Mining the epistatic gene loci for complex diseases is an important research topic in recent years. The existing epistasis detection methods have the shortcomings of high complexity, low efficiency, high false positive rate, inability to deal with large-scale genome-wide data, etc. In this work, we propose a epistasis mining method which using k-tree optimizing Bayesian network (BN) (KtreeBnIp). Firstly, it constructs the k-tree including large-scale of SNP loci and phenotype by sampling the Dandelion code uniformly. Then it decomposes k-tree into different k-cliques by using the graph decomposition algorithm based on neighbor nodes. For the nodes in each k-clique, the integer linear programming optimizing Bayesian network (ILPBN) is used to obtain the sub-network quickly and accurately. Finally, it merges all the sub-networks to obtain the whole network. It repeats the above operations (k-tree generation, k-tree decomposition, network generation) several times and gets the final network according to the frequency of edges, and thus to obtain the epistatic loci affecting phenotype. We compare KtreeBnIp with the current popular epistasis mining algorithms using both simulated and real age-related macular degeneration(AMD) datasets. Experiment results show that KtreeBnIp has better epistasis detection accuracy, higher F1-score, and lower false positive rate. Most importantly, it can be used in large-scale of SNPs for epistasis detection.

Identifying the function of DNA sequences accurately is an essential and challenging task in the genomic field. Until now, deep learning has been widely used in the functional analysis of DNA sequences, including DeepSEA, DanQ, DeepATT and TBiNet. However, these methods have the problems of high computational complexity and not fully considering the distant interactions among chromatin features, thus affecting the prediction accuracy. In this work, we propose a hybrid deep neural network model, called DeepFormer, based on convolutional neural network (CNN) and flow-attention mechanism for DNA sequence function prediction. In DeepFormer, the CNN is used to capture the local features of DNA sequences as well as important motifs. Based on the conservation law of flow network, the flow-attention mechanism can capture more distal interactions among sequence features with linear time complexity. We compare DeepFormer with the above four kinds of classical methods using the commonly used dataset of 919 chromatin features of nearly 4.9 million noncoding DNA sequences. Experimental results show that DeepFormer significantly outperforms four kinds of methods, with an average recall rate at least 7.058% higher than other methods. Furthermore, we confirmed the effectiveness of DeepFormer in capturing functional variation using Alzheimer's disease, pathogenic mutations in alpha-thalassemia and modification in CCCTC-binding factor (CTCF) activity. We further predicted the maize chromatin accessibility of five tissues and validated the generalization of DeepFormer. The average recall rate of DeepFormer exceeds the classical methods by at least 1.54%, demonstrating strong robustness.

MOTIVATION: The question of how to construct gene regulatory networks has long been a focus of biological research. Mutual information can be used to measure nonlinear relationships, and it has been widely used in the construction of gene regulatory networks. However, this method cannot measure indirect regulatory relationships under the influence of multiple genes, which reduces the accuracy of inferring gene regulatory networks. APPROACH: This work proposes a method for constructing gene regulatory networks based on mixed entropy optimizing context-related likelihood mutual information (MEOMI). First, two entropy estimators were combined to calculate the mutual information between genes. Then, distribution optimization was performed using a context-related likelihood algorithm to eliminate some indirect regulatory relationships and obtain the initial gene regulatory network. To obtain the complex interaction between genes and eliminate redundant edges in the network, the initial gene regulatory network was further optimized by calculating the conditional mutual inclusive information (CMI2) between gene pairs under the influence of multiple genes. The network was iteratively updated to reduce the impact of mutual information on the overestimation of the direct regulatory intensity. RESULTS: The experimental results show that the MEOMI method performed better than several other kinds of gene network construction methods on DREAM challenge simulated datasets (DREAM3 and DREAM5), three real Escherichia coli datasets (E.coli SOS pathway network, E.coli SOS DNA repair network and E.coli community network) and two human datasets. AVAILABILITY AND IMPLEMENTATION: Source code and dataset are available at https://github.com/Dalei-Dalei/MEOMI/ and http://122.205.95.139/MEOMI/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

How to construct accurate gene regulatory networks $(GRN)$ has important significance for the research of functional genomics. Several existing gene regulatory network construction methods have the problems of low accuracy and unable to handle large scale network effectively. In order to solve the above problems, this work proposes a gene regulatory network construction method based on k-tree sampling and decomposition (KtreeGRN). It transforms the problem of gene network construction into generating codes. Firstly, it constructs the k-tree structure of the gene regulatory network through sampling dandelion codes uniformly. Then, the k-tree is decomposed into several k-cliques using the tree decomposition algorithm based on the minimum degree selection. It constructs the sub-networks for the genes in each k-clique using the mixed entropy optimizing mutual information method. Finally, it obtains the whole gene regulatory network through merging all the sub-networks. It repeats the above operations (k-tree generation, k-tree decomposition, network generation) several times and gets the final gene regulatory network according to the frequency of edges. Experimental results show that KtreeGRN performs better than other several kinds of gene network construction methods on the simulated dataset of DREAM challenge and two real datasets of Escherichia-coil (E. coli SOS pathway network, E. coli SOS DNA repair network).

Proposing a more effective and accurate epistatic loci detection method is of great significance in improving crop quality, disease treatment, etc. Due to the characteristics of high accuracy and processing non-linear relationship, Bayesian network (BN) has been widely used in constructing the network of SNPs and phenotypes and thus to mine epistasis. However, the shortcoming of BN is that the search space is too large and unable to process large-scale SNPs. In this work, we propose a kind of epistasis mining method using Markov Chain Monte Carlo (MCMC) sampling optimizing Bayesian network (EpiMCBN). Firstly, we use the space of node order composed of SNPs and phenotype to replace the space of network structure. Then MCMC algorithm is used to do sampling to generate multiple different initial orders in linear space or partial space. We use Markov state transition matrix to transfer the initial samples along the Markov chain, thus obtaining multiple order samples. Then we use the $\alpha$-BICBN scoring function to score the Bayesian networks corresponding to these node orders. Through estimating the probability of edge occurrence in the Bayesian networks, we get an approximate Bayesian network of SNPs and phenotype, then obtain the epistatic loci affecting phenotype. Finally, we compare EpiMCBN with the current popular epistasis mining algorithms using both simulated and real age-related macular disease (AMD) datasets. Experiment results show that EpiMCBN has better epistasis detection accuracy, lower false positive rate, and higher F1-score compared to other methods. Availability and implementation: Source code and dataset are available at: http://122.205.95.139/EpiMCBN/.

Proposing a more effective and accurate epistatic loci detection method in large-scale genomic data has important research significance for improving crop quality, disease treatment, etc. Due to the characteristics of high accuracy and processing non-linear relationship, Bayesian network (BN) has been widely used in constructing the network of SNPs and phenotype traits and thus to mine epistatic loci. However, the shortcoming of BN is that it is easy to fall into local optimum and unable to process large-scale of SNPs. In this work, we transform the problem of learning Bayesian network into the optimization of integer linear programming (ILP). We use the algorithms of branch-and-bound and cutting planes to get the global optimal Bayesian network (ILPBN), and thus to get epistatic loci influencing specific phenotype traits. In order to handle large-scale of SNP loci and further to improve efficiency, we use the method of optimizing Markov blanket to reduce the number of candidate parent nodes for each node. In addition, we use α-BIC that is suitable for processing the epistatis mining to calculate the BN score. We use four properties of BN decomposable scoring functions to further reduce the number of candidate parent sets for each node. Experiment results show that ILPBN can not only process 2-locus and 3-locus epistasis mining, but also realize multi-locus epistasis detection. Finally, we compare ILPBN with several popular epistasis mining algorithms by using simulated and real Age-related macular disease (AMD) dataset. Experiment results show that ILPBN has better epistasis detection accuracy, F1-score and false positive rate in premise of ensuring the efficiency compared with other methods. Availability: Codes and dataset are available at: http://122.205.95.139/ILPBN/.

Wenzhong Shia , Pengfei Chenab* , Xiaoqi Shenc & Jianxiao Liuaa Department of Land Surveying and Geo-Informatics, The Hong Kong Polytechnic University, Hong Kong, Chinab School of Geospatial Engineering and Science, Sun Yat-Sen University, Guangzhou, Guangdong, Chinac School of Environment Science and Spatial Informatics, China University of Mining and Technology, Xuzhou, Jiangsu, ChinaWenzhong Shi is the Head and Chair Professor of Geographical Information Science and Remote Sensing in the Department of Land Surveying and Geo-informatics at the Hong Kong Polytechnic University. His research interests include GIScience and remote sensing, focusing on uncertainties and quality control of spatial data, satellite images and LiDAR data, 3D modelling, and human dynamics.Pengfei Chen works as an assistant professor in School of Geospatial Engineering and Science at Sun Yat-Sen University. His research interests include human mobility modeling, geospatial artificial intelligence and spatial uncertainty.Xiaoqi Shen is a PhD candidate in School of Environment Science and Spatial Informatics at China University of Mining and Technology. His research interest includes human mobility analysis, geographical data mining, and geospatial artificial intelligence.Jianxiao Liu currently is a research assistant in the Department of Land Surveying and Geo-informatics at the Hong Kong Polytechnic University. His research interests include urban and rural planning, human mobility, and geographical data mining.CONTACT Pengfei Chen [email protected] movement uncertainty is of profound significance in promoting effective trajectory analysis and mining. However, classic uncertainty models are limited by rigid assumptions on moving speed and distance, which ignores the stochastic nature of individual’s travel behaviour. This study introduces a novel method using adaptive ellipses to represent the movement uncertainty in a planar space under the framework of time geography. Two models are established by considering different error sources in trajectory data. The first model captures the uncertainty caused by sampling error, and the second one further, takes the measurement error into account. The Minkowski distance metric is adopted to determine the size of uncertainty ellipses, in which the Minkowski parameter is optimized for each segment in the raw trajectory on the basis of the geometric characteristics extracted. Compared with state-of-the-art methods on five real-life trajectory datasets, the proposed method is proved to produce more effective uncertain regions, which significantly reduce the redundant uncertain area, while retaining at a comparative level of actual movement coverage. Given the heterogeneity of human mobility patterns, this study provides a robust and applicable solution for adaptively modelling individual’s movement uncertainty, which is expected to benefit trajectory-related applications in various scenarios.

How to mine the interaction between SNPs (namely epistasis) efficiently and accurately must be considered when to tackle the complexity of underlying biological mechanisms. In order to overcome the defect of low learning efficiency and local optimal, this work proposes an epistasis mining method using artificial fish swarm optimizing Bayesian network (AFSBN). This method uses the characteristics of global optimization, good robustness and fast convergence about the artificial fish swarm algorithm, and uses the algorithm into the heuristic search strategy of Bayesian network. The initial network structure can be evolved through the manipulations of foraging behavior, clustering behavior, tail-chasing behavior and random behavior. This algorithm chooses different behaviors to modify the network state according to the changing of surrounding environment and the states of partners. It realizes the interaction between each artificial fish and its neighboring environment, and finally finds the optimal network in the population. We compared AFSBN with other existing algorithms on both simulated and real datasets. The experimental results demonstrate that our method outperforms others in epistasis detection accuracy in the case of not affecting the efficiency basically for different datasets.

Epistasis is an important factor affecting complex disease, some powerful methods should be proposed in order to improve large-scale genome-wide data mining efficiency. There are some problems exist in epistasis detection methods, such as low efficiency, low accuracy and inability to process large number of SNPs. In this work, we propose a k-tree optimizing Bayesian network (BN) epistasis mining method that can deal with large-scale of SNPs. Firstly, we construct the k-tree including large-scale of SNP loci and phenotype traits by sampling the Dandelion code uniformly. Then the k-tree is decomposed into different k-cliques by using the degree selection tree decomposition algorithm. In different k-cliques, the optimized fast incremental association BN learning method (omb-Fast) is used to learn the sub-Bayesian network quickly and accurately. Finally, all the sub-networks are merged to obtain the whole network. The above operations (k-tree generation, k-tree decomposition, network generation) are done several times, and thus to obtain the epistatic loci affecting phenotype traits. The simulated experiments validate the effective of our method. Experiment results show that the proposed method has better epistasis detection accuracy, lower false positive rate and higher F1-score on the basis of ensuring the efficiency compared to other methods. Most importantly, it can be used into large-scale of SNPs in the whole-genome for epistasis detection.

BACKGROUND: Gene selection refers to find a small subset of discriminant genes from the gene expression profiles. How to select genes that affect specific phenotypic traits effectively is an important research work in the field of biology. The neural network has better fitting ability when dealing with nonlinear data, and it can capture features automatically and flexibly. In this work, we propose an embedded gene selection method using neural network. The important genes can be obtained by calculating the weight coefficient after the training is completed. In order to solve the problem of black box of neural network and further make the training results interpretable in neural network, we use the idea of knockoffs to construct the knockoff feature genes of the original feature genes. This method not only make each feature gene to compete with each other, but also make each feature gene compete with its knockoff feature gene. This approach can help to select the key genes that affect the decision-making of neural networks. RESULTS: We use maize carotenoids, tocopherol methyltransferase, raffinose family oligosaccharides and human breast cancer dataset to do verification and analysis. CONCLUSIONS: The experiment results demonstrate that the knockoffs optimizing neural network method has better detection effect than the other existing algorithms, and specially for processing the nonlinear gene expression and phenotype data.

How to mine the gene regulatory relationship and construct gene regulatory network (GRN) is of utmost interest within the whole biological community, however, which has been consistently a challenging problem since the tremendous complexity in cellular systems. In present work, we construct gene regulatory network using an improved three-phase dependency analysis algorithm (TPDA) Bayesian network learning method, which includes the steps of Drafting, Thickening, and Thinning. In order to solve the problem of learning result is not reliable due to the high order conditional independence test, we use the entropy estimation approach of Gaussian kernel probability density estimator to calculate the (conditional) mutual information between genes. The experiment on the public benchmark data sets show the improved method outperforms the other nine kinds of Bayesian network learning methods when to process the data with large sample size, with small number of discrete values, and the frequency of different discrete values is about same. In addition, the improved TPDA method was further applied on a real large gene expression data set on RNA-seq from a global collection with 368 elite maize inbred lines. Experiment results show it performs better than the original TPDA method and the other nine kinds of Bayesian network learning algorithms significantly.

Mining epistatic gene locus which influence complex disease has great research significance. Bayesian network (BN) has been widely used in many researches of epistasis mining. However, Bayesian network methods have disadvantages of being easily trapped into local optimum, low learning efficiency and not being able to handle large-scale network. In this work, we propose an epistasis mining approach based on artificial bee colony algorithm optimizing Bayesian network (BnBeeEpi). We apply artificial bee colony algorithm into the heuristic search strategy of Bayesian network, and then use two kinds of BN scoring functions (BIC and MIT) to calculate the network fitness value to avoid overfitting and reduce false positive rate. Moreover, we introduce decomposable BIC scoring to solve the large-scale network learning problem. Finally, we compare BnBeeEpi with current popular epistasis mining algorithms by using both simulated and real datasets. Experiment results show that omb-Fast has very short running time with its accuracy is as good as other methods, and BnBeeEpi has better F1-score and lower false positive rate compared to others. Availability and implementation: codes and visualization platform are available at: http://106.14.132.202/.

SUMMARY: CRISPR-Local is a high-throughput local tool for designing single-guide RNAs (sgRNAs) in plants and other organisms that factors in genetic variation and is optimized to generate genome-wide sgRNAs. CRISPR-Local outperforms other sgRNA design tools in the following respects: (i) designing sgRNAs suitable for non-reference varieties; (ii) screening for sgRNAs that are capable of simultaneously targeting multiple genes; (iii) saving computational resources by avoiding repeated calculations from multiple submissions and (iv) running offline, with both command-line and graphical user interface modes and the ability to export multiple formats for further batch analysis or visualization. We have applied CRISPR-Local to 71 public plant genomes, using both CRISPR/Cas9 and CRISPR/cpf1 systems. AVAILABILITY AND IMPLEMENTATION: CRISPR-Local can be freely downloaded from http://crispr.hzau.edu.cn/CRISPR-Local/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Mining epistatic loci which affects specific phenotypic traits is an important research issue in the field of biology. Bayesian network (BN) is a graphical model which can express the relationship between genetic loci and phenotype. Until now, it has been widely used into epistasis mining in many research work. However, this method has two disadvantages: low learning efficiency and easy to fall into local optimum. Genetic algorithm has the excellence of rapid global search and avoiding falling into local optimum. It is scalable and easy to integrate with other algorithms. This work proposes an epistasis mining approach based on genetic tabu algorithm and Bayesian network (Epi-GTBN). It uses genetic algorithm into the heuristic search strategy of Bayesian network. The individual structure can be evolved through the genetic operations of selection, crossover and mutation. It can help to find the optimal network structure, and then further to mine the epistasis loci effectively. In order to enhance the diversity of the population and obtain a more effective global optimal solution, we use the tabu search strategy into the operations of crossover and mutation in genetic algorithm. It can help to accelerate the convergence of the algorithm. RESULTS: We compared Epi-GTBN with other recent algorithms using both simulated and real datasets. The experimental results demonstrate that our method has much better epistasis detection accuracy in the case of not affecting the efficiency for different datasets. CONCLUSIONS: The presented methodology (Epi-GTBN) is an effective method for epistasis detection, and it can be seen as an interesting addition to the arsenal used in complex traits analyses.

Web service composition is to assemble single services, and it can realize the function that cannot be completed by the single services. How to composite services according to user's requests quickly and accurately is a key problem to be solved in the era of service-oriented software engineering (SOSE). On the basis of storing service information, we get execution dependency relations among services in the consideration of service interface (Input and Output) and capability (Precondition and Effect) firstly. Then the join operation in relational database (RDB) is used to determine the service composition path quickly according to user's requests. The semantic reasoning relationship between concepts and the concept status path are used to do the calculation, and it helps to enhance the accuracy of service composition. The effectiveness of the proposed method is validated through experiments.

In the era of service-oriented software engineering (SOSE), service clustering is used to organize Web services, and it can help to enhance the efficiency and accuracy of service discovery. In order to improve the efficiency and accuracy of service clustering, this paper uses the self-join operation in relational database (RDB) to realize Web service clustering. Based on storing service information, it does the self-join operation towards the Input, Output, Precondition, Effect (IOPE) tables of Web services, which can enhance the efficiency of computing services similarity. The semantic reasoning relationship between concepts and the concept status path are used to do the calculation, which can improve the calculation accuracy. Finally, we use experiments to validate the effectiveness of the proposed methods.

With the increasing number of Web services available on the Internet, how to recommend Web services to interested users effectively and efficiently remains to be a big challenge. At present, collaborative filtering (CF) is the most widely used technique in the design of recommender systems to handle information overload. For Web services, however, it is difficult for user to collect personalized QoS (Quality of Service)data and other explicit feedbacks such as ratings. In most cases, only a part of the implicit feedbacks (e.g., watchlist) is available in service registry. In this paper, we leverage implicit feedback from user's watchlist to build a CF-based recommender system for Web service. Our main contribution is to transform implicit feedbacks into explicit ratings to improve the accuracy of service recommendation. More specifically, we first construct binary user-service rating matrix according to the implicit feedback from the watchlist. Then, temporal and tag preference are combined into the original rating matrix to generate a more accurate pseudo rating matrix, which can reflect users' different preference on services in their own watchlists. Finally, we use traditional user-based CF method to produce a personalized service recommendation list with corresponding pseudo ratings. Moreover, the empirical experiments based on ProgrammableWeb show that compared with traditional log-based CF method, the recommender system with temporal and tag preference is more accurate and precise.

As one foundational technology of cloud computing, services computing is playing a critical role to enable provisioning of software as a service (SaaS). However, how to effectively and efficiently discover proper available services from the cloud of resources remains a big challenge. This paper reports our continuous efforts on semantic services discovery. We extend the Support Vector Machine (SVM)-based text clustering technique in the context of service-oriented categorization in a service repository, and propose an iterative process to incrementally enrich domain ontology. A popular Web 2.0 mashup platform is used as a testbed; and preliminary evaluation results are reported.

With the development of web service applications, how to realize on-demand services selection according to user's personal requirements is a hot research work in modern times. WS-CDL (Web Services Choreography Description Language) is a W3C candidate recommendation for the description of peer-to-peer collaborations for the participants in web services interaction. However, the goal of the service choreography is lacking and we can't understand the information exchange from the whole point. This paper extends WS-CDL from the aspect of goal that services to implement, and this can lay the foundation of on-demand service selection. Finally, the feasibility of the proposed method is validated through a case study.

Cloud computing has been attracting much attention since its birth. How to cloudify software systems especially legacy applications in the cloud era is becoming increasingly important. Based on RGPS meta-model framework and International standards-ISO/IEC 19763, an architecture for cloudification of legacy applications is proposed, which consists of three parts: a Web portal, a SaaS service supermarket, and a SaaS application development platform. In this paper, we take an open source software as an example to illustrate the proposed approach. Based on the architecture and supporting techniques on software virtualization and multi-tenancy, we develop a prototype Cloud CRM to demonstrate the basic procedure for cloudification of legacy applications, as well as the feasibility of the proposed approach.