network for the Wnt-related signaling pathways. The results show that GTB-PPI can significantly improve accuracy of PPI prediction. The code and datasets of GTB-PPI can be downloaded from https://github.com/QUST-AIBBDRC/GTB-PPI/.

Comparison of machine learning tools for the prediction of AMD based on genetic, age, and diabetes-related variables in the Chinese population

Introduction

Age-related macular degeneration (AMD) is the main cause of visual impairment and the most important cause of blindness in older people. However, there is currently no effective treatment for this disease, so it is necessary to establish a risk model to predict AMD development.

Methods

This study included a total of 202 subjects, comprising 82 AMD patients and 120 control subjects. Sixty-six single-nucleotide polymorphisms (SNPs) were identified using the MassArray assay. Considering 14 independent clinical variables as well as SNPs, four predictive models were established in the training set and evaluated by the confusion matrix, area under the receiver operating characteristic (ROC) curve (AUROC). The difference distributions of the 14 independent clinical features between the AMD and control groups were tested using the chi-squared test. Age and diabetes were adjusted using logistic regression analysis and the “genomic-control” method was used for multiple testing correction.

Results

Three SNPs (rs10490924, OR = 1.686, genomic-control corrected p-value (GC) = 0.030; rs2338104, OR = 1.794, GC = 0.025 and rs1864163, OR = 2.125, GC = 0.038) were significant risk factors for AMD development. In the training set, four models obtained AUROC values above 0.72.

Conclusions

We believe machine learning tools will be useful for the early prediction of AMD and for the development of relevant intervention strategies.

Bird Eye View of Protein Subcellular Localization Prediction

Proteins are made up of long chain of amino acids that perform a variety of functions in different organisms. The activity of the proteins is determined by the nucleotide sequence of their genes and by its 3D structure. In addition, it is essential for proteins to be destined to their specific locations or compartments to perform their structure and functions. The challenge of computational prediction of subcellular localization of proteins is addressed in various in silico methods. In this review, we reviewed the progress in this field and offered a bird eye view consisting of a comprehensive listing of tools, types of input features explored, machine learning approaches employed, and evaluation matrices applied. We hope the review will be useful for the researchers working in the field of protein localization predictions.

scGMAI: a Gaussian mixture model for clustering single-cell RNA-Seq data based on deep autoencoder

The rapid development of single-cell RNA sequencing (scRNA-Seq) technology provides strong technical support for accurate and efficient analyzing single-cell gene expression data. However, the analysis of scRNA-Seq is accompanied by many obstacles, including dropout events and the curse of dimensionality. Here, we propose the scGMAI, which is a new single-cell Gaussian mixture clustering method based on autoencoder networks and the fast independent component analysis (FastICA). Specifically, scGMAI utilizes autoencoder networks to reconstruct gene expression values from scRNA-Seq data and FastICA is used to reduce the dimensions of reconstructed data. The integration of these computational techniques in scGMAI leads to outperforming results compared to existing tools, including Seurat, in clustering cells from 17 public scRNA-Seq datasets. In summary, scGMAI is an effective tool for accurately clustering and identifying cell types from scRNA-Seq data and shows the great potential of its applicative power in scRNA-Seq data analysis. The source code is available at https://github.com/QUST-AIBBDRC/scGMAI/.

NonClasGP-Pred: robust and efficient prediction of non-classically secreted proteins by integrating subset-specific optimal models of imbalanced data

Non-classically secreted proteins (NCSPs) are proteins that are located in the extracellular environment, although there is a lack of known signal peptides or secretion motifs. They usually perform different biological functions in intracellular and extracellular environments, and several of their biological functions are linked to bacterial virulence and cell defence. Accurate protein localization is essential for all living organisms, however, the performance of existing methods developed for NCSP identification has been unsatisfactory and in particular suffer from data deficiency and possible overfitting problems. Further improvement is desirable, especially to address the lack of informative features and mining subset-specific features in imbalanced datasets. In the present study, a new computational predictor was developed for NCSP prediction of gram-positive bacteria. First, to address the possible prediction bias caused by the data imbalance problem, ten balanced subdatasets were generated for ensemble model construction. Then, the F-score algorithm combined with sequential forward search was used to strengthen the feature representation ability for each of the training subdatasets. Third, the subset-specific optimal feature combination process was adopted to characterize the original data from different aspects, and all subdataset-based models were integrated into a unified model, NonClasGP-Pred, which achieved an excellent performance with an accuracy of 93.23 %, a sensitivity of 100 %, a specificity of 89.01 %, a Matthew's correlation coefficient of 87.68 % and an area under the curve value of 0.9975 for ten-fold cross-validation. Based on assessment on the independent test dataset, the proposed model outperformed state-of-the-art available toolkits. For availability and implementation, see: http://lab.malab.cn/~wangchao/softwares/NonClasGP/.

Computational prediction of species-specific yeast DNA replication origin via iterative feature representation

Deoxyribonucleic acid replication is one of the most crucial tasks taking place in the cell, and it has to be precisely regulated. This process is initiated in the replication origins (ORIs), and thus it is essential to identify such sites for a deeper understanding of the cellular processes and functions related to the regulation of gene expression. Considering the important tasks performed by ORIs, several experimental and computational approaches have been developed in the prediction of such sites. However, existing computational predictors for ORIs have certain curbs, such as building only single-feature encoding models, limited systematic feature engineering efforts and failure to validate model robustness. Hence, we developed a novel species-specific yeast predictor called yORIpred that accurately identify ORIs in the yeast genomes. To develop yORIpred, we first constructed optimal 40 baseline models by exploring eight different sequence-based encodings and five different machine learning classifiers. Subsequently, the predicted probability of 40 models was considered as the novel feature vector and carried out iterative feature learning approach independently using five different classifiers. Our systematic analysis revealed that the feature representation learned by the support vector machine algorithm (yORIpred) could well discriminate the distribution characteristics between ORIs and non-ORIs when compared with the other four algorithms. Comprehensive benchmarking experiments showed that yORIpred achieved superior and stable performance when compared with the existing predictors on the same training datasets. Furthermore, independent evaluation showcased the best and accurate performance of yORIpred thus underscoring the significance of iterative feature representation. To facilitate the users in obtaining their desired results without undergoing any mathematical, statistical or computational hassles, we developed a web server for the yORIpred predictor, which is available at: http://thegleelab.org/yORIpred.

Estimation of the LDL subclasses in ischemic stroke as a risk factor in a Chinese population

Background

Acute ischemic stroke (AIS) is one of the leading causes of mortality and long-term disability worldwide. Our study aims to clarify the role of low-density lipoproteins (LDL) subclasses in the occurrence of AIS and develop a risk xprediction model based on these characteristics to identify high-risk people.

Methods

Five hundred and sixty-six patients with AIS and 197 non-AIS controls were included in this study. Serum lipids and other baseline characteristics including fasting blood glucose (GLU), serum creatinine (Scr), and blood pressure were investigated in relation to occurrence of AIS. The LDL subfractions were classified and measured with the Lipoprint System by a polyacrylamide gel electrophoresis technique.

Results

Levels of LDL-3, LDL-4 and LDL-5 subclasses were significantly higher in the AIS group compared to the non-AIS group and lower level of LDL-1 was prevalent in the AIS patients. Consistently, Spearman correlation coefficient demonstrated that sd-demonevels, especially LDL-3 and LDL-4 levels, were significantly positively correlated with AIS. Furthermore, there is a significant positive correlation between small dense LDL (sd-LDL, that is LDL-3 to 7) levels and serum lipids including total cholesterol (TC), Low density lipoprotein cholesterol (LDL-C), and Triglyceride (TG). Increased LDL-3 and LDL-4 as well as decreased LDL-1 and LDL-2 were correlated to the occurrence of AIS, even in the people with normal LDL-C levels. A new prediction model including 12 variables can accurately predict the AIS risk in Chinese patients (AUC = 0.82 ± 0.04).

Conclusions

Levels of LDL subclasses should be considered in addition to serum LDL-C in assessment and management of AIS. A new prediction model based on clinical variables including LDL subtractions can help clinicians identify high of AIS, even in the people with norm.

Computational prediction and interpretation of cell-specific replication origin sites from multiple eukaryotes by exploiting stacking framework

Origins of replication sites (ORIs), which refers to the initiative locations of genomic DNA replication, play essential roles in DNA replication process. Detection of ORIs' distribution in genome scale is one of key steps to in-depth understanding their regulation mechanisms. In this study, we presented a novel machine learning-based approach called Stack-ORI encompassing 10 cell-specific prediction models for identifying ORIs from four different eukaryotic species (Homo sapiens, Mus musculus, Drosophila melanogaster and Arabidopsis thaliana). For each cell-specific model, we employed 12 feature encoding schemes that cover nucleic acid composition, position-specific and physicochemical properties information. The optimal feature set was identified from each encoding individually and developed their respective baseline models using the eXtreme Gradient Boosting (XGBoost) classifier. Subsequently, the predicted scores of 12 baseline models are integrated as a novel feature vector to train XGBoost and develop the final model. Extensive experimental results show that Stack-ORI achieves significantly better performance as compared with their baseline models on both training and independent datasets. Interestingly, Stack-ORI consistently outperforms existing predictor in all cell-specific models, not only on training but also on independent test. Moreover, our novel approach provides necessary interpretations that help understanding model success by leveraging the powerful SHapley Additive exPlanation algorithm, thus underlining the most important feature encoding schemes significant for predicting cell-specific ORIs.

iMRM: a platform for simultaneously identifying multiple kinds of RNA modifications

MOTIVATION

RNA modifications play critical roles in a series of cellular and developmental processes. Knowledge about the distributions of RNA modifications in the transcriptomes will provide clues to revealing their functions. Since experimental methods are time consuming and laborious for detecting RNA modifications, computational methods have been proposed for this aim in the past five years. However, there are some drawbacks for both experimental and computational methods in simultaneously identifying modifications occurred on different nucleotides.

RESULTS

To address such a challenge, in this article, we developed a new predictor called iMRM, which is able to simultaneously identify m6A, m5C, m1A, ψ and A-to-I modifications in Homo sapiens, Mus musculus and Saccharomyces cerevisiae. In iMRM, the feature selection technique was used to pick out the optimal features. The results from both 10-fold cross-validation and jackknife test demonstrated that the performance of iMRM is superior to existing methods for identifying RNA modifications.

AVAILABILITY AND IMPLEMENTATION

A user-friendly web server for iMRM was established at http://www.bioml.cn/XG_iRNA/home. The off-line command-line version is available at https://github.com/liukeweiaway/iMRM.

CONTACT

greatchen@ncst.edu.cn.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Computational Prediction of Protein Arginine Methylation Based on Composition-Transition-Distribution Features

Arginine methylation is one of the most essential protein post-translational modifications. Identifying the site of arginine methylation is a critical problem in biology research. Unfortunately, biological experiments such as mass spectrometry are expensive and time-consuming. Hence, predicting arginine methylation by machine learning is an alternative fast and efficient way. In this paper, we focus on the systematic characterization of arginine methylation with composition-transition-distribution (CTD) features. The presented framework consists of three stages. In the first stage, we extract CTD features from 1750 samples and exploit decision tree to generate accurate prediction. The accuracy of prediction can reach 96%. In the second stage, the support vector machine can predict the number of arginine methylation sites with 0.36 R-squared. In the third stage, experiments carried out with the updated arginine methylation site data set show that utilizing CTD features and adopting random forest as the classifier outperform previous methods. The accuracy of identification can reach 82.1 and 82.5% in single methylarginine and double methylarginine data sets, respectively. The discovery presented in this paper can be helpful for future research on arginine methylation.

T4SE-XGB: Interpretable Sequence-Based Prediction of Type IV Secreted Effectors Using eXtreme Gradient Boosting Algorithm

Type IV secreted effectors (T4SEs) can be translocated into the cytosol of host cells via type IV secretion system (T4SS) and cause diseases. However, experimental approaches to identify T4SEs are time- and resource-consuming, and the existing computational tools based on machine learning techniques have some obvious limitations such as the lack of interpretability in the prediction models. In this study, we proposed a new model, T4SE-XGB, which uses the eXtreme gradient boosting (XGBoost) algorithm for accurate identification of type IV effectors based on optimal features based on protein sequences. After trying 20 different types of features, the best performance was achieved when all features were fed into XGBoost by the 5-fold cross validation in comparison with other machine learning methods. Then, the ReliefF algorithm was adopted to get the optimal feature set on our dataset, which further improved the model performance. T4SE-XGB exhibited highest predictive performance on the independent test set and outperformed other published prediction tools. Furthermore, the SHAP method was used to interpret the contribution of features to model predictions. The identification of key features can contribute to improved understanding of multifactorial contributors to host-pathogen interactions and bacterial pathogenesis. In addition to type IV effector prediction, we believe that the proposed framework can provide instructive guidance for similar studies to construct prediction methods on related biological problems. The data and source code of this study can be freely accessed at https://github.com/CT001002/T4SE-XGB.

Meta-i6mA: an interspecies predictor for identifying DNA N6-methyladenine sites of plant genomes by exploiting informative features in an integrative machine-learning framework

DNA N6-methyladenine (6mA) represents important epigenetic modifications, which are responsible for various cellular processes. The accurate identification of 6mA sites is one of the challenging tasks in genome analysis, which leads to an understanding of their biological functions. To date, several species-specific machine learning (ML)-based models have been proposed, but majority of them did not test their model to other species. Hence, their practical application to other plant species is quite limited. In this study, we explored 10 different feature encoding schemes, with the goal of capturing key characteristics around 6mA sites. We selected five feature encoding schemes based on physicochemical and position-specific information that possesses high discriminative capability. The resultant feature sets were inputted to six commonly used ML methods (random forest, support vector machine, extremely randomized tree, logistic regression, naïve Bayes and AdaBoost). The Rosaceae genome was employed to train the above classifiers, which generated 30 baseline models. To integrate their individual strength, Meta-i6mA was proposed that combined the baseline models using the meta-predictor approach. In extensive independent test, Meta-i6mA showed high Matthews correlation coefficient values of 0.918, 0.827 and 0.635 on Rosaceae, rice and Arabidopsis thaliana, respectively and outperformed the existing predictors. We anticipate that the Meta-i6mA can be applied across different plant species. Furthermore, we developed an online user-friendly web server, which is available at http://kurata14.bio.kyutech.ac.jp/Meta-i6mA/.

DeepPred-SubMito: A Novel Submitochondrial Localization Predictor Based on Multi-Channel Convolutional Neural Network and Dataset Balancing Treatment

Mitochondrial proteins are physiologically active in different compartments, and their abnormal location will trigger the pathogenesis of human mitochondrial pathologies. Correctly identifying submitochondrial locations can provide information for disease pathogenesis and drug design. A mitochondrion has four submitochondrial compartments, the matrix, the outer membrane, the inner membrane, and the intermembrane space, but various existing studies ignored the intermembrane space. The majority of researchers used traditional machine learning methods for predicting mitochondrial protein localization. Those predictors required expert-level knowledge of biology to be encoded as features rather than allowing the underlying predictor to extract features through a data-driven procedure. Besides, few researchers have considered the imbalance in datasets. In this paper, we propose a novel end-to-end predictor employing deep neural networks, DeepPred-SubMito, for protein submitochondrial location prediction. First, we utilize random over-sampling to decrease the influence caused by unbalanced datasets. Next, we train a multi-channel bilayer convolutional neural network for multiple subsequences to learn high-level features. Third, the prediction result is outputted through the fully connected layer. The performance of the predictor is measured by 10-fold cross-validation and 5-fold cross-validation on the SM424-18 dataset and the SubMitoPred dataset, respectively. Experimental results show that the predictor outperforms state-of-the-art predictors. In addition, the prediction of results in the M983 dataset also confirmed its effectiveness in predicting submitochondrial locations.

Identification of Human Enzymes Using Amino Acid Composition and the Composition of k-Spaced Amino Acid Pairs

Enzymes are proteins that can efficiently catalyze specific biochemical reactions, and they are widely present in the human body. Developing an efficient method to identify human enzymes is vital to select enzymes from the vast number of human proteins and to investigate their functions. Nevertheless, only a limited amount of research has been conducted on the classification of human enzymes and nonenzymes. In this work, we developed a support vector machine- (SVM-) based predictor to classify human enzymes using the amino acid composition (AAC), the composition of k-spaced amino acid pairs (CKSAAP), and selected informative amino acid pairs through the use of a feature selection technique. A training dataset including 1117 human enzymes and 2099 nonenzymes and a test dataset including 684 human enzymes and 1270 nonenzymes were constructed to train and test the proposed model. The results of jackknife cross-validation showed that the overall accuracy was 76.46% for the training set and 76.21% for the test set, which are higher than the 72.6% achieved in previous research. Furthermore, various feature extraction methods and mainstream classifiers were compared in this task, and informative feature parameters of k-spaced amino acid pairs were selected and compared. The results suggest that our classifier can be used in human enzyme identification effectively and efficiently and can help to understand their functions and develop new drugs.

A novel riboswitch classification based on imbalanced sequences achieved by machine learning

Riboswitch, a part of regulatory mRNA (50-250nt in length), has two main classes: aptamer and expression platform. One of the main challenges raised during the classification of riboswitch is imbalanced data. That is a circumstance in which the records of a sequences of one group are very small compared to the others. Such circumstances lead classifier to ignore minority group and emphasize on majority ones, which results in a skewed classification. We considered sixteen riboswitch families, to be in accord with recent riboswitch classification work, that contain imbalanced sequences. The sequences were split into training and test set using a newly developed pipeline. From 5460 k-mers (k value 1 to 6) produced, 156 features were calculated based on CfsSubsetEval and BestFirst function found in WEKA 3.8. Statistically tested result was significantly difference between balanced and imbalanced sequences (p < 0.05). Besides, each algorithm also showed a significant difference in sensitivity, specificity, accuracy, and macro F-score when used in both groups (p < 0.05). Several k-mers clustered from heat map were discovered to have biological functions and motifs at the different positions like interior loops, terminal loops and helices. They were validated to have a biological function and some are riboswitch motifs. The analysis has discovered the importance of solving the challenges of majority bias analysis and overfitting. Presented results were generalized evaluation of both balanced and imbalanced models, which implies their ability of classifying, to classify novel riboswitches. The Python source code is available at https://github.com/Seasonsling/riboswitch.

A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features

The thermostability of proteins is a key factor considered during enzyme engineering, and finding a method that can identify thermophilic and non-thermophilic proteins will be helpful for enzyme design. In this study, we established a novel method combining mixed features and machine learning to achieve this recognition task. In this method, an amino acid reduction scheme was adopted to recode the amino acid sequence. Then, the physicochemical characteristics, auto-cross covariance (ACC), and reduced dipeptides were calculated and integrated to form a mixed feature set, which was processed using correlation analysis, feature selection, and principal component analysis (PCA) to remove redundant information. Finally, four machine learning methods and a dataset containing 500 random observations out of 915 thermophilic proteins and 500 random samples out of 793 non-thermophilic proteins were used to train and predict the data. The experimental results showed that 98.2% of thermophilic and non-thermophilic proteins were correctly identified using 10-fold cross-validation. Moreover, our analysis of the final reserved features and removed features yielded information about the crucial, unimportant and insensitive elements, it also provided essential information for enzyme design.

Extremely-randomized-tree-based Prediction of N6-Methyladenosine Sites in Saccharomyces cerevisiae

Introduction

N⁶-methyladenosine (m6A) is one of the most common post-transcriptional modifications in RNA, which has been related to several biological processes. The accurate prediction of m6A sites from RNA sequences is one of the challenging tasks in computational biology. Several computational methods utilizing machine-learning algorithms have been proposed that accelerate in silico screening of m6A sites, thereby drastically reducing the experimental time and labor costs involved.

Methodology

In this study, we proposed a novel computational predictor termed ERT-m6Apred, for the accurate prediction of m6A sites. To identify the feature encodings with more discriminative capability, we applied a two-step feature selection technique on seven different feature encodings and identified the corresponding optimal feature set.

Results

Subsequently, performance comparison of the corresponding optimal feature set-based extremely randomized tree model revealed that Pseudo k-tuple composition encoding, which includes 14 physicochemical properties significantly outperformed other encodings. Moreover, ERT-m6Apred achieved an accuracy of 78.84% during cross-validation analysis, which is comparatively better than recently reported predictors.

Conclusion

In summary, ERT-m6Apred predicts Saccharomyces cerevisiae m6A sites with higher accuracy, thus facilitating biological hypothesis generation and experimental validations.

Review and comparative analysis of machine learning-based phage virion protein identification methods

Phage virion protein (PVP) identification plays key role in elucidating relationships between phages and hosts. Moreover, PVP identification can facilitate the design of related biochemical entities. Recently, several machine learning approaches have emerged for this purpose and have shown their potential capacities. In this study, the proposed PVP identifiers are systemically reviewed, and the related algorithms and tools are comprehensively analyzed. We summarized the common framework of these PVP identifiers and constructed our own novel identifiers based upon the framework. Furthermore, we focus on a performance comparison of all PVP identifiers by using a training dataset and an independent dataset. Highlighting the pros and cons of these identifiers demonstrates that g-gap DPC (dipeptide composition) features are capable of representing characteristics of PVPs. Moreover, SVM (support vector machine) is proven to be the more effective classifier to distinguish PVPs and non-PVPs.

Prediction of Extracellular Matrix Proteins by Fusing Multiple Feature Information, Elastic Net, and Random Forest Algorithm

Extracellular matrix (ECM) proteins play an important role in a series of biological processes of cells. The study of ECM proteins is helpful to further comprehend their biological functions. We propose ECMP-RF (extracellular matrix proteins prediction by random forest) to predict ECM proteins. Firstly, the features of the protein sequence are extracted by combining encoding based on grouped weight, pseudo amino-acid composition, pseudo position-specific scoring matrix, a local descriptor, and an autocorrelation descriptor. Secondly, the synthetic minority oversampling technique (SMOTE) algorithm is employed to process the class imbalance data, and the elastic net (EN) is used to reduce the dimension of the feature vectors. Finally, the random forest (RF) classifier is used to predict the ECM proteins. Leave-one-out cross-validation shows that the balanced accuracy of the training and testing datasets is 97.3% and 97.9%, respectively. Compared with other state-of-the-art methods, ECMP-RF is significantly better than other predictors.