Accurate immune cell composition profiling is crucial for understanding immunological dynamics and disease mechanisms. Bulk RNA sequencing (bulk RNA-seq) is widely employed due to its cost-effectiveness and scalability; however, it lacks the resolution to identify cell-specific gene expression. To address this limitation, we propose a self-attention enhanced deep learning model designed for precise immune cell deconvolution from bulk RNA-seq data. We systematically annotated immune cell types from four single-cell RNA-seq (scRNA-seq) peripheral blood mononuclear cell (PBMC) datasets and validated these annotations against established automated identification tools (SingleR, Seurat, scPred, ScType). Leveraging these annotations, we generated realistic pseudo-bulk RNA-seq training samples using Dirichlet-distribution-based composition sampling, significantly enhancing the model’s performance, particularly for rare cell populations. Comparative evaluations demonstrated that our self-attention enhanced deep learning model consistently outperformed existing approaches, including CIBERSORTx and Scaden, achieving lower prediction errors and higher correlations on benchmark PBMC datasets. Integrating multi-head self-attention allowed the model to dynamically capture intricate dependencies among gene expression features, substantially improving deconvolution accuracy for specific cell subsets. While demonstrating robust performance on PBMC datasets, we acknowledge that broader validation is essential due to potential limitations in generalizability across different tissue types and conditions. Our study highlights the potential of self-attention mechanisms and realistic training data generation strategies to enhance computational deconvolution techniques, providing valuable tools for clinical diagnostics and translational immunology research.

Kinases play a critical role in regulating fundamental cellular processes, including metabolism, signal transduction, and cell growth, primarily through phosphorylation. The dysregulation of kinase activity is implicated in various diseases, highlighting the urgent need for robust and interpretable methodologies to profile this activity. Current approaches frequently depend on overly complex or limited datasets, lack generalizability, or fail to provide meaningful biological insights into the mechanisms governing kinase activity. To address these challenges, we developed an explainable deep learning framework that leverages mass spectrometry-based phosphoproteomics data to profile kinase activity effectively. Our study systematically evaluated deep neural networks (DNNs) and convolutional neural networks (CNNs), incorporating a diverse set of feature inputs, including phosphorylation sites and kinase-substrate relationships. A notable finding was that a three-layer CNN, optimized through rigorous feature selection techniques, demonstrated superior performance, achieving substantial improvements in prediction accuracy and stability when compared to established methods such as kinase-substrate enrichment analysis (KSEA) and the kinase activity ranking pipeline (KARP). We integrated Shapley additive explanations (SHAP) values to enhance interpretability, illuminating biologically significant phosphorylation sites. For example, PAK2-related phosphorylation sites associated with the progression of colon adenocarcinoma and CAMK2D sites integral to adrenergic signaling were identified, thereby effectively linking computational predictions to established molecular pathways. This research illustrates the potential of explainable artificial intelligence in advancing kinase activity profiling by providing accurate and interpretable predictions. Our framework is valuable for elucidating disease mechanisms and identifying therapeutic targets, facilitating broader applications in precision medicine.

Rapid identification of antibiotic-resistant infections is crucial, as antimicrobial resistance is a global health crisis. Yet, conventional antibiotic susceptibility tests (AST) often require days to yield results. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has emerged as a rapid, cost-effective tool for bacterial identification and shows promise for resistance profiling by detecting spectral biomarkers. In this study, we harness MALDI-TOF MS with machine learning and deep learning to predict ciprofloxacin resistance across four Gram-negative bacteria, Escherichia coli, Klebsiella pneumoniae, Acinetobacter baumannii, and Acinetobacter nosocomialis, using a cross-species "basket-wise" approach. We extracted features from mass spectra using kernel density estimation-based peak detection and m/z binning, then trained a random forest (RF) classifier and a convolutional neural network (CNN) to distinguish ciprofloxacin-resistant and susceptible isolates. To interpret the models, we employed dual feature importance analyses: gradient-weighted class activation mapping (Grad-CAM) for the CNN to highlight critical m/z regions and an ensemble RF-based method to identify significant peak features. The CNN achieved higher overall accuracy than the RF, especially in three of four species, while the ensemble RF approach identified interpretable sets of around 20 important m/z peaks per organism. Several informative peaks overlapped between species, indicating some common resistance-associated spectral signatures. However, no single universal marker was found across all species. These findings demonstrate an AI-enhanced MALDI-TOF MS framework for rapid AMR detection, yielding accurate predictions and interpretable spectral markers. The approach highlights clinical potential to guide effective therapy and bolster antimicrobial stewardship, particularly for underrepresented pathogens such as A. nosocomialis.

The emergence of multidrug-resistant bacteria is a critical global crisis that poses a serious threat to public health, particularly with the rise of multidrug-resistant Staphylococcus aureus. Accurate assessment of drug resistance is essential for appropriate treatment and prevention of transmission of these deadly pathogens. Early detection of drug resistance in patients is critical for providing timely treatment and reducing the spread of multidrug-resistant bacteria. This study aims to develop a novel risk assessment framework for S. aureus that can accurately determine the resistance to multiple antibiotics. The comprehensive 7-year study involved ˃20 000 isolates with susceptibility testing profiles of six antibiotics. By incorporating mass spectrometry and machine learning, the study was able to predict the susceptibility to four different antibiotics with high accuracy. To validate the accuracy of our models, we externally tested on an independent cohort and achieved impressive results with an area under the receiver operating characteristic curve of 0. 94, 0.90, 0.86 and 0.91, and an area under the precision-recall curve of 0.93, 0.87, 0.87 and 0.81, respectively, for oxacillin, clindamycin, erythromycin and trimethoprim-sulfamethoxazole. In addition, the framework evaluated the level of multidrug resistance of the isolates by using the predicted drug resistance probabilities, interpreting them in the context of a multidrug resistance risk score and analyzing the performance contribution of different sample groups. The results of this study provide an efficient method for early antibiotic decision-making and a better understanding of the multidrug resistance risk of S. aureus.

Recent studies have demonstrated that the matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) could be used to detect superbugs, such as methicillin-resistant Staphylococcus aureus (MRSA). Due to an increasingly clinical need to classify between MRSA and methicillin-sensitive Staphylococcus aureus (MSSA) efficiently and effectively, we were motivated to develop a systematic pipeline based on a large-scale dataset of MS spectra. However, the shifting problem of peaks in MS spectra induced a low effectiveness in the classification between MRSA and MSSA isolates. Unlike previous works emphasizing on specific peaks, this study employs a binning method to cluster MS shifting ions into several representative peaks. A variety of bin sizes were evaluated to coalesce drifted or shifted MS peaks to a well-defined structured data. Then, various machine learning methods were performed to carry out the classification between MRSA and MSSA samples. Totally 4858 MS spectra of unique S. aureus isolates, including 2500 MRSA and 2358 MSSA instances, were collected by Chang Gung Memorial Hospitals, at Linkou and Kaohsiung branches, Taiwan. Based on the evaluation of Pearson correlation coefficients and the strategy of forward feature selection, a total of 200 peaks (with the bin size of 10 Da) were identified as the marker attributes for the construction of predictive models. These selected peaks, such as bins 2410-2419, 2450-2459 and 6590-6599 Da, have indicated remarkable differences between MRSA and MSSA, which were effective in the prediction of MRSA. The independent testing has revealed that the random forest model can provide a promising prediction with the area under the receiver operating characteristic curve (AUC) at 0.8450. When comparing to previous works conducted with hundreds of MS spectra, the proposed scheme demonstrates that incorporating machine learning method with a large-scale dataset of clinical MS spectra may be a feasible means for clinical physicians on the administration of correct antibiotics in shorter turn-around-time, which could reduce mortality, avoid drug resistance and shorten length of stay in hospital in the future.

BACKGROUND: A mass spectrometry-based assessment of methicillin resistance in Staphylococcus aureus would have huge potential in addressing fast and effective prediction of antibiotic resistance. Since delays in the traditional antibiotic susceptibility testing, methicillin-resistant S. aureus remains a serious threat to human health. RESULTS: Here, linking a 7 years of longitudinal study from two cohorts in the Taiwan area of over 20 000 individually resolved methicillin susceptibility testing results, we identify associations of methicillin resistance with the demographics and mass spectrometry data. When combined together, these connections allow for machine-learning-based predictions of methicillin resistance, with an area under the receiver operating characteristic curve of >0.85 in both the discovery [95% confidence interval (CI) 0.88-0.90] and replication (95% CI 0.84-0.86) populations. CONCLUSIONS: Our predictive model facilitates early detection for methicillin resistance of patients with S. aureus infection. The large-scale antibiotic resistance study has unbiasedly highlighted putative candidates that could improve trials of treatment efficiency and inform on prescriptions.

In recent years, antimicrobial peptides (AMPs) have become an emerging area of focus when developing therapeutics hot spot residues of proteins are dominant against infections. Importantly, AMPs are produced by virtually all known living organisms and are able to target a wide range of pathogenic microorganisms, including viruses, parasites, bacteria and fungi. Although several studies have proposed different machine learning methods to predict peptides as being AMPs, most do not consider the diversity of AMP activities. On this basis, we specifically investigated the sequence features of AMPs with a range of functional activities, including anti-parasitic, anti-viral, anti-cancer and anti-fungal activities and those that target mammals, Gram-positive and Gram-negative bacteria. A new scheme is proposed to systematically characterize and identify AMPs and their functional activities. The 1st stage of the proposed approach is to identify the AMPs, while the 2nd involves further characterization of their functional activities. Sequential forward selection was employed to extract potentially informative features that are possibly associated with the functional activities of the AMPs. These features include hydrophobicity, the normalized van der Waals volume, polarity, charge and solvent accessibility-all of which are essential attributes in classifying between AMPs and non-AMPs. The results revealed the 1st stage AMP classifier was able to achieve an area under the receiver operating characteristic curve (AUC) value of 0.9894. During the 2nd stage, we found pseudo amino acid composition to be an informative attribute when differentiating between AMPs in terms of their functional activities. The independent testing results demonstrated that the AUCs of the multi-class models were 0.7773, 0.9404, 0.8231, 0.8578, 0.8648, 0.8745 and 0.8672 for anti-parasitic, anti-viral, anti-cancer, anti-fungal AMPs and those that target mammals, Gram-positive and Gram-negative bacteria, respectively. The proposed scheme helps facilitate biological experiments related to the functional analysis of AMPs. Additionally, it was implemented as a user-friendly web server (AMPfun, http://fdblab.csie.ncu.edu.tw/AMPfun/index.html) that allows individuals to explore the antimicrobial functions of peptides of interest.

BACKGROUND: Group B streptococcus (GBS) is an important pathogen that is responsible for invasive infections, including sepsis and meningitis. GBS serotyping is an essential means for the investigation of possible infection outbreaks and can identify possible sources of infection. Although it is possible to determine GBS serotypes by either immuno-serotyping or geno-serotyping, both traditional methods are time-consuming and labor-intensive. In recent years, the matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) has been reported as an effective tool for the determination of GBS serotypes in a more rapid and accurate manner. Thus, this work aims to investigate GBS serotypes by incorporating machine learning techniques with MALDI-TOF MS to carry out the identification. RESULTS: In this study, a total of 787 GBS isolates, obtained from three research and teaching hospitals, were analyzed by MALDI-TOF MS, and the serotype of the GBS was determined by a geno-serotyping experiment. The peaks of mass-to-charge ratios were regarded as the attributes to characterize the various serotypes of GBS. Machine learning algorithms, such as support vector machine (SVM) and random forest (RF), were then used to construct predictive models for the five different serotypes (Types Ia, Ib, III, V, and VI). After optimization of feature selection and model generation based on training datasets, the accuracies of the selected models attained 54.9-87.1% for various serotypes based on independent testing data. Specifically, for the major serotypes, namely type III and type VI, the accuracies were 73.9 and 70.4%, respectively. CONCLUSION: The proposed models have been adopted to implement a web-based tool (GBSTyper), which is now freely accessible at http://csb.cse.yzu.edu.tw/GBSTyper/, for providing efficient and effective detection of GBS serotypes based on a MALDI-TOF MS spectrum. Overall, this work has demonstrated that the combination of MALDI-TOF MS and machine intelligence could provide a practical means of clinical pathogen testing.

BACKGROUND: Functional RNA molecules participate in numerous biological processes, ranging from gene regulation to protein synthesis. Analysis of functional RNA motifs and elements in RNA sequences can obtain useful information for deciphering RNA regulatory mechanisms. Our previous work, RegRNA, is widely used in the identification of regulatory motifs, and this work extends it by incorporating more comprehensive and updated data sources and analytical approaches into a new platform. METHODS AND RESULTS: An integrated web-based system, RegRNA 2.0, has been developed for comprehensively identifying the functional RNA motifs and sites in an input RNA sequence. Numerous data sources and analytical approaches are integrated, and several types of functional RNA motifs and sites can be identified by RegRNA 2.0: (i) splicing donor/acceptor sites; (ii) splicing regulatory motifs; (iii) polyadenylation sites; (iv) ribosome binding sites; (v) rho-independent terminator; (vi) motifs in mRNA 5'-untranslated region (5'UTR) and 3'UTR; (vii) AU-rich elements; (viii) C-to-U editing sites; (ix) riboswitches; (x) RNA cis-regulatory elements; (xi) transcriptional regulatory motifs; (xii) user-defined motifs; (xiii) similar functional RNA sequences; (xiv) microRNA target sites; (xv) non-coding RNA hybridization sites; (xvi) long stems; (xvii) open reading frames; (xviii) related information of an RNA sequence. User can submit an RNA sequence and obtain the predictive results through RegRNA 2.0 web page. CONCLUSIONS: RegRNA 2.0 is an easy to use web server for identifying regulatory RNA motifs and functional sites. Through its integrated user-friendly interface, user is capable of using various analytical approaches and observing results with graphical visualization conveniently. RegRNA 2.0 is now available at http://regrna2.mbc.nctu.edu.tw.

mRNA polyadenylation Is an essential mechanism in human genes and is direct linked to the termination of transcription. Alternative polyadenylation changes the length of the mature mRNA's 3'UTR. Since 3'UTRs have been shown to contain regulatory elements that control mRNA functioning, alternative polyadenylation plays an important role in controlling the expression of human genes. Prediction of polyadenylation sites can help with the identification of genes and aid our understanding of the mechanisms of alternative polyadenylation. In this study, we constructed a system for mRNA polyadenylation site prediction in human genes using SVM and based on an analysis of the sequence alignment between pair-end diTags (PET) and genome sequences. The PET sequences were mapped to the reference genome more accurate compared to earlier methods. We also analyzed single-site type and multiple-site type sequences PET sequence datasets and found that the frequencies of each nucleotide were different when the single-site type and multiple-site type PET sequences were compared.

Current pathway presentation method to the biologist is static graph. The analysis of differentially expressed genes using microarray gene expression data can help to find factors that affect diseases. However, the differentially expressed genes that are identified may be too large in number and it's difficult for biologist to pinpoint the correlations between genes and crucial points on pathway interactively. In this study, we propose a method that attempts to avoid this problem and allows the discovery of greater biological meaning than the traditional method. We select a gene pair set of interacting genes in a biological pathway and investigate the correlation in expression between the gene pairs under different condition (such as relapsed and non-relapsed breast cancer) using microarray gene expression data. We tested the approach using breast cancer relapsed and non-relapsed datasets in order to demonstrate that our method is both useful and reliable; very stable results were obtained when the same microarray platform was used. We finally use an interface to display correlations within a biological pathway.

Riboswitches are cis-acting genetic regulatory elements within a specific mRNA, and can regulate both transcription and translation by interacting with their corresponding metabolites. Recently, more and more riboswitches were identified and investigated about their roles in regulatory functions in different species. Both of the sequence contexts and structural conformations are important characteristics of riboswitches. None of previous developed tools, such Covariance Models (CMs), Riboswitch finder, and RibEx, provides a web server for efficiently searching homologous instances to known riboswitches and considers two crucial characteristics of each riboswitch, such as structural conformations and sequence contexts of functional regions. Therefore, we developed a systematic method to identify twelve kinds of riboswitches. The method is implemented and provided as a web server, RiboSW, to efficiently and conveniently identify riboswitches within messenger RNA sequences. RiboSW is now available on the web at http://bioinfo.csie.ncu.edu.tw/RiboSW/. The predictive accuracy of the proposed method is comparable with other previous tools. The efficiency of the proposed method for identifying riboswitches was improved in order to achieve a reasonable computational time required for the prediction. That makes it possible to have an accurate and convenient web server for biologists to obtain their analyzing results in a given mRNA sequence.

During gene expression, transcription factors are unable to bind to a transcription binding site (TFBS) involved in regulation if DNA methylation has occurred at the TFBS. Methyl-CpG-binding proteins may also occupy the TFBS and prevent the functioning of a transcription factor. Thus, the methylation status of CpG sites is an important issue when trying to understand gene regulation and shows strong correlation with the TFBS involved. In addition, CpG islands would seem to undergo cell-specific and tissue-specific me-thylation. Such differential methylation is presented at numerous genetic loci that are essential for development. Current DNA methylation site prediction tools need to be improved so that they include TFBS features and have greater accuracy in terms of the DNA region that is involved in methylation. We developed models that compare the differences across these regions and tissues. The TFBSs, DNA properties and DNA distribution were used as features for this classification. From the results, we found some TFBSs that were able to discriminate whether a sequence was methylated or not. The sensitivity, specificity and accuracy estimated using 10-fold cross validation were 90.8%, 80.54%, and 86.07%, respectively. Thus, for these four regions and twelve tissues, the performance levels (ACC) were all greater than 80%. We propose that the differential features or methylations vary between the different regions because the features common to each DNA region made up only 50% of the top 70 features. An online predictor based on EpiMeP is available at http://140.115.51.41/EpiMeP/. Supplementary file is available at http://140.115.51.41/EpiMeP/supplementary.doc.

Recently, the H5N1 virus has had an increasingly important impact on human life. This is because more and more people are becoming infected with this virus, and the possibility of a serious pandemic with human to human transmission is looming. This might occur if the genome of this influenza virus mutates either by antigenic drift or by antigenic shift, especially if there is a mutation of the hemagglutinin (HA) glycoprotein. The HA is the surface glycoprotein, and it binds to sialic acid of the host cell surface receptor. Thus, the combination of HA and sialic acid are central to whether influenza virus infects humans. In this study, we selected 497 HA protein sequences from the National Center for Biotechnology Information (NCBI) Influenza Resource database, and used a decision tree method to identify discriminative amino acids in the HA protein sequences that may possibly influence the binding of HA to sialic acid. Four such amino acid positions at 54, 55, 241, and 281 were identified and these may play an important role in infection by H5N1 influenza virus.

In this study, we proposed a new methodology to combine the paper slides, computer screen and a digital pen in order to improve teachers teaching and students learning in the traditional classroom. In students' learning, we design the new course slides which have kept the habit students had a class in the past, and integrate the advantage of real-time online support. In instructors' teaching, instructors can immediately get students' learning status, and change their teaching strategies by recommendations of the system.

Relational robot is an intelligent conversational agent designed to be a learning secretary for student to notify student of using our learning Web site on time. In this paper we discuss some modules of relational robot to let it play a learning secretary role; it contains interaction module (verbal communication unit, nonverbal communication unit and process unit), internet interface, and communication interface with structure dialogue

Viral infection poses a major problem for public health, horticulture, and animal husbandry, possibly causing severe health crises and economic losses. Viral infections can be identified by the specific detection of viral sequences in many ways. The microarray approach not only tolerates sequence variations of newly evolved virus strains, but can also simultaneously diagnose many viral sequences. Many chips have so far been designed for clinical use. Most are designed for special purposes, such as typing enterovirus infection, and compare fewer than 30 different viral sequences. None considers primer design, increasing the likelihood of cross hybridization to similar sequences from other viruses. To prevent this possibility, this work establishes a platform and database that provides users with specific probes of all known viral genome sequences to facilitate the design of diagnostic chips. This work develops a system for designing probes online. A user can select any number of different viruses and set the experimental conditions such as melting temperature and length of probe. The system then returns the optimal sequences from the database. We have also developed a heuristic algorithm to calculate the probe correctness and show the correctness of the algorithm. (The system that supports probe design for identifying viruses has been published on our web page http://bioinfo.csie.ncu.edu.tw/.)

Identification of transcriptional regulatory sites plays an important role in the investigation of gene regulation. For this propose, we designed and implemented a data warehouse to integrate multiple heterogeneous biological data sources with data types such as text-file, XML, image, MySQL database model, and Oracle database model. The utility of the biological data warehouse in predicting transcriptional regulatory sites of coregulated genes was explored using a synexpression group derived from a microarray study. Both of the binding sites of known transcription factors and predicted over-represented (OR) oligonucleotides were demonstrated for the gene group. The potential biological roles of both known nucleotides and one OR nucleotide were demonstrated using bioassays. Therefore, the results from the wet-lab experiments reinforce the power and utility of the data warehouse as an approach to the genome-wide search for important transcription regulatory elements that are the key to many complex biological systems.

Multiplex Polymerase Chain Reaction (PCR) experiments are used for amplifying several segments of the target DNA simultaneously and thereby to conserve template DNA, reduce the experimental time, and minimize the experimental expense. The success of the experiment is dependent on primer design. However, this can be a dreary task as there are many constrains such as melting temperatures, primer length, GC content and complementarity that need to be optimized to obtain a good PCR product. Motivated by the lack of primer design tools for multiplex PCR genotypic assay, we propose a multiplex PCR primer design tool using a genetic algorithm, which is a stochastic approach based on the concept of biological evolution, biological genetics and genetic operations on chromosomes, to find an optimal selection of primer pairs for multiplex PCR experiments. The presented experimental results indicate that the proposed algorithm is capable of finding a series of primer pairs that obeies the design properties in the same tube.

Cervical cancer is common among women all over the world. Although infection with high-risk types of human papillomavirus (HPV) has been identified as the primary cause of cervical cancer, only some of those infected go on to develop cervical cancer. Obviously, the progression from HPV infection to cancer involves other environmental and host factors. Recent population-based twin and family studies have demonstrated the importance of the hereditary component of cervical cancer, associated with genetic susceptibility. Consequently, SNP markers and microsatellites should be considered genetic factors for determining what combinations of genetic factors are involved in precancerous changes to cervical cancer. This study employs a Bayesian network and four different decision tree algorithms, and compares the performance of these learning algorithms. The results of this study raise the possibility of investigations that could identify combinations of genetic factors, such as SNPs and microsatellites, that influence the risk associated with common complex multifactorial diseases, such as cervical cancer. The web site associated with this study is http://dblab8.csie.ncu.edu.tw/FactorAnalysis/.

UNLABELLED: Included in Prokaryotic Growth Temperature database (PGTdb) are a total of 1334 temperature data from 1072 prokaryotic organisms, Bacteria and Archaea: PGTdb integrates microbial growth temperature data from literature survey with their nucleotide/protein sequence and protein structure data from related databases. A direct correlation is observed between the average growth temperature of an organism and the melting temperature of proteins from the organism. Therefore, this database is useful not only for microbiologists to obtain cultivation condition, but also for biochemists and structure biologists to study the correlation between protein sequences/structures and their thermostability. In addition, the taxonomy and ribosomal RNA sequence(s) of an organism are linked through NCBI Taxonomy and the Ribosomal RNA Operon Copy Number Database umdb, respectively. PGTdb is the only integrated database on the Internet to provide the growth temperature data of the prokaryotes and the combined information of their nucleotide/protein sequences, protein structures, taxonomy and phylogeny. AVAILABILITY: http://pgtdb.csie.ncu.edu.tw

Cervical cancer is common among women all over the world. Although infection with high-risk types of human papillomavirus (HPV) has been identified as the primary cause of cervical cancer, only some of those infected go on to develop cervical cancer. Obviously, the progression from HPV infection to cancer involves other environmental and host factors. Recent population-based twin and family studies have demonstrated the importance of the hereditary component of cervical cancer, associated with genetic susceptibility. Consequently, single-nucleotide polymorphism (SNP) markers and microsatellites should be considered genetic factors for determining what combinations of genetic factors are involved in precancerous changes to cervical cancer. This study employs a Bayesian network and four different decision tree algorithms, and compares the performance of these learning algorithms. The results of this study raise the possibility of investigations that could identify combinations of genetic factors, such as SNPs and microsatellites, that influence the risk associated with common complex multifactorial diseases, such as cervical cancer.

Very large-scale gene expression analysis, i.e., UniGene and dbEST, are provided to find those genes with significantly differential expression in specific tissues. The differentially expressed genes in a specific tissue are potentially regulated concurrently by a combination of transcription factors. This study attempts to mine putative binding sites on how combinations of the known regulatory sites homologs and over-represented repetitive elements are distributed in the promoter regions of considered groups of differentially expressed genes. We propose a data mining approach to statistically discover the significantly tissue-specific combinations of known site homologs and over-represented repetitive sequences, which are distributed in the promoter regions of differential gene groups. The association rules mined would facilitate to predict putative regulatory elements and identify genes potentially co-regulated by the putative regulatory elements.

Mobile techniques make information transform more efficiently. Many learning strategies will become more effectively under the supporting of significant and real time information. Three kinds of information awareness mechanisms using mobile devices are proposed to assist students to promote learning performance of students. Mechanisms for improving learning status awareness, schedule awareness, and mentor awareness are developed to recommend students what should learn, what should do and who could be of help for a problem by transferring information through cell phone short message to students. Experiments have been performed and the results show that most students satisfy with system's recommendation and these awareness mechanisms have positive affect to students.

Approximately 43% of the human genome is occupied by repetitive elements. Even more, around 51% of the rice genome is occupied by repetitive elements. The analysis presented here indicates that repetitive elements in complete genomes may have been very important in the evolutionary genomics. In this study, a database, called the Repeat Sequence Database, is first designed and implemented to store complete and comprehensive repetitive sequences. See http://rsdb.csie.ncu.edu.tw for more information. The database contains direct, inverted and palindromic repetitive sequences, and each repetitive sequence has a variable length ranging from seven to many hundred nucleotides. The repetitive sequences in the database are explored using a mathematical algorithm to mine rules on how combinations of individual binding sites are distributed among repetitive sequences in the database. Combinations of transcription factor binding sites in the repetitive sequences are obtained and then data mining techniques are applied to mine association rules from these combinations. The discovered associations are further pruned to remove insignificant associations and obtain a set of associations. The mined association rules facilitate efforts to identify gene classes regulated by similar mechanisms and accurately predict regulatory elements. Experiments are performed on several genomes including C. elegans, human chromosome 22, and yeast.

We discuss the scheduling problems of a sports league and propose a new approach to solve these problems by applying evolution strategy. A schedule in a sports league must satisfy many constraints on timing, such as the number of games played between every pair of teams, the bounds on the number of consecutive home (or away) games for each team, every pair of teams must have played each other in the first half of the season, and so on. In addition to finding a feasible schedule that meets all the timing restrictions, the problem addressed has the additional complexity of having the objective of minimizing travel costs and every team having a balanced number of games at home. We formalize the scheduling problem into an optimization problem and adopt the concept of evolution strategy to solve it. We define the travel cost and distance cost for teams in the sports league by referring to Major League Baseball (MLB) in the United States and focus on the scheduling problem in MLB. Using the new method, it is more efficient at finding better results than previous approaches.

In a mobile environment, a decision maker would usually like to query times, locations, and moving behaviors of certain mobile terminals. For example, a mobile transaction would like to know where is the next cell of its mobile clients and the probabilities that the mobile clients will move to the cells so that it could reserve appropriate channels for them. As a result, tracking the changing times and positions of mobile terminals capable of continuous movement is becoming increasingly necessary. However, current personal communication services (PCS) networks can only offer current maintained location information of non-idle mobile terminals. Pertinent researches predict location based on tangent velocity approaches. Location prediction based on tangent velocity is effective only within a short time interval. In this paper, we propose a model to model the moving behavior of each mobile terminal. From the moving behavior, we estimate and propose several location prediction functions for location query. The experimental results show that our proposed location prediction functions are accurate enough for regular moving mobile terminals.

Web based peer assessment for portfolios has been used as an innovative assessment methods to reuse students portfolio for refining learning. However, without sophisticated support to articulate the assessment contexture about portfolios, students cannot communicate with explicit learning concepts and think reflectively for refinement of their learning. This study attempts to utilize knowledge acquisition and data mining techniques to solve the problems in support contexture awareness. By way of knowledge acquisition and data mining techniques, web peer assessment systems obtain students' personal theories during assessing portfolios. Thereby, teachers and students can go online to fully exchange personal theories, thus allowing them to think reflectively for refinement of learning.

Due to the insufficiency of available bandwidth resources and the continuously growing demand for cellular communication services, the channel assignment problem becomes increasingly important. To trace the optimal assignment, several heuristic strategies have been proposed. So far, most of them focus on the small-scale systems containing no more than 25 cells and they use an anachronistic cost model that does not satisfy the requirements of most existing cellular operators to measure the solution quality. Solving the small-scale channel assignment problems could not be applied into existing large scale cellular networks' practice. This article proposes a decomposition approach to solve the fixed channel assignment problem (FCAP) for large-scale cellular networks through partitioning the whole cellular network into several smaller sub-networks and then a sequential branch-and-bound algorithm is designed to solve the FCAP for them sequentially. The key issue of partition is to minimize the dependences of the sub-networks so that the proposed heuristics for solving smaller problems will suffer fewer constraints in searching better assignments. The experimental results show that the proposed algorithms perform well and we applied our algorithms in finding better assignments for the cellular network of the Taiwan Cellular Cooperation in ChungLi city.

Some structural motifs, like tetra-loops, in ribosomal RNA are known to functionally implicate in virtually every aspect of protein synthesis. Our aim in this study is to discover common structural motifs (CSMs), which are related to specific domains or functions, within the secondary structures of ribosomal RNAs in a data set constructed. After applying data mining techniques to mine the common structural motifs, a machine learning approach is used to find significant discriminating common structural motifs from groups of organisms. By applying to several data sets constructed in this study, it suggests that the CSMs can provide effective information to classify organisms and help biologists understand the functions of ribosomal RNA. From the experiments of the classification of organisms and the construction of phylogenetic trees by CSMs mined, we find our approach is promising.

A flexible ligand docking protocol based on evolutionary algorithms is investigated. The proposed approach integrates decreasing-based mutations and self-adaptive mutations with differential evolution. This approach possesses global and local search strategies to balance the trade-off between exploitation and exploration of the search. The proposed approach is applied to a dihydrofolate reductase enzyme with the anti-cancer drug methotrexate and two analogues of antibacterial drug trimethoprim. Numerical results indicate that the new approach is very robust.

Due to the poverty of available bandwidth resources and increasing demand for cellular communication services, the problem of channel assignment becomes increasingly important. To trace optimal assignment, several algorithms have been proposed to minimize the amount of required channels. However, the total number of available frequencies are given and fixed in many situations. A new cost model is required for assigning channel in the cellular networks with limited bandwidth. We analyze the cost of each assignment in the view of damages from blocking calls and interfered by other frequencies. Furthermore, we formulate a new optimization problem for the fixed channel assignment problem by incorporating the limited bandwidth constraint into its cost model. To minimize the cost function, we adopt genetic approach to propose an evolutionary approach. Experimental results show that the cost function does reflect the quality of different assignments and also show that our algorithm does improve the solution quality significantly.

We propose a new location tracking strategy called behavior-based strategy (BBS) based on each mobile's moving behavior. With the help of data mining technologies the moving behavior of each mobile could be mined from long-term collection of the mobile's moving logs. From the moving behavior of each mobile, we first estimate the time-varying probability of the mobile and then the optimal paging area of each time region is derived. To reduce unnecessary computation, we consider the location tracking and computational cost and then derive a cost model. A heuristics is proposed to minimize the cost model through finding the appropriate moving period checkpoints of each mobile. The experimental results show our strategy outperforms fixed paging area strategy currently used in the GSM system and time-based strategy for highly regular moving mobiles.

A robust evolutionary approach, called the Family Competition Evolutionary Algorithm (FCEA), is described for the synthesis of optical thin-film designs. Based on family competition and adaptive rules, the proposed approach consists of global and local strategies by integrating decreasing mutations and self-adaptive mutations. The method is applied to three different optical coating designs with complex spectral quantities. Numerical results indicate that the proposed approach performs very robustly and is very competitive with other approaches.

This investigation presents evolution strategies to solve quadratic assignment problems. The proposed algorithm applies family competition and clustering to enhance the solution quality. Several problems obtained from QAPLIB (Burkard et al., 1991) were tested and experimental results were compared with other approaches in the literature. Experimental results also show that the proposed algorithm is promising.

In this paper, we present a new evolutionary technique to train three general neural networks. Based on family competition principles and adaptive rules, the proposed approach integrates decreasing-based mutations and self-adaptive mutations to collaborate with each other. Different mutations act as global and local strategies respectively to balance the trade-off between solution quality and convergence speed. Our algorithm is then applied to three different task domains: Boolean functions, regular language recognition, and artificial ant problems. Experimental results indicate that the proposed algorithm is very competitive with comparable evolutionary algorithms. We also discuss the search power of our proposed approach.

A data warehouse stores lots of materialized views to provide efficient decision-support or OLAP queries. The view-selection problem addresses the selection of a fittest set of materialized views under the limitation of storage space forms a challenge in data warehouse research. In this paper, we present genetic algorithms to choose materialized views. We also use experiments to demonstrate the power of our approach.

This investigation presents an adapted genetic algorithm to resolve simple plant location problems. The proposed algorithm applies a clustering technique as mutation guidance and a novel local search method to enhance the solution quality. The proposed algorithm is then applied to the fifteen test problems taken from Beasley's OR-Library (J.E. Beasley, 1990). Empirical results indicate that the error rate of the proposed adapted GA is less than 0.3 percent. In addition, the computational time is bounded by a polynomial function of the problem size.

The paper presents an evolutionary technique to train neural networks in tasks requiring learning behavior. Based on family competition principles and adaptive rules, the proposed approach integrates decreasing-based mutations and self-adaptive mutations. Different mutations act global and local strategies separately to balance the trade-off between solution quality and convergence speed. The algorithm proposed herein is applied to two different task domains: Boolean functions and artificial ant problem. Experimental results indicate that in all tested problems, the proposed algorithm performs better than other canonical evolutionary algorithms, such as genetic algorithms, evolution strategies, and evolutionary programming. Moreover, essential components such as mutation operators and adaptive rules in the proposed algorithm are thoroughly analyzed.

The hierarchical network structure significantly reduces the size and maintenance cost of routing table for huge networks. But in ad hoc networks, no fixed host leads to the challenge of the hierarchical structure, since the topology information needs to be updated dynamically due to membership changes caused by mobility. To construct the hierarchical structure of physical locations, we adopt a cluster infrastructure to partition the network into different groups for physical location maintenance. In order to construct the hierarchical structure of logical locations, all hosts are divided into several domains, each one of them has one corresponding domain location server to record all of the member physical locations (cluster locations). With the hierarchical structure, most necessary routing information can be ignored.

As database and workflow technologies are used to manage business processes, decision-makers of enterprises must query the execution of business processes to understand and refine these processes for expected throughput and quality. Introducing the representation of business processes in the database schema allows the execution histories of business processes to be maintained in the database for supporting queries on how business processes are executed. In this work, we incorporate finite state machines into the entity-relationship (ER) model for representing business processes in the database schema. Analytical systems can be developed in the proposed representation to assist decision-makers in observing the business processes.

This paper explores the use of neural networks to control robots in tasks requiring sequential and learning behavior. We propose a family competition evolutionary algorithm (FCEA) to evolve networks that can integrate these different types of behavior in a smooth and continuous manner. The approach integrates self-adaptive Gaussian mutation, self-adaptive Cauchy mutation, decreasing-based Gaussian mutation, and family competition. In order to illustrate the power of the approach, we apply this approach to two different task domains: the "artificial ant" problem and a sequential behavior problem - an agent learns to play football. From the experimental results, we find our approach performs much better than other evolutionary algorithms in these two tasks. Based on the results from our experiments, it is shown that our approach can evolve neural networks to provide a means of integrating, sequencing and learning within a single control system.

Error-correcting graph isomorphism has been found useful in numerous pattern recognition applications. This paper presents a genetic-based search approach that adopts genetic algorithms as the searching criteria to solve the problem of error-correcting graph isomorphism. By applying genetic algorithms, some local search strategies are amalgamated to improve convergence speed. Besides, a selection operator is proposed to prevent premature convergence. The proposed approach has been implemented to verify its validity. Experimental results reveal the superiority of this new technique than several other well-known algorithms.