Bioinformatics Research and Applications: 17th International Symposium, ISBRA 2021, Shenzhen, China, November 26–28, 2021, Proceedings (Lecture Notes in Computer Science)

This book constitutes the proceedings of the 17th International Symposium on Bioinformatics Research and Applications, ISBRA 2021, held in Shenzhen, China, in November 2021. The 51 full papers presented in this book were carefully reviewed and selected from 135 submissions. They were organized in topical sections named: AI and disease; computational proteomics; biomedical imaging; drug screening and drug-drug interaction prediction; Biomedical data; sequencing data analysis. Preface Organization Contents AI and Disease MKL-LP: Predicting Disease-Associated Microbes with Multiple-Similarity Kernel Learning-Based Label Propagation 1 Introduction 2 Materials and Methods 2.1 Human Microbe-Disease Associations 2.2 Similarity Calculation 2.3 Multiple Kernel Learning 2.4 Label Propagation for Predicting Novel Associations 3 Results and Discussion 3.1 Parameter Selection 3.2 Performance Evaluation 3.3 Case Study 4 Conclusion References Immune-Microbiota Crosstalk Underlying Inflammatory Bowel Disease 1 Introduction 2 Materials and Methods 2.1 Sample and Data Description 2.2 Annotation of the 16S rDNA Sequencing Data of the Mucosa-Luminal Interface Microbiota 2.3 Microbial Community Detection 2.4 Proteome Data of the Host Tissue (Human Colon or Ileum) 2.5 Gene Set Enrichment Analysis 2.6 The Activeness of Each Gene Pathway 3 Results 3.1 Pathways Enriched in IBD Patients 3.2 OTU Communities Within Human Gut 3.3 Multiple Health Beneficial Genera Are Negatively Correlated with Inflammation-Relevant MHC Pathways 4 Conclusions References Epidemic Vulnerability Index for Effective Vaccine Distribution Against Pandemic 1 Introduction 2 Methodology 2.1 Mortality and Clinical Features 2.2 Infection Rate and Network Centrality 3 Experiment 4 Conclusion References Exploiting Multi-granular Features for the Enhanced Predictive Modeling of COPD Based on Chinese EMRs 1 Introduction 2 Methodology 2.1 Concept Recognition 2.2 Concept Relation Extraction 2.3 Attribute-Value Pair Extraction 2.4 Construction of a COPD Risk Prediction Model 3 Experiments 3.1 Datasets 3.2 Evaluation Criteria and Experimental Settings 3.3 Overall Evaluation Results 4 Conclusion References Task-Oriented Feature Representation for Spontaneous Speech of AD Patients 1 Introduction 2 Related Work 3 Problem Description 4 Model Architecture 4.1 The Encoding Network 4.2 Task Orientation 5 Experimentation and Discussion 5.1 Dataset 5.2 Training Parameters Selection 5.3 Experimental Result 5.4 Discussion 6 Conclusion and Future Work References Identification of Protein Markers Predictive of Drug-Specific Survival Outcome in Cancers 1 Introduction 2 Results 2.1 Significant Proteins Predictive of Drug-Specific Survival 2.2 Literature Support of Predictive Protein Markers 2.3 Correlation Between Predictive Proteins and Their Coding Genes 3 Materials and Methods 3.1 Data Access 3.2 Data Preprocessing 3.3 Survival Analysis 3.4 Literature Search 4 Conclusion References Diabetic Retinopathy Grading Base on Contrastive Learning and Semi-supervised Learning 1 Introduction 2 Method 2.1 Contrastive Learning Based DR Grading 2.2 Contrastive Learning Based Lesion Segmentation 2.3 Pseudo-label Attention and Deep Supervision 3 Experimental Results 3.1 Settings 3.2 Implementation Details 3.3 DR Grading with Contrastive Learning 3.4 Lesion Segmentation with Contrastive Learning 3.5 Comparison with SOTA Methods 4 Discussion References Reinforcement Learning for Diabetes Blood Glucose Control with Meal Information 1 Introduction 2 Related Works 3 The RL-Meal 3.1 Problem Definition 3.2 Details 3.3 Algorithm Description 4 Experiments and Results 4.1 Experimental Setup 4.2 Results 5 Conclusion References Predicting Microbe-Disease Association via Tripartite Network and Relation Graph Convolutional Network 1 Introduction 2 Material and Methods 2.1 Material 2.2 Methods 3 Experiments and Results 3.1 Model Analysis 3.2 Comparison with Other Methods 4 Case Studies 5 Conclusion References Combining Model-Based and Model-Free Reinforcement Learning Policies for More Efficient Sepsis Treatment 1 Introduction 2 Background 2.1 RL 2.2 Off-Policy Evaluation 3 Related Work 4 Methods 4.1 Data Preprocessing and Problem Formulation 4.2 The Policy Mixture Framework 5 Experiment Results 6 Conclusion References An Efficient Two-Stage Fusion Network for Computer-Aided Diagnosis of Diabetic Foot 1 Introduction 2 Related Works 2.1 Computer-Aided Diagnosis of Diabetic Foot 2.2 Machine Learning Methods in DFU Diagnosis 2.3 Deep Learning Methods in DFU Diagnosis 3 Method and Experiment 3.1 Fusion Network 3.2 Object Detection Module 3.3 Dataset 4 Results and Analysis 4.1 Experimental Results 4.2 Detailed Analysis 5 Conclusion and Discussion References A Heterogeneous Graph Convolutional Network-Based Deep Learning Model to Identify miRNA-Disease Association 1 Introduction 2 Materials 3 Methods 3.1 Heterogeneous Network Construction 3.2 Node Attributes Preparation 3.3 Node Embedding with Graph Convolution Network 3.4 Link Prediction from Embedding 4 Results 4.1 Randomly Zeroing Cross-Validation 4.2 Multi-column Zeroing Cross-Validation and Multi-row Zeroing Cross-Validation 5 Conclusion References Identification of Gastric Cancer Immune Microenvironment Related Genes with Poor Prognosis and Tumor Immune Infiltration 1 Introduction 2 Materials and Methods 2.1 GC Datasets and Immune-Related Genes Datasets (IRG) 2.2 Calculation of Immune Score and Stromal Score 2.3 Identification of Differentially Expressed Genes (DEGs) 2.4 Univariate Cox Survival Analysis 2.5 Random Forest Variable Selection 2.6 Reverse Transcription-Quantitative Polymerase Chain Reaction (RT-qPCR) 3 Results 3.1 Correlation of the Expression of CLRN3 with the Survival and Clinic-Pathological Staging of GC Patients 3.2 Correlation of CLRN3 with the Proportion of TICs 3.3 CLRN3 Had Potential to Be a Microenvironment Indicator 3.4 Experimental Verification of the Expression of CLRN3 3.5 The Transcription Factors (TFs) of CLRN3 4 Discussion 5 Conclusion References A k-mer Based Approach for SARS-CoV-2 Variant Identification 1 Introduction 2 Related Work 3 Proposed Approach 3.1 k-mers Computation 3.2 Kernel PCA 3.3 Machine Learning Classifiers 3.4 Baseline Model 4 Dataset Description and Preprocessing 5 Experimental Evaluation 5.1 Importance of Amino Acid Positions 6 Conclusion and Future Directions References A Novel Network Representation of SARS-CoV-2 Sequencing Data 1 Introduction 2 Methods 3 Results 3.1 Datasets 3.2 Assortativity Analysis 3.3 Transmission Network Analysis 3.4 Scalability Analysis 4 Conclusion References Computational Proteomics A Sequence-Based Antibody Paratope Prediction Model Through Combing Local-Global Information and Partner Features 1 Introduction 2 Materials and Methods 2.1 Datasets 2.2 Input Features 2.3 Input Representation 2.4 Model Architecture 2.5 Training Details 3 Results 3.1 Model Results 3.2 Effect of Global Features and Partner Features 4 Conclusion References SuccSPred: Succinylation Sites Prediction Using Fused Feature Representation and Ranking Method 1 Introduction 2 Material and Method 2.1 Dataset 2.2 Feature Representation 2.3 Computational Model 2.4 Performance Evaluation 3 Experiment 4 Conclusion References BindTransNet: A Transferable Transformer-Based Architecture for Cross-Cell Type DNA-Protein Binding Sites Prediction 1 Introduction 2 Method 2.1 Feature Representation 2.2 Model Architecture 2.3 Transfer Learning 2.4 Model Implementation 3 Experiment 3.1 Experiment Setup 3.2 Experiment Results and Analysis 4 Conclusion References Overlapping Protein Complexes Detection Based on Multi-level Topological Similarities 1 Introduction 2 Materials 3 Methods 3.1 Constructing Local Similarity Matrix 3.2 Constructing Global Similarity Matrix 3.3 Detecting Protein Complexes Based on Core-attachment 4 Results 4.1 Evaluation Metrics 4.2 Comparison with Other Methods 5 Conclusions References LPI-FKLGCN: Predicting LncRNA-Protein Interactions Through Fast Kernel Learning and Graph Convolutional Network 1 Introduction 2 Materials and Methods 2.1 Generate Multiple Base Kernels for LncRNAs and Proteins 2.2 Kernel Fusion by Fast Kernel Learning 2.3 Construction of a Heterogeneous LncRNA-Protein Network 2.4 Encoding by Multi-Layer Graph Convolution Network 2.5 Decoding and Final Prediction 3 Results 3.1 Datasets and Experimental Environments 3.2 Parameter Impact Analysis 3.3 The Influence of FKL and GCN in LPI Prediction 3.4 Comparison with the Baseline Methods 3.5 Case Study 4 Discussion References Biomedical Imaging Prediction of Protein Subcellular Localization from Microscopic Images via Few-Shot Learning 1 Introduction 2 Methods 2.1 Few-Shot Learning 2.2 Contrastive Representation Learning 2.3 Loss Functions 3 The Backbone CNN Model 4 Results 4.1 Dataset 4.2 Experimental Settings 4.3 Evaluation 4.4 Batch Selection 4.5 Prediction Results of ArcNet 4.6 Comparison Between ContrastiveNet and ArcNet 5 Conclusions and Future Work References A Novel Pseudo-Labeling Approach for Cell Detection Based on Adaptive Threshold 1 Introduction 2 Method 2.1 Method Overview 2.2 Pseudo-Labeling Based on Adaptive Threshold (PLAT) 2.3 Retrain and Patch Attention 3 Experiment 3.1 Dataset 3.2 Experiment Detail 3.3 Experiment Result 3.4 Ablation Study 3.5 Comparison to Previous Work 4 Conclusion References Parameter Transfer Learning Measured by Image Similarity to Detect CT of COVID-19 1 Introduction 2 Related Work 2.1 Image and Conclusion Model 2.2 Attention Model 2.3 Activate Function 3 Method 3.1 Parameter Transfer Learning 3.2 Parallel Channel and Spatial Attention Mechanism 3.3 Swish Activation Function 4 Experiment 4.1 Dataset and Code Availability 4.2 Experimental Results 5 Conclusion References A Novel Prediction Framework for Two-Year Stroke Recurrence Using Retinal Images 1 Introduction 2 Materials and Methods 2.1 Data Collection 2.2 Data Preprocessing 2.3 Feature Selection 2.4 Classification Methods 3 Results and Discussion 3.1 Experiment Environment 3.2 Prediction Performance for Two-Year Stroke Recurrence 4 Conclusion References The Classification System and Biomarkers for Autism Spectrum Disorder: A Machine Learning Approach 1 Introduction 2 Materials and Methods 2.1 Datasets 2.2 Data Preprocessing 2.3 Feature Selection 2.4 Model Construction and Analysis 3 Results and Discussions 3.1 The Performance of Classification System on the Single Site Dataset 3.2 The Performance of Classification System on Dataset of the Multi-sites 3.3 Analysis About the Biomarker of ASD 4 Conclusions References LiteTrans: Reconstruct Transformer with Convolution for Medical Image Segmentation 1 Introduction 2 Related Work 3 Method 3.1 Revisit Convolution and Self-attention 3.2 Local-Global Self-Attention (LGSA) 3.3 Multi-branch Module Composed of Convolution and LGSA 3.4 LiteTrans: Reconstruct Transformer with Multi-branch Module 4 Experiments 4.1 Datasets 4.2 Implementation Details 4.3 Experiment Results 4.4 Ablation Study 5 Conclusion References CFCN: A Multi-scale Fully Convolutional Network with Dilated Convolution for Nuclei Classification and Localization 1 Introduction 2 Methods 2.1 CFCN Architecture 2.2 Focal Loss for Multi-class Nuclei Classification 3 Experiments 3.1 Dataset 3.2 Implementation Details 3.3 Results 4 Conclusions References Image to Image Transfer Makes Malpositioned Teeth Orderly 1 Introduction 2 Related Works 2.1 Generative Adversarial Network 2.2 Image-to-Image Translation 2.3 Text to Image Synthesis 2.4 Image to Text Generation 3 Dataset 4 Our Method: Orthod-GAN 4.1 Malpositioned-Teeth-Image to Malpositioned-Teeth-Code 4.2 Malpositioned-Teeth-Code to Neat-Teeth-Code 4.3 Neat-Teeth-Code to Neat-Teeth-Image 5 Experiments 5.1 Baselines 5.2 Comparison with Baselines 6 Conclusion References MIFS: A Peer-to-Peer Medical Images Storage and Sharing System Based on Consortium Blockchain 1 Introduction 2 Related Work 2.1 PACS 2.2 Blockchain-Based Storage System 3 Specific Design 3.1 System Architecture 3.2 Authentication and Access Control 3.3 Peer-to-Peer File Storage 3.4 Retrieval, Authorization, and Sharing 4 Implementation and Evaluation 4.1 System Implementation 4.2 Performance 5 Conclusion References StarLace: Nested Visualization of Temporal Brain Connectivity Data 1 Introduction 2 Related Work 3 Proposed Technology 3.1 Motivation 3.2 System Overview 3.3 Pipeline of Data Processing 3.4 Visual Design 4 Result 5 Expert Evaluation 6 Conclusion References Batch Weighted Nuclear-Norm Minimization for Medical Image Sequence Segmentation 1 Introduction 2 Methodology 2.1 Background and Notations 2.2 Loss Function 2.3 Measuring Similarity with Matrix Rank 2.4 Weighted Nuclear Norm Minimization 2.5 Sequential Segmentation with WNNM-Based Regularizer 3 Experimental Results 3.1 Synthetic Dataset 3.2 Ultrasound Tongue Contour Extraction 3.3 LiTS 2017 CT Dataset 3.4 BraTS 2019 Dataset 3.5 Effect of 4 Conclusion References Drug Screening and Drug-Drug Interaction Prediction Predicting Drug Drug Interactions by Signed Graph Filtering-Based Convolutional Networks 1 Introduction 2 Signed DDIs Networks and Prediction Problems 3 The Proposed SGFCN for DDIs Prediction 3.1 Graph Convolutional Networks and Graph Filtering 3.2 Drug Representation via Signed Graph Filtering-Based Convolutional Networks (SGFCN) 3.3 The End-to-End Learning Framework for DDIs on Signed Networks 4 Experimental Settings and Analyses 4.1 Datasets and Signed DDIs Networks 4.2 Training Settings 4.3 Results and Analyses 5 Conclusions References Drug-Target Interaction Prediction Based on Gaussian Interaction Profile and Information Entropy 1 Introduction 2 Related Works 3 Method 3.1 Problem Formalization 3.2 The Gaussian Interaction Profile Similarity Improvement 3.3 Similarity Fusion Coefficient Based on Information Entropy 3.4 Similarity Fusion 4 Result Analysis and Discussion 4.1 Datasets 4.2 Evaluation Metrics and Experimental Setup 4.3 Results 5 Conclusion References A Deep Learning Approach Based on Feature Reconstruction and Multi-dimensional Attention Mechanism for Drug-Drug Interaction Prediction 1 Introduction 2 Methods 2.1 Data Description and Preprocessing 2.2 Feature Reconstruction 2.3 Attention-DNN Construction 3 Experiments and Results 3.1 Evaluation Metrics 3.2 Experimental Setup 3.3 Influence of Feature Reconstruction 3.4 Influence of Multi-dimensional Attention Mechanism 3.5 Comparison with Existing State-of-the-Art Methods 4 Conclusion References OrgaNet: A Deep Learning Approach for Automated Evaluation of Organoids Viability in Drug Screening 1 Introduction 2 Related Work 3 Proposed Method 3.1 Dataset Construction 3.2 Problem Formulation 3.3 OrgaNet and Loss Function 4 Experiments and Results 4.1 Evaluation Metrics 4.2 Results 5 Conclusion References HGDD: A Drug-Disease High-Order Association Information Extraction Method for Drug Repurposing via Hypergraph 1 Introduction 2 Methods 2.1 Hypergraph Introduction 2.2 Overview 2.3 Initialization of Drug-Disease Hypergraph 2.4 Extraction of High-Order Information from Hypergraph 2.5 Association Prediction 3 Experiment and Discussion 3.1 Datasets and Settings 3.2 Comparison with Other Models 3.3 Case Study 4 Conclusion References IDOS: Improved D3DOCK on Spark 1 Introduction 2 Distributed (Spark) Implementation 2.1 Cluster Design 2.2 Receptor Data Broadcast and Assemble 2.3 Ligand Data Division and Task Distribution 2.4 Parallelization 2.5 Docking and Aggregate 2.6 Some Positive Measures 3 Results 3.1 Experimental Setup 3.2 The Accuracy 3.3 Speed Up Performance 3.4 Scalability 4 Conclusion and Discussion References Biomedical Data A New Deep Learning Training Scheme: Application to Biomedical Data 1 Introduction 2 Successive k-fold Cross-Validation 3 Experiments and Results 3.1 Experimental Datasets 3.2 Benchmark Methods 3.3 Experimental Settings 3.4 Drug-Target Affinity Prediction 3.5 Tumor Segmentation 3.6 Image Classification 4 Conclusion and Future Work References EEG-Based Emotion Recognition Fusing Spacial-Frequency Domain Features and Data-Driven Spectrogram-Like Features 1 Introduction 2 Methodology 2.1 Spacial-Frequency Matrix 2.2 Scaling Convolutional Layer 2.3 Proposed SFMS-Net 3 Experiments 3.1 Dataset 3.2 Experiment Settings and Evaluation 3.3 Experimental Results 3.4 Ablation Studies 4 Conclusion References ECG Arrhythmia Detection Based on Hidden Attention Residual Neural Network 1 Introduction 2 Methods 2.1 Data Prepocessing 2.2 Feature Extraction 2.3 Classification 3 Experiment 3.1 Data Source 3.2 Implementation Details 3.3 Results 4 Discussion 5 Conclusion References EEG-Based Depression Detection with a Synthesis-Based Data Augmentation Strategy 1 Introduction 2 Related Work 3 Method 3.1 Individual Diversity Sampling 3.2 Signal Decomposition and Recombination 4 Experiments and Results 4.1 Data Acquisition and Preprocessing 4.2 Experimental Settings 4.3 Comparison with Mainstream Signal Synthesis Methods 4.4 Comparison with Other Subject Sampling Strategies 5 Conclusion References Sequencing Data Analysis Joint CC and Bimax: A Biclustering Method for Single-Cell RNA-Seq Data Analysis 1 Introduction 2 Related Works 2.1 Bimax Model 2.2 The Mean Square Residual 3 JCB Algorithm 4 Experiments 4.1 Datasets 4.2 Evaluation Metric 4.3 Parameters Selection 4.4 Biclustering Results 5 Conclusion References Improving Protein-protein Interaction Prediction by Incorporating 3D Genome Information 1 Introduction 2 Methods 2.1 PPI and 3D Genome Data 2.2 Chromatin 3D Modeling 2.3 Encoding Scheme and Prediction Methods 3 Results 4 Discussion References Boosting Metagenomic Classification with Reads Overlap Graphs 1 Introduction 2 Methods 2.1 Pre-processing 2.2 Graph Transformation and Reduction 2.3 Label Propagation 3 Results 3.1 Performance Evaluations 4 Conclusions References ScDA: A Denoising AutoEncoder Based Dimensionality Reduction for Single-cell RNA-seq Data 1 Introduction 2 Methods and Materials 2.1 ScDA 2.2 Evaluation Metric 2.3 Datasets 3 Results 3.1 Comparison of the Clustering Performance 3.2 Comparison of the Run Time 4 Conclusion References Others PickerOptimizer: A Deep Learning-Based Particle Optimizer for Cryo-Electron Microscopy Particle-Picking Algorithms 1 Introduction 2 Materials and Methods 2.1 Classification Datasets 2.2 Algorithm 2.3 Training Details 2.4 Evaluation Metrics 3 Results 3.1 Classification Performance 3.2 Comparison with Other Pruning Approaches 3.3 Use Case 4 Conclusions References SkeIn: Sketchy-Intensive Reading Comprehension Model for Multi-choice Biomedical Questions 1 Introduction 2 Dataset 2.1 Task Formulation 2.2 Data Analysis 3 Method 3.1 Encoder 3.2 Dual-Path Multi-Head Co-Attention 3.3 Decoder 4 Experiment 4.1 Experimental Settings 4.2 Results 5 Conclusion References DNA Image Storage Using a Scheme Based on Fuzzy Matching on Natural Genome 1 Introduction 2 Data and Software Availability 2.1 Image Dataset 2.2 Genome Files and Source Code 2.3 Testing Environment 3 Methods 3.1 Overview of Genome-Based DNA Storage System 3.2 Transforming Nucleotide Sequences into Digital Sequences 3.3 Image Denoising 3.4 Image Super-Resolution 3.5 The Compression Rate 4 Result 4.1 The Comparison Among Three Transforming Rules 4.2 Image Processing 4.3 The Compression Rate of the Genome-Based DNA Storage System 4.4 Factors to Influence the Result of Image’s Fuzzy Matching in Genomes 5 Conclusion References Prediction of Virus-Receptor Interactions Based on Similarity and Matrix Completion 1 Introduction 2 Materials 3 PreVRIs for Discovering Virus-Receptor Interactions 3.1 Virus Similarity Kernel 3.2 Receptor Similarity Kernel 3.3 Similarity Kernel Learning 3.4 Matrix Completion for Discovering Virus-Receptor Interactions 4 Experimental Results and Discussion 4.1 Performance Evaluation 4.2 Comparison with Related Methods 4.3 Case Study 5 Conclusion References An Efficient Greedy Incremental Sequence Clustering Algorithm 1 Introduction 2 Background 3 Methods and Evaluation Matrices 3.1 Pre-filtering 3.2 Modified Short Word Filtering 3.3 Data Packing 3.4 Parallelization 3.5 Evaluation Metrics 4 Result and Discussion 4.1 Datasets and Experimental Setting 4.2 Performance Improvement by Filtering 4.3 Performance Improvement by Data Packing 4.4 Heterogeneous Parallelism 4.5 Comparison with Other Clustering Tools 5 Conclusion References Correlated Evolution in the Small Parsimony Framework 1 Introduction 2 Small Parsimony 3 Correlation 4 Experimental Analysis 5 Discussion References Author Index This book constitutes the refereed proceedings of the 9th International Symposium on Bioinformatics Research and Applications, ISBRA 2013, held in Charlotte, NC, USA, in May 2013. The 25 revised full papers presented together with 4 invited talks were carefully reviewed and selected from 46 submissions. The papers cover a wide range of biomedical databases and data integration, high-performance bio-computing, biomolecular imaging, high-throughput sequencing data analysis, bio-ontologies, molecular evolution, comparative genomics and phylogenomics, molecular modeling and simulation, pattern discovery and classification, computational proteomics, population genetics, data mining and visualization, software tools and applications.