Cyber Threat Intelligence (CTI) helps organizations understand the tactics, techniques, and procedures used by potential cyber criminals to defend against cyber threats. To protect the core systems and services of organizations, security analysts must analyze information about threats and vulnerabilities. However, analyzing large amounts of data requires significant time and effort. To streamline this process, we propose an enhanced architecture, BERT-CRF, by removing the BiLSTM layer from the conventional BERT-BiLSTM-CRF model. This model leverages the strengths of deep learning-based language models to extract critical threat intelligence and novel information from threats effectively. In our BERT-CRF model, the token embeddings generated by BERT are directly fed into the Conditional Random Field (CRF) layer for efficient Named Entity Recognition (NER), thus preventing the need for an intermediate BiLSTM layer. We train and evaluate the model with three publicly available threat entity databases. We also collect open-source threat intelligence data from recent years for evaluating the applicability of the constructed model in a real-world environment. Furthermore, we compare our model with the most popular GPT-3.5 and the most downloaded open-source BERT question-and-answer models. Through this study, our proposed model demonstrated robust usability and outperformed other models, signifying its potential for application in CTI. In a real-world scenario, our model achieved an accuracy of 82.64%, while with malware-specific threat intelligence data, it achieved an impressive accuracy of 93.95%. The code for this research is publicly available at https://github.com/stwater20/ner_bert_crf_open_version.

DNA methylation plays an important role in the regulation of gene expression, and aberrant changes in epigenetic regulation can be detected in cancer development or progress. In this study, genome-wide DNA methylation profiles and electronic medical records were combined to identify effective DNA methylation biomarkers for colorectal cancer (CRC). Statistical analytics and deep learning approaches were integrated to explore associated novel biomarkers. These identified biomarkers could facilitate accurate diagnosis of CRC and monitor disease progression for a testing subject, and provide suggestions for appropriate medical treatments. Firstly, DNA methylation profiles were analyzed through a standard pipeline to discover significantly differential methylation biomarkers as primary biomarkers. Incorporating Taiwan’s National Health Insurance Research Database (NHIRD), associated comorbidities of CRC were identified and associated disease genes were considered as secondary biomarkers. The intersection of primary and secondary biomarkers was performed to obtain CRC-relevant biomarker candidates, and gene ontology annotations were used to calculate candidate biomarker-to-biomarker distances. Based on the formulated gene-pair distance matrix for all candidate biomarkers, we applied clustering algorithms to categorize candidate biomarkers into different functional groups. Corresponding coefficients for each biomarker within a functional cluster were ranked through an attention-based recurrent neural network. The weighting coefficient for each biomarker was further taken as a reference for designing a practical testing toolkit. Finally, we obtained 11 important biomarkers from 5 functional clustered groups which were established by the 141 biomarker candidates screened from 450k DNA methylation probes.

The short-term mortality risk of Acute Myocardial Infarction (AMI) remains high all over the world. Accurately predicting risk of death may assist physicians’ decisions on medical treatments and advance care planning for patients. The Taiwan National Health Insurance Research Database (NHIRD) was applied to construct a model to predict whether a diagnosed AMI patient could turn into a severe condition within a short interval. We used IQR and propensity score matching to select control group subjects, and Kaplan Meier method to define a duration boundary for distinguishing two patient groups. Both chi-square test and p-value were applied to evaluate and screen statistical significance of each feature. An XGBoost model was applied to construct a prediction model and compare prediction results at different stages. System performance was optimized by tuning parameters based on increasing recall rate as much as possible. The proposed prediction system can determine mortality risk of patients immediately and prevent delays in treatment. The results showed that although various attempts were tried to optimize the system with different feature combinations, the prediction performance was yet unsatisfied due to NHIRD providing limited longitudinal diagnosed records and medicine information only. No biometric feature was collected in the NHIRD and it only achieved a low recall rate of 56%. It is firmly believed that features obtained from biometric screening test could provide important factors for a better prediction system and precision treatment in practical clinical applications.

BACKGROUND: The correct establishment of the barcode classification system for fish can facilitate biotaxonomists to distinguish fish species, and it can help the government to verify the authenticity of the ingredients of fish products or identify unknown fish related samples. The Cytochrome c oxidation I (COI) gene sequence in the mitochondria of each species possesses unique characteristics, which has been widely used as barcodes in identifying species in recent years. Instead of using COI gene sequences for primer design, flanking tRNA segments of COI genes from 2618 complete fish mitochondrial genomes were analyzed to discover suitable primers for fish classification at taxonomic family level. The minimal number of primer sets is designed to effectively distinguish various clustered groups of fish species for identification applications. Sequence alignment analysis and cross tRNA segment comparisons were applied to check and ensure the primers for each cluster group are exclusive. RESULTS: Two approaches were applied to improve primer design and re-cluster fish species. The results have shown that exclusive primers for 2618 fish species were successfully discovered through in silico analysis. In addition, we applied sequence alignment analysis to confirm that each pair of primers can successfully identify all collected fish species at the taxonomic family levels. CONCLUSIONS: This study provided a practical strategy to discover unique primers for each fishery species and a comprehensive list of exclusive primers for extracting COI barcode sequences of all known fishery species. Various applications of verification of fish products or identification of unknown fish species could be effectively achieved.

Heart failure is one of the most common causes of death for all different racial groups all over the world. It is noticed that a high proportion of patients diagnosed with heart failure (HF) within a short period of time after suffering from myocardial infarction (MI). This study is designed to translate existing real world data to real world evidence by exploring associations between historical comorbidities and heart failure diseases. Machine learning technologies were applied to predict whether patients with myocardial infarction would develop heart failure within a specific time period, and to remind patients how to strengthen personal self-care to avoid the transition towards heart failure or postpone the occurrence of the preceding events. In this study, patients with heart failure after myocardial infarction were divided into two groups according to a median age of 71 years old, and corresponding prediction models were constructed for two different age groups respectively. Three different machine learning technologies, namely logistic regression, random forest, and XGBoost were used to construct prediction models and a 5-fold cross-validation was applied to evaluate prediction accuracy and stability of prediction models. The results of our proposed method reveal that if a prediction model was constructed without age stratification, the constructed prediction model provided inferior performance compared to stratified groups by employing identical features. The analytical results from three different machine learning techniques consistently supported that the prediction models of myocardial infarction resulted in accelerated transition towards heart failure within a specific interval should be constructed by stratifying age groups first, and then training the corresponding data for better system performance.

BACKGROUND: The Iridoviridae family is categorized into five genera and clustered into two subfamilies: Alphairidovirinae includes Lymphocystivirus, Ranavirus (GIV), and Megalocystivirus (TGIV), which infect vertebrate hosts and Betairidovirinae includes Iridovirus and Chloriridovirus, which infect invertebrate hosts. Clustered Iridoviridae subfamilies possess host-specific characteristics, which can be considered as exclusive features for in-silico prediction of effective epitopes for vaccine development. A voting mechanism-based linear epitope (LE) prediction system was applied to identify and endorse LE candidates with a minimum length requirement for each clustered subfamily RESULTS: The experimental results showed that four conserved epitopes among the Iridovirideae family, one exclusive epitope for invertebrate subfamily and two exclusive epitopes for vertebrate family were predicted. These predicted LE candidates were further validated by ELISA assays for evaluating the strength of antigenicity and cross antigenicity. The conserved LEs for Iridoviridae family reflected high antigenicity responses for the two subfamilies, while exclusive LEs reflected high antigenicity responses only for the host-specific subfamily CONCLUSIONS: Host-specific characteristics are important features and constraints for effective epitope prediction. Our proposed voting mechanism based system provides a novel approach for in silico LE prediction prior to vaccine development, and it is especially powerful for analyzing antigen sequences with exclusive features between two clustered groups.

BACKGROUND: Transcriptomic sequencing (RNA-seq) related applications allow for rapid explorations due to their high-throughput and relatively fast experimental capabilities, providing unprecedented progress in gene functional annotation, gene regulation analysis, and environmental factor verification. However, with increasing amounts of sequenced reads and reference model species, the selection of appropriate reference species for gene annotation has become a new challenge. METHODS: We proposed a novel approach for finding the most effective reference model species through taxonomic associations and ultra-conserved orthologous (UCO) gene comparisons among species. An online system for multiple species selection (MSS) for RNA-seq differential expression analysis was developed, and comprehensive genomic annotations from 291 reference model eukaryotic species were retrieved from the RefSeq, KEGG, and UniProt databases. RESULTS: Using the proposed MSS pipeline, gene ontology and biological pathway enrichment analysis can be efficiently achieved, especially in the case of transcriptomic analysis of non-model organisms. The results showed that the proposed method solved problems related to limitations in annotation information and provided a roughly twenty-fold reduction in computational time, resulting in more accurate results than those of traditional approaches of using a single model reference species or the large non-redundant reference database. CONCLUSIONS: Selection of appropriate reference model species helps to reduce missing annotation information, allowing for more comprehensive results than those obtained with a single model reference species. In addition, adequate model species selection reduces the computational time significantly while retaining the same order of accuracy. The proposed system indeed provides superior performance by selecting appropriate multiple species for transcriptomic analysis compared to traditional approaches.

Heart failure (HF) is one of the most common causes of hospitalization for people over 65 years old, and over half of patients who diagnosed with severe HF conditions at first time cannot survive over 5 years. It is also noticed that rehospitalization rate of heart failure patients may increase to 50% in three months and the mortality rate from 33% to 50% within five years if HF patients are not treated with proper medication and physical therapies. Here we provide a classification system and an early warning mechanism for detecting HF disease based on integrating 6-minute walking test (6MWT), Internet of medical thing devices, and cloud computing technologies. This study performed 6MWTs for 50 HF patients accompanied by medical staffs for recording walkway distance, walking heart rate, and resting heart rate. All retrieved features and classified functional levels of heart organ of HF patients are trained as a target referencing dataset. According to the selected features and trained results, the clustered information representing various heart conditions is applied for detecting potential HF patients at earlier stages. In addition, the newly obtained self-exercise rehabilitation records from HF patients in a real-time manner will be compared to her/his previous 6MWT patterns. The compared differences are considered as important information for doctors to arrange medical treatment and adjusted physical therapy during the next follow-up visit in hospital.

Next Generation Sequencing technology has become an efficient approach for analyzing functional genomics and facilitates discovery of molecular biomarkers for various biological applications. This study is aim to discover all simple sequence repeat (SSR) biomarker candidates and associated genetic functions from two different tilapia strains. The two transcriptome datasets, Nile Tilapia and Mozambique Tilapia, were retrieved from the public SRA database for transcriptome assembly and functional biomarker discovery. We adopted Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases for gene function annotations. A total of 116,781 SSRs and 126,324 SSRs were identified from these two tilapia strains, and there were 31,950 SSR variations could be detected within annotated genes. To illustrate effective and efficient performance of our proposed system, we used the keywords of “skeletal system development” as an example for discovering associated SSR markers. There are three genes annotated with function of skeletal system development, and which possess 5 different SSRs within the gene sequences. Under such a particular functional constraint, a gene cluster can be automatically defined and all linked polymorphic SSRs could be identified between the two different species. These identify linked SSR markers are important for further analysis on genetic regulation, genetic disease detection, gene therapy, or cultivation of various agricultural products.

Molecular biomarkers are important and commonly used as classification features for detecting function performances of target genes and distinguishing characteristics of different species or strains. As expeditious development in Next Generation Sequencing technologies, whole genome sequencing delivers a comprehensive view and new insights for novel species. However, utilization of whole genome sequences to discover genetic biomarkers for specific functional gene groups still possess many challenges. To solve this problem, we utilized Gene Ontology definitions and Ensembl orthologous information to decide a set of associated functional genes according to user's query keywords. Then, previously assembled NGS contigs were matched against the selected gene set from model species, and an SSR searching algorithm and a cross-species comparison mechanism were performed to reveal potential polymorphic SSR biomarkers. Each identified SSR would be annotated and classified by important attributes, such as lengths, genetic locations, fundamental repeat patterns, tolerant rates of retrieved SSRs. The developed system could identify novel SSR biomarkers for a specified functional gene group regarding conserved and unique features between the query and the target model species, and which could facilitate biologists in finding potential markers for elucidating biological functions and distinguishable features of a novel species.

For de novo transcriptome analysis, choosing a closest reference model specie in terms of evolutionary distance is a general approach for gene mapping and genome annotations. However, not every selected reference model species possesses comprehensive genome annotations and curated information, and the total number of mapped genes from the selected reference species could not be fully expected either. Due to inefficient mapped genes from the selected reference model species, the following functional pathway analysis on transcriptome datasets would be seriously affected. To solve this problem, we proposed an improved approach based on multiple reference model species selection, especially for KEGG pathway analysis on differentially expressed genes. Applying union operations on individually mapped genes from different selected species, we could significantly promote the integrity of gene mapping results in KEGG pathways and provide realistic P-values for each identified pathway. Furthermore, based on mapped genes and KGML datasets, we applied various gray-levels, colors and shapes to present gene expression conditions on each biological pathway. Taking NGS transcriptomic datasets from an unknown Antarctic green alga species as an experimental example and selecting three published known species including Chlamydomonas reinhardtii, Chlorella variabilis, and Coccomyxa subellipsoidea as candidate reference species, we compared the results of pathway enrichment analysis by adopting different selections of reference species. We found that integrating all mapped genes from various model species provided a better result compared to using any single reference species. Some missed important biological pathways could be retrieved under an identical threshold setting of P-value, such as Ribosome, Pyrimidine metabolism and ABC transporters pathways. Therefore, we believe appropriate selection of multiple reference species is necessary and significant for transcriptome analysis on de novo species.

Heart failure (HF) is a major public health problem with an increasing prevalence that has tremendous impact to patients all over the world. However, the signs and symptoms of HF in the early stages are not clear, so it is relatively difficult to prevent or predict. HF is also one of complicated diseases, there are yet no strict standards to classify various types of HF patients based on various tendencies in each cause. To discover new evidences and associated medications for each type of HF according to some known features, we could divide HF patients into four major subgroups including systolic, diastolic, valvular, and non-specified types. We have performed statistical analysis and verified prediction results by several selected features for each clustered group from CGMH patient medical records. The constructed reference models could provide necessary measurements and prognosis indications for HF patients to prevent from heart deterioration. Based on automatic medication record analysis, we applied SVM tools to train and classify the associations for different HF types. The predicted results achieved an accuracy rate of 75.26% through a 10-folds cross validation mechanism. In addition, the proposed system effectively predicted patient's survival year and life expectancy with accuracy rates between 80%-87% under different parameter settings.

Ubiquitous Internet environment and matured key technologies of sensing devices provide diversified applications of Internet of Thing (IoT). IoT in intelligent healthcare is one of the major topics for convincing powerful functions of IoT regarding the promotion of national health policies, strategies, and plans. However, according to statistics reports, annual global mortality for cardiovascular diseases remained as high as 29%, of which coronary artery disease (CAD) is accounted for 42%. Unfortunately, almost half of the patients not knowing they are at risk of cardiovascular diseases until serious symptoms occurred. Here, we proposed an early warning system by integrating IoT, big data, and cloud computing technologies which physically linked among personal communication devices of patient/doctor, cloud systems, and hospital medical information systems. The system collects personal heart rate and metabolic equivalent (MET) features from calibrated fitness devices wore by users, and corresponding fitted curves are dynamically calculated and compared to previously trained curve patterns from 213 patients with heart diseases and 124 healthy individuals. The prototype system was already successfully developed and validated by a couple of testing cases with excellent performance. To ensure the robustness of the proposed system, now we are evaluating system reliability and trying to collect more testing samples for further validating effectiveness of the proactive healthcare mechanism.

Grouper is one of the most important aquaculture fish in the world. Due to high density of grouper farming, the aquaculture industry is threatened by iridovirus infections in recent years. In this study, we used next-generation sequencing technologies to analyze gene differential expression of groupers infected by two different types of iridoviruses, including Grouper Iridovirus of Taiwan (TGIV) of Megalocytivirus and Grouper Iridovirus (GIV) of Ranavirus. Zebra fish (Danio rerio) was selected as the reference model for gene functional annotations and pathway analyses. According to different fold change levels and identified associated KEGG pathways from defined gene sets, we found exclusive biological characteristics of groupers infected by two different iridoviruses. The results have shown that two specific gene sets, including the first set of PSMB2, CASP3, and U2AF2 for TGIV and the second set of PTGES, PDE3, and THBS1 for GIV were significantly and exclusively expressed. Furthermore, we also found two important pathways, Proteasome and ECM-receptor interaction pathways, were identified by employing gene set/pathway enrichment analysis. These results could facilitate biologists realizing comprehensive gene regulation of groupers infected by different iridoviruses and developing specific vaccines for grouper aquaculture.

Gene expression data analysis from high throughput approaches is complicated due to its abundant scales. It is especially difficult to analyze and categorize temporal differential expression of gene groups and discover interactive relationship among genes from single gene observation. Hence, integration of temporal gene expression data and biological pathway information may provide a powerful approach for understanding dynamic gene regulatory networks. In this study, we adopted the expression data from a series of zebra fish embryo developmental stages, and integrated with KEGG biological pathways to display dynamic gene expression through both heat map and 3D visualization methods in a temporal version. Through filtering analysis, all drastically varied gene expression could be identified at different time and it reflected how these key genes play important roles during embryogenic development. The designed system could retrieve important information for temporal gene expression constrained to a specific group of genes.

A conformational epitope (CE) in an antigenic protein is composed of amino acid residues that are spatially near each other on the antigen's surface but discontinuously appeared in sequence. CEs bind their complementary Para topes which are located on the surface areas of B-cell receptors and/or antibodies. CE predication is a challenging research topic, which plays an important and fundamental role for vaccine design and immuno-biological experiments. The aim of this study is to primarily analyze CE epitopes of 163 non-redundant antigen structures based on physical and chemical properties of surface residues. In terms of physical properties, distinguishable geometric features of solid angle and surface rate of amino acids are observed, while for chemical properties, higher energy, hydrophilicity and polarity from verified epitope regions are discovered as well. All these recognized features from verified CEs were thoroughly analyzed and compared for supporting a CE prediction system.

One of the most challenging research topics in developing application software for computational immunology is to predict B-cell epitopes on antigenic protein structural surfaces correctly. Although there have been long-term research history in both linear and conformational epitope prediction, it is yet far from being satisfied for perfect solutions. Especially, several developed systems in the past few years for predicting conformational epitopes neither reach high-accuracy performance, nor for efficient simulations. Therefore, an effective and efficient prediction tool for epitope analysis plays an important role for growth and development in immune-related applications, such as vaccine design and disease prevention. In this paper, we designed an intelligent system based on a set of combinatorial features including amino acid types and physicochemical characteristics of each residue. We also proposed a novel geometric spiral vector on structural surface for matching similarities of conformational epitopes. The simulation results achieved an average sensitivity of 65.7%, an average specificity of 86.1%, an average positive prediction value of 51.1%, and an average accuracy of 83.5% for a non-redundant dataset containing 53 antigenic proteins. Experimental results show a superior performance of our proposed system compared to currently published computational techniques in the fields of antigen-antibody interaction analysis.

BACKGROUND: Protein structures are flexible and often show conformational changes upon binding to other molecules to exert biological functions. As protein structures correlate with characteristic functions, structure comparison allows classification and prediction of proteins of undefined functions. However, most comparison methods treat proteins as rigid bodies and cannot retrieve similarities of proteins with large conformational changes effectively. RESULTS: In this paper, we propose a novel descriptor, local average distance (LAD), based on either the geodesic distances (GDs) or Euclidean distances (EDs) for pairwise flexible protein structure comparison. The proposed method was compared with 7 structural alignment methods and 7 shape descriptors on two datasets comprising hinge bending motions from the MolMovDB, and the results have shown that our method outperformed all other methods regarding retrieving similar structures in terms of precision-recall curve, retrieval success rate, R-precision, mean average precision and F1-measure. CONCLUSIONS: Both ED- and GD-based LAD descriptors are effective to search deformed structures and overcome the problems of self-connection caused by a large bending motion. We have also demonstrated that the ED-based LAD is more robust than the GD-based descriptor. The proposed algorithm provides an alternative approach for blasting structure database, discovering previously unknown conformational relationships, and reorganizing protein structure classification.

Each gene regulatory network (GRN) involves a set of DNA components and indirectly interacts each other to control gene expression rates. Especially, some sub-networks appear in a GRN holding feedback loop or coupled feedback loop mechanisms, which possess specific relationships and aim to maintain certain functional conditions. Many cases have been reported and experimentally verified the existence of feedback and coupled feedback loops in transcription regulatory networks. However, most feedback loops were not shown directly from the pathways reported by KEGG database. In this study, we integrated all possible cross-regulatory pathways within KEGG maps through examining the feedback relationship among all genes. A total of 427 biological pathway maps were verified individually and an on-line web based system was designed for identifying a specified gene for its possible feedback relationship. In addition, paralogous gene information was also considered for discovering potential coupled feedback loop relationships. There are 50 newly identified feedback loops and 124 coupled feedback loops were found in this study, of which 48 (96%) feedback loops were constructed from different KEGG maps. The built cross-linked pathways could facilitate biologists in understanding how the unveiled dynamic patterns of gene regulation maintain the balance of a specific biological function.

Simple Sequence Repeats (SSRs) are abundant in genome sequences and become popular biomarkers for genetic studies. Several SSRs were proved essential for gene regulation, abnormal repeat patterns of these critical SSRs might cause lethal diseases. The Next Generation Sequencing technologies provided efficient approaches for SSR polymorphism detection. However, inefficient and manually curated processes were unavoidable for identifying SSR markers in previous approaches. An automatic and efficient system for detecting polymorphic SSRs at genomic scales was proposed without manual curated and examining works. The workflow accepted multiple NGS sequencing datasets and started with assembly by de novo or reference mapping approaches. The consensus sequences were then obtained from previously assembled contigs, and calibrated coordinates in each individual contig were aligned according to the selected reference sequences. Next, the mining SSR mechanism was designed to retrieve all potential polymorphic SSRs whenever the circumstances were occurred due to insertion or deletion mechanisms. The 1000 genomes Trio projects were employed as the testing sequence datasets, and the CODIS SSR markers and 9 well known disease-related SSR motifs were verified as the testing targets. The results have shown the proposed method could identify the known polymorphic SSRs as well as novel SSR markers when there was no sequencing or mapping errors within the consensus sequences. The proposed method employed NGS technologies to identify SSR polymorphism and accelerate related researches, which facilitates novel SSR biomarker selection and regulatory elements discovery.

BACKGROUND: A conformational epitope (CE) in an antigentic protein is composed of amino acid residues that are spatially near each other on the antigen's surface but are separated in sequence; CEs bind their complementary paratopes in B-cell receptors and/or antibodies. CE predication is used during vaccine design and in immuno-biological experiments. Here, we develop a novel system, CE-KEG, which predicts CEs based on knowledge-based energy and geometrical neighboring residue contents. The workflow applied grid-based mathematical morphological algorithms to efficiently detect the surface atoms of the antigens. After extracting surface residues, we ranked CE candidate residues first according to their local average energy distributions. Then, the frequencies at which geometrically related neighboring residue combinations in the potential CEs occurred were incorporated into our workflow, and the weighted combinations of the average energies and neighboring residue frequencies were used to assess the sensitivity, accuracy, and efficiency of our prediction workflow. RESULTS: We prepared a database containing 247 antigen structures and a second database containing the 163 non-redundant antigen structures in the first database to test our workflow. Our predictive workflow performed better than did algorithms found in the literature in terms of accuracy and efficiency. For the non-redundant dataset tested, our workflow achieved an average of 47.8% sensitivity, 84.3% specificity, and 80.7% accuracy according to a 10-fold cross-validation mechanism, and the performance was evaluated under providing top three predicted CE candidates for each antigen. CONCLUSIONS: Our method combines an energy profile for surface residues with the frequency that each geometrically related amino acid residue pair occurs to identify possible CEs in antigens. This combination of these features facilitates improved identification for immuno-biological studies and synthetic vaccine design. CE-KEG is available at http://cekeg.cs.ntou.edu.tw.

BACKGROUND: Protein-ligand interactions are key processes in triggering and controlling biological functions within cells. Prediction of protein binding regions on the protein surface assists in understanding the mechanisms and principles of molecular recognition. In silico geometrical shape analysis plays a primary step in analyzing the spatial characteristics of protein binding regions and facilitates applications of bioinformatics in drug discovery and design. Here, we describe the novel software, PLB-SAVE, which uses parallel processing technology and is ideally suited to extract the geometrical construct of solid angles from surface atoms. Representative clusters and corresponding anchors were identified from all surface elements and were assigned according to the ranking of their solid angles. In addition, cavity depth indicators were obtained by proportional transformation of solid angles and cavity volumes were calculated by scanning multiple directional vectors within each selected cavity. Both depth and volume characteristics were combined with various weighting coefficients to rank predicted potential binding regions. RESULTS: Two test datasets from LigASite, each containing 388 bound and unbound structures, were used to predict binding regions using PLB-SAVE and two well-known prediction systems, SiteHound and MetaPocket2.0 (MPK2). PLB-SAVE outperformed the other programs with accuracy rates of 94.3% for unbound proteins and 95.5% for bound proteins via a tenfold cross-validation process. Additionally, because the parallel processing architecture was designed to enhance the computational efficiency, we obtained an average of 160-fold increase in computational time. CONCLUSIONS: In silico binding region prediction is considered the initial stage in structure-based drug design. To improve the efficacy of biological experiments for drug development, we developed PLB-SAVE, which uses only geometrical features of proteins and achieves a good overall performance for protein-ligand binding region prediction. Based on the same approach and rationale, this method can also be applied to predict carbohydrate-antibody interactions for further design and development of carbohydrate-based vaccines. PLB-SAVE is available at http://save.cs.ntou.edu.tw.

More than 5% of adults suffer from different types of kidney disease, and millions of people die prematurely from cardiovascular diseases associated with chronic kidney disease (CKD) in each year. The best way to reduce death caused by kidney disease is early prophylaxis and treatment, and which could be achieved through accurate and reliable diagnoses at the early stage. Among various diagnostic methods, ultrasonographic diagnosis is a low-cost, convenient, non-invasive, and timeliness method. Most importantly, this type inspection would not cause extra burden for patients who suffer kidney diseases. This paper presents a computer-aided diagnosis tool based on analyzing ultrasonography images, and the developed system could detect and classify different stages of CKD. The image processing techniques focus on detecting the atrophy of kidney and the proportion of fibrosis conditions within kidney tissues. The system includes image in painting, noise filtering, contour detection, local contrast enhancement, tissue clustering, and quantitative indicator measuring for distinguishing various stages of CKD. This study has collected thousands of ultrasonic images from patients with kidney diseases, and the selected representative CKD images were applied to be pre-analyzed and trained for comparison. The calculated transition locations as reference indicators could provide physicians an auxiliary and objective computer-aid diagnosis tool for CKD identification and classification.

Human ribonuclease A (hRNaseA) super family members have similar biological functions such as catalytic activities against specific RNA substrates. However, these enzymes with high sequence similarity may exhibit divergent physiological functions other than RNase activity, for example, angio genesis and innate immunity. In our investigation, a novel heparin-binding motif (HBM), RWRCK, identified from hRNase3 contributed to specific protein-heparin/heparan sulfate (HS) interaction. Based on this core HBM sequence, a 10-amino acid heparin binding peptide (HBPRNase3), NYRWRCKNQN, has been designed and characterized. Employing Clustal W2 and Uniprot Blastn program, such HBP pattern is found to be conserved in human and higher primates. Multiple sequence alignment of 13 members of human RNase A family reveals that HBP regions in hRNase2 and hRNase8 share 80% and 50% sequence identity to HBPRNase3, but the corresponding sequences of Gorilla and Pan troglodytes RNase3 are 100% identical, strongly suggesting that HBPRNase3is conserved in higher primates along with species evolution. Interestingly, the putative HBPRNase2, NYQRRCKNQN, shows much lower heparin binding activity than HBPRNase3. In summary, HBPRNase3is not a conserved motif in RNaseA superfamily, but it is a unique motif presenting in higher primates to play a crucial role in molecular interaction to heparin and HS.

Ribonucleases (RNases) constitute a cellular function of exo- and endo-nucleases in all living organisms include vertebrate, bacteria, mold and plant. Human ribonuclease A (hRNaseA) super family consists of thirteen members with high structure similarities. Among which hRNase3 has multiple functions such as RNase, cytotoxicity, and heap ran sulfate (HS) binding activity. Employing Clustal W2 multiple sequence alignment of 13 human RNase A family members revealed that hRNase3 shares 67% identity and 76% similarity with hRNase2, much higher than the others. Besides, both RNase 2 and RNase3 are high conserved among primates. In terms of hRNase3-heparin/heparan sulfate (HS) interaction, three putative heparin binding regions (HBRs),34RWRCK38(HBR1),75RSRFR79(HBR2) and101RPGRR105(HBR3) have been predicted due to presence of positive charge clusters. Furthermore, these regions are also characterized as unique peptide regions within all 13 RNase sequences by REMUS system. Our experimental data derived from protein engineering and binding assays have demonstrated that these HBRs indeed involve in characteristic functions of hRNase3. In summary, through evolution hRNaseA super family has gained novel functions which may be preserved in the unique region or domain to account for additional molecular interactions.

Glaucoma is a serious ocular disease and leads blindness if it couldn't be detected and treated in proper way. The diagnostic criteria for glaucoma include intraocular pressure measurement, optic nerve head evaluation, retinal nerve fiber layer and visual field defect. The observation of optic nerve head, cup to disc ratio and neural rim configuration are important for early detecting glaucoma in clinical practice. However, the broad range of cup to disc ratio is difficult to identify early changes of optic nerve head, and different ethnic groups possess various features in optic nerve head structures. Hence, it is still important to develop various detection techniques to assist clinicians to diagnose glaucoma at early stages. In this study, we developed an automatic detection system which contains two major phases: the first phase performs a series modules of digital fundus retinal image analysis including vessel detection, vessel in painting, cup to disc ratio calculation, and neuro-retinal rim for ISNT rule, the second phase determines the abnormal status of retinal blood vessels from different aspect of view. In this study, the novel idea of integrating image in painting and active contour model techniques successfully assisted the correct identification of cup and disk regions. Several clinical fundus retinal images containing normal and glaucoma images were applied to the proposed system for demonstration.

Tag cloud is an efficient visualization technique to help users rapidly understand the significant components within a tag based dataset. To make practical use of the visualization method, we proposed a cross-comparison mechanism for discovering Simple Sequence Repeat (SSR) biomarkers by annotating retrieved tag entries in various colors and font sizes to differentiate conservative status of two species groups. The identified SSR tag entries can represent the conservative and exclusive features among various species with respect to a specified functional gene set. Grouped SSR features of the same kind species can be classified and displayed simultaneously in order to compare evolutionary relationships. In addition, we applied different parameter settings for SSR discovery algorithms to generate multiscale tag clouds which provide different levels of significance of discovered repeat segments. In this study, several significant SSR patterns within eight skeletal development genes from 10 different species were illustrated and classified as notable features for the previously defined two groups of mammalian and fishery species. The proposed multiscale SSR tag clouds were demonstrated to be able to facilitate biologists on searching notable SSR biomarkers or putative gene regulatory elements in this study.

An epitope activates B cells to amplify and induce antibodies which can neutralize the foreign molecules, particles and pathogens. It also plays a crucial role in developing synthetic peptides for vaccination. Identification of epitopes using biological screening approaches is time consuming and high cost. Therefore, bioinformatics approaches are developed to enhance the speed of identifying the epitopes and conserve time. Herein, a combinatorial methodology based on physico-chemical properties and SVM (Support Vector Machine) techniques was proposed to address the aim of this study. Datasets of epitope and non epitope segments with 2, 3 and 4 residues in length were trained and applied as statistical features of SVM. After training, three datasets including one curated and two public ones were employed to evaluate the performance of the proposed system which was also compared with four existing LE predictors, BepiPred, ABCpred, BCPred and FBCPred. Our proposed system has presented better specificity, accuracy, and positive prediction value (PPV) in most testing cases. High specificity and PPV of a linear epitope prediction can lead to an efficient and effective design on biological experiments.

Ribonuclease 3 (RNase3) is a member of the human RNaseA superfamily. It is led by a 27-residue signal peptide and secreted into body fluid to eliminate the infection of microbial pathogens in innate immune systems. The signal peptide of RNase3 (RNase3sp) not only directs protein secretion via endoplasmic reticulum (ER) and Golgi apparatus, but also its N-terminal 17 amino acid residues (RNase3sp1-17) induces the overexpression of epidermal growth factor receptor (EGFR). In this study, phage display technology was employed to screen the phage peptides which interacted with RNase3sp1-17. The peptide sequences were efficiently analyzed by proposed similarity search systems to retrieve potential candidates which possessed consensus residues conserved with the phage peptides. The retrieved potential candidates were subsequently input into the STRING database for discovering the interaction relationship among the candidates and EGFR. Finally, four potential RNase3sp1-17 interacting candidates, Notch1, GATA2, CBL, and RPX were selected, and the interaction network centered by Notch1 was generated. According to the interaction network and our 3 hypotheses, we attempt to solve the signaling pathway of EGFR overexpression triggered by RNase3sp1-17 by molecular biological experiments.

Repetitive substructures within a protein play an important role in understanding protein folding and stability, biological function, and genome evolution. About 25% of all proteins contain repeat structures for eukaryote species and most of them do not have the resolved structural information yet. Therefore, this study aimed to design a comprehensive system for identifying internal repeats either from a protein sequence or structural information. In this study, we have curated a set of internal repeat units as a benchmark dataset for performing both sequence and structural alignment with respect to the query sequence or structure. Except for the traditional BLAST algorithms on amino acid sequence or the optimal structural superposition approaches on structures, a novel method employing the predicted secondary structure element information for internal repeat identification was proposed. Sequences were firstly transformed into Length Encoded Secondary Structure (LESS) profiles and followed by autocorrelation analyses. From the primary experimental results, the developed Internal Repeat Identification System (IRIS) can successfully identify internal repeats from those known protein structures, and the web system is freely available at http://iris.cs.ntou.edu.tw/.

A novel approach based on solid angles and CUDA parallel technologies for detection of protein docking regions is proposed in this paper. A key feature of a solid angle reveals the geometrical characteristics of protein surface structures and provides an efficient criterion for identifying the anchor amino acids on surface regions. These solid angles as well as several corresponding physical properties facilitate rapid evaluation on matching shape complementary from two interactive proteins. According to the performance of the proposed algorithms on enzyme-inhibitor protein complexes, the retrieved and clustered potential docking regions and the counterpart of centralized anchor amino acids can be effectively determined. Most of identified candidate regions on protein surfaces reveal either with concave or convex characteristics, and the results conform to the regular conditions of protein docking phenomena with respect to enzyme-inhibitor complexes. More importantly, the main goal of this paper not only identifies possible protein docking regions by evaluating solid angle characteristics, but also provides a feasible and efficient way to identify possible binding regions by employing the CUDA parallel computing architecture. The system evaluation has shown that the proposed algorithms achieved an accuracy rate of 62.85% for identifying possible docking regions on two interacted proteins and an average of one-half saving in running time requirements by incorporating CUDA parallel algorithms.

MOTIVATION: Carbohydrate-binding modules (CBMs) share similar secondary and tertiary topology, but their primary sequence identity is low. Computational identification of ligand-binding residues allows biologists to better understand the protein-carbohydrate binding mechanism. In general, functional characterization can be alternatively solved by alignment-based manners. As alignment accuracy based on conventional methods is often sensitive to sequence identity, low sequence identity among query sequences makes it difficult to precisely locate small portions of relevant features. Therefore, we propose a feature-incorporated alignment (FIA) to flexibly align conserved signatures in CBMs. Then, an FIA-based target-template prediction model was further implemented to identify functional ligand-binding residues. RESULTS: Arabidopsis thaliana CBM45 and CBM53 were used to validate the FIA-based prediction model. The predicted ligand-binding residues residing on the surface in the hypothetical structures were verified to be ligand-binding residues. In the absence of 3D structural information, FIA demonstrated significant improvement in the estimation of sequence similarity and identity for a total of 808 sequences from 11 different CBM families as compared with six leading tools by Friedman rank test.

Conformational Epitopes (CEs) are composed of discontinuous segments folding with polypeptide chains, which play an important role in vaccine designs and immuno-biological experiments. It has been estimated that more than 90% of B-cell epitopes are in a conformational structure, and therefore prediction of CEs becomes an important task for practical applications. The proposed system is the first tool that adopts three-dimensional mathematical morphology to define protein structural surface regions. Based on the surface information, spatial distance relationship and antigenic characteristics of amino acids are considered as key features for CEs candidate identification. In this paper, thirty three antigen-antibody complexes with verified CEs were adopted to evaluate prediction accuracy, and the performance was compared with PEPOP. To observe the influence of chemical property scales and surface ratios, a complete evaluation by various parameter settings were performed as well. It was discovered that the sensitivity was improved by selecting hydrophilicity rather than hydrophobicity features, while the specificity favored using hydrophobicity instead of hydrophilicity feature in general. The prediction results of the proposed systemachieves an overall accuracy of 75% and its performance is comparable to existing tools.

Simple sequence repeats (SSRs) play as important genetic markers in genome mapping projects and polymorphic research. In this study, we have designed a database which facilitates the searching and position verification of SSRs and provides comparative genomics information among various species. However, performing in silico analysis of biological data sometimes attempts to result in high false positive rates. In order to promote the specificity of discovered SSRs from the proposed system, we take advantage of evolutionarily conserved segments among sequences from various species. Users are able to choose specific species as targets to filter out SSRs which are not located in conserved regions. Screening processes narrow down candidate SSRs and improve the performance of specificity. In this database, there are eleven representative species collected for comparative genomics analysis. Taking the comparison between zebrafish and medaka fish as an example, 60,363 SSRs from zebrafish genome are found in conserved regions in which 6.20% SSRs are located in protein-coding regions, 0.20% in 5psilaUTR, 0.87% in 3psilaUTR, 51.26% in intron, and 41.45% in intergenic region. Each SSR is precisely allocated and annotated in this database for further applications.

BACKGROUND: Members of a protein family often have highly conserved sequences; most of these sequences carry identical biological functions and possess similar three-dimensional (3-D) structures. However, enzymes with high sequence identity may acquire differential functions other than the common catalytic ability. It is probable that each of their variable regions consists of a unique peptide motif (UPM), which selectively interacts with other cellular proteins, rendering additional biological activities. The ability to identify and localize such UPMs is paramount in recognizing the characteristic role of each member of a protein family. RESULTS: We have developed a reinforced merging algorithm (RMA) with which non-gapped UPMs were identified in a variety of query protein sequences including members of human ribonuclease A (RNaseA), epidermal growth factor receptor (EGFR), matrix metalloproteinase (MMP), and Sma-and-Mad related protein families (Smad). The UPMs generally occupy specific positions in the resolved 3-D structures, especially the loop regions on the structural surfaces. These motifs coincide with the recognition sites for antibodies, as the epitopes of four monoclonal antibodies and two polyclonal antibodies were shown to overlap with the UPMs. Most of the UPMs were found to correlate well with the potential antigenic regions predicted by PROTEAN. Furthermore, an accuracy of 70% can be achieved in terms of mapping a UPM to an epitope. CONCLUSION: Our study provides a bioinformatic approach for searching and predicting potential epitopes and interacting motifs that distinguish different members of a protein family.

The aim of the designed system is mainly to develop the environment for grouping and interactive learning mechanism performing on e-schoolbag, which preserves traditional academic mechanisms nowadays in Taiwan and enhances the teaching and learning activities among teachers and students in a digital manner. The system is developed based on the traditional teaching and learning models in order to smoothen the transition from paper-based materials into digital document format in each educational activity, and its foundation is the conviction that students learn more by doing and discussing than by watching and listening. The essence of the developed system lies in increased interaction at all levels for mathematical learning programs, from peer-to-peer discussions to one-on-one communications among students and teachers. With the assistance of the interactive learning system in the e-learning era, we expect that students are able to develop positive, optimistic, creative, and cooperative abilities beyond the confines of traditional learning mechanisms.

Automated Chinese official document processing techniques and the public-key based cryptographic methodology are used to improve the efficiency of the existing Chinese executive information systems. They provide executive officials more room to achieve the best decision-making. One of the best solutions in document automation and message transmission is to combine low-cost computing power with stable communication capability. In this paper a computer-based architecture for intelligent Chinese official document processing is proposed. The automation of form analysis, document processing, data filing, security control, and information retrieval are discussed in detail. A digital multisignature technology is employed in this system to meet the basic requirements of data security and trust handling. By using such a technology, users will have confidence in utilizing information transmission systems without security suspicion.

A new algorithm for minimizing a morphological skeleton entitled boundary-constrained skeleton minimization (BCSM), as well as a new algorithm for reconstructing an original image from its minimized skeletal structure termed boundary-constrained skeleton reconstruction (BCSR), are proposed. The new algorithms are shown to reduce data storage requirements from (N+1) binary images represented as separate skeleton subsets with their corresponding indices, to 2 binary images composed of a binary morphological skeleton and its corresponding morphological boundary structure. In addition to a reduction in memory storage, BCSM and BCSR result in substantial savings in computational complexity. The proposed algorithms are evaluated in the context of image analysis and coding, and their performance is compared to previous algorithms proposed by Serra (1982, 1988) and by Maragos and Schafer (1986). Sample evaluations indicate a greater than 22-fold savings in computational requirements and 11-fold reduction in memory requirements.>

The authors introduce a lossless data compression technique based on the morphological skeleton representation. Mathematical morphology is a methodology for image analysis which provides a means for describing the geometrical structure of an image quantitatively. A morphological skeleton representation is a useful means of illustrating the geometrical properties of an image (such as shape, size, and orientations). Since it is capable of extracting the minimum underlying geometry of an image, it can also reduce the entropy of an image (data compression). In the present work, the compression ratio is calculated for the evaluation of a new data compression technique using a sample radiograph. This example illustrates that a compression ratio of 1.72 can be achieved through the use of morphological skeleton representation. Two new algorithms, boundary-constrained skeleton minimization and boundary-constrained skeleton reconstruction, are also presented for improving the performance of the morphological-based coding scheme.>

A new algorithm for exact reconstruction of an original image from its morphological skeleton and edge structure is presented. The new technique results in improved computational performance and reduced memory storage requirements. The algorithm is also shown to have better performance characteristics than two previous algorithms in terms of data storage, data transmission rate, and computational complexity. If the size of the maximal element of a skeleton point is larger than the size of a unit structuring element, the algorithm will result in at least a 25% computational savings. The computational savings increases nonlinearly as the size of the maximal element increases.>