This paper proposes a parallel branch and bound algorithm that efficiently runs on the Grid. The proposed algorithm is parallelized with the hierarchical master-worker paradigm in order to efficiently compute fine-grain tasks on the Grid. The hierarchical algorithm performs master-worker computing in two levels, computing among PC clusters on the Grid and that among computing nodes in each PC cluster, and reduces communication overhead by localizing frequent communication in tightly coupled computing resources, or a PC cluster. On each PC cluster, granularity of tasks dispatched to computing nodes is adaptively adjusted to obtain the best performance. The algorithm is implemented on the Grid testbed by using GridRPC middleware, Ninf-G and Ninf. In the implementation, communication among PC clusters is securely performed via Ninf-G using the Grid Security Infrastructure, and fast communication in each PC cluster is performed via Ninf. The experimental results showed that parallelization with the hierarchical master-worker paradigm using combination of Ninf-G and Ninf effectively utilized computing resources on the Grid in order to run a fine-grain application. The results also showed that the adaptive task granularity control automatically gave the same or better performance compared to performance with manual control.

View full abstract

This study proposes asynchronous MPI, a simple and effective parallel programming model for SMP clusters, to reimplement the High PerformanceLinpack benchmark. The proposed model forces processors of an SMP node to work in different phases, thereby avoiding unneccessary communication and computation bottlenecks. As a result, we can achieve significant improvements in performance with a minimal programming effort. In comparison with a de-facto flat MPI solution, our algorithm can yield a 20.6% performance improvement for a 16-node cluster of Xeon dual-processor SMPs.

View full abstract

This paper proposes a novel approach to discover dynamic laws and models represented by simultaneous time differential equations including hidden states from time series data measured in an objective process. This task has not been addressed in the past work though it is essentially important in scientific discovery since any behaviors of objective processes emerge in time evolution. The promising performance of the proposed approach is demonstrated through the analysis of synthetic data.

View full abstract

We propose a framework in this paper for transforming programs with templates based on term rewriting. The programs are given by term rewriting systems. We discuss how to validate the correctness of program transformation within our framework. We introduce a notion of developed templates and a simple method of constructing such templates without explicit use of induction. We then show that in any transformation of programs using the developed templates, their correctness can be verified automatically. The correctness of program transformation within our framework is discussed based on operational semantics. We also present some examples of program transformations in our framework.

View full abstract

We specified a mechanism with which Mobile IPv6 and IPsec/IKE can work together efficiently. The interaction is necessary for updating the endpoint address of an IPsec tunnel in accordance with movement performed by a mobile node. Based on an analysis of needs for interaction between Mobile IPv6 and IPsec/IKE, we designed and implemented a mechanism that is an extension to the PF_KEY framework. The proposed mechanism allows Mobile IPv6 to inform IPsec/IKE of the movement so that necessary updates to the security policy database and security association database can be taken by IPsec/IKE.This notification helps IKE to update its internal state. The mechanism is also applicable to the other scenarios, such as NEMO, Mobile VPN and its variants.

View full abstract

Gene classification based on orthologous relations is an important problem to understand species-universal or species-specific conservation of genes in genomes associated with phenotype in species. In the present study, we proposed a classification system of genes based on configuration of networks concerning bidirectional best-hit relations (called orthologous relation group), which makes it possible to compare multiple genomes. We have applied this method to five Bacillus species (B. subtilis, B. anthracis, B. cereus, B. halodurans, andB. thuringiensis). With regards to the the five species, 4, 776 orthologous relation groups have been obtained, and those are classified into 113 isomorphic groups. An isomorphic group may contain only orthologs or a combination of orthologs and paralogs. Gene functions and the conservativeness are discussed in view of configuration of orthologous relation groups.

View full abstract

Many attempts have so far been made at modeling RNA secondary structure by formal grammars. In a grammatical approach, secondary structure prediction can be viewed as parsing problem. However, there may be many different derivation trees for an input sequence. Thus, it is necessary to have a method of extracting biologically realistic derivation trees among them. One solution to this problem is to extend a grammar to a probabilistic model and find the most likely derivation tree, and another is to take free energy minimization into account. One simple formalism for describing RNA folding is context-free grammars(CFGs), but it is known that CFGs cannot represent pseudoknots. Therefore, several formal grammars have been proposed for modeling RNA pseudoknotted structure. In this paper, we focus on multiple context-free grammars (MCFGs), which are natural extension of CFGs and can represent pseudoknots, and extend MCFGs to a probabilistic model called stochastic MCFG (SMCFG). We present a polynomial time parsing algorithm for finding the most probable derivation tree, which is applicable to RNA secondary structure prediction including pseudoknots. Also, we propose a probability parameter estimation algorithm based on the EM (expectation maximization) algorithm. Finally, we show some experimental results on RNA pseudoknot prediction using the SMCFG parsing algorithm, which show good prediction accuracy.

View full abstract

We are developing PROFESS, a system to assist with the extraction of protein functional site information from the literature related to protein structural analysis. In this system, the sentences with functional information are first extracted. This paper proposes the complementary use of the protein structure data, keywords and patterns to extract the target sentences. In the proposed method, the sentences in the literature are expressed in vector using these three features, which are learnt by the SVM. As the accuracy of the SVM depends on the number of effective vector elements, we propose a method to automatically extract patterns to add as new vector elements and obtain a higher value in accuracy. There is a problem of matching of the patterns to the sentences when any proper noun tag is expressed adjacent to residue tag. We defined two rules to eliminate these unnecessary tags so that the patterns can match to the sentences. The proposed method was applied to five documents related to structural analysis of protein for extracting sentences with protein functional information, where eight literatures were used for the feedback for each of the experiment literatures. The average recall value and F value were 0.96 and 0.69, respectively. It was confirmed that the increase of the number of the vector elements lead to a higher performance in the sentence extraction.

View full abstract

A Protein-Protein Interaction network, what we call a PPI network is considered as an important source of information for prediction of protein functions. However, it is quite difficult to analyze such networks for their complexity. We expected that if we could develop a good visualizing method for PPI networks, we could predict protein functions visually because of the close relation between protein functions and protein interactions. Previously, we proposed one, which is based on clustering concepts, by extracting clusters defined as relatively densely connected group of nodes. But the results of visualization of a network differ very much depending on the clustering algorithm. Therefore, in this paper, we compare the outcome of two different clustering algorithms, namely DPClus and Newman algorithms, by applying them to a PPI network, and point out some advantages and limitations of both.

View full abstract

Comparative analyses of the metabolic networks among different species provide important information regarding the evolution of organisms as well as pharmacological targets. In this paper, a method is proposed for comparing metabolic networks based on enzymatic reactions within different species. Specifically, metabolic networks are handled as sets of enzymatic reactions. Based on the presence or absence of metabolic reactions, the metabolic network of an organism is represented by a bit string comprised of the digits “1” and “0, ” called the “reaction profile.” Then, the degree of similarity between bit strings is defined, followed by clustering of metabolic networks by different species. By applying our method to the metabolic networks of 33 representative organisms selected from bacteria, archaea, and eukaryotes in the MetaCyc database, a phylogenetic tree was reconstructed that represents the similarity of metabolic network based on metabolic phenotypes.

View full abstract

Multifactorial diseases, such as lifestyle-related diseases, for example, cancer, diabetes mellitus, and myocardial infarction, are believed to be caused by the complex interactions between various environmental factors on a polygenic basis. In addition, it is believed that genetic risk factors for the same disease differ on an individual basis according to their susceptible environmental factors. In the present study, to predict the development of myocardial infarction (MI) and classify the subjects into personally optimum development patterns, we have extracted risk factor candidates (RFCs) that comprised a state that is a derivative form of polymorphisms and environmental factors using a statistical test. We then selected the risk factors using a criterion for detecting personal group (CDPG), which is defined in the present study. By using CDPG, we could predict the development of MI in blinded subjects with an accuracy greater than 75%. In addition, the risk percentage for MI was higher with an increase in the number of selected risk factors in the blinded data. Since sensitivity using the CDPG was high, it can be an effective and useful tool in preventive medicine and its use may provide a high quality of life and reduce medical costs.

View full abstract

We used microarrays to examine individual-based differences in gene expression in primary cultures of renal tubular cells derived from Japanese subjects. The subjects had solitary tumors in the kidney or urinary tract, which were diagnosed pathologically as renal cell carcinoma or transitional cell carcinoma. Renal tissue samples collected from a non-tumorous portion of the tissue were regarded as normal tissues, as there were no abnormal microscopic findings and no evidence of renal dysfunction from the clinical laboratory data. The genome-wide gene expression profiles of nine human renal cell cultures were analyzed using the Affymetrix GeneChip HG-U133A and HG-U133B arrays. Approximately 8, 500 transcripts exhibited significant differential expression (p < 0.05) among the subjects, and the coefficients of variation for 1, 338 transcripts were greater than 50%. Some of these transcripts encode drug-metabolizing enzymes (e.g., UGT1A8 and UGT1A9) or sodium/phosphate cotransporters (e.g., PDZK1). These data provide the basis for toxicogenomic studies using primary cultured renal cortical cells from Japanese subjects.

View full abstract

The recently developed array-based comparative genomic hybridization(array CGH) technique measures DNA copy number aberrations that occur as causes or consequences of cell diseases such as cancers. Conventional array CGH analysis classifies DNA copy number aberrations into three categories: no significant change, significant gain, and significant loss. However, recent improvements in microarray measurement precision enable more quantitative analysis of copy number aberrations. We propose a method, called comb fitting, that extracts a quantitative interpretation from array CGH data. We also propose modifications that allow us to apply comb fitting to cases featuring heterogeneity of local aberrations in DNA copy numbers. By using comb fitting, we can correct the baseline of the fluorescence ratio data measured by array CGH and simultaneously translate them into the amount of changed copy numbers for each small part of the chromosome, such as 0, ±1, ±2, ···. Comb fitting is applicable even when a considerable amount of contamination by normal cells exists and when heterogeneity in the ploidy number cannot be neglected.

View full abstract

The development of physiological cell models to support the understanding of biological mechanisms gains increasingly importance. Due to the complexity of biological systems, whole cell models, which are composed of many imported component models of functional elements, get quite complex, making modifications difficult. Here, we propose a method to enhance structural changes of cell models, employing the markup languages of CellML and our original PMSML (Physiological Model Structure Markup Language), in addition to a new ontology for cell physiological modelling, the Cell Model Ontology. In particular, a method to make references from CellML files to the ontology and a method to assist with manipulation of model structures using PMSML together with the Cell Model Ontology are reported. Using these methods two software utilities, an interactive ontology ID assigner, the CellML Ontologizer, and a graphical cell model editor, the Cell Structure Editor, are implemented. Experimental results proved that the proposed method and the implemented software are useful for the modification of physiological models.

View full abstract

Demands on efficient drug design have been increasing with the advancement of computing technology and bioinformatics. A variety of information technologies pertaining to drug design have been proposed recently and such technology mostly contributes to drug design research. Molecular docking simulation is a promising application for drug design, and can be realized with current information technology. However although docking simulation and the related information technology have advanced in recent years, scientists still have difficulty finding a suitable parameter set of docking simulations for accuracy of simulation. The parameter-tuning step takes a long time, and existing computing technology can hardly assist in this step. This is because the parameter-tuning step involves factors that are difficult to automate with computers. In this paper, we propose a new architecture for assisting procedures that require the decisions of scientists, especially when they need to tune parameters in a docking simulation.

View full abstract

Error Correction Code (ECC) is widely used to detect and correct soft errors in VLSI caches. Maintaining ECC on a per-word basis, which is preferable in caches with word-based access, is expensive. This paper proposes Zigzag-HVP, a cost-effective technique to detect and correct soft errors in such caches. Zigzag-HVP utilizes horizontal-vertical parity (HVP). HVP maintains the parity of a data array both horizontally and vertically. Basic HVP can detect and correct a single bit error (SBE), but not a multi-bit error (MBE). By dividing the data array into multiple HVP domains and interleaving bits of different domains, a spatial MBE can be converted to multiple SBEs, each of which can be detected and corrected by the corresponding parity domain. Vertical parity updates and error recovery in Zigzag-HVP can be efficiently executed through modifications to the cache data paths, write-buffer, and Built-In Self Test. The evaluation results indicate that the area and power overheads of Zigzag-HVP caches are lower than those of ECC-based ones.

View full abstract

We propose a new algorithm that is effective for objects that are shared among threads but are not contended for in SMP environments. We can remove the overhead of the serialization between lock and other non-lock operations and avoid the latency of complex atomic operations in most cases. We established the safety of the algorithm by using a software tool called Spin. The experimental results from our benchmarking on an SMP machine using Intel Xeon processors revealed that our algorithm could significantly improve efficiency by 80% on average compared to using complex atomic instruction.

View full abstract

Sequential pattern mining is a useful technique used to discover frequent subsequences as patterns in a sequence database. Depending on the application, sequence databases vary by number of sequences, number of individual items, average length of sequences, and average length of potential patterns. In addition, to discover the necessary patterns in a sequence database, the support threshold may be set to different values. Thus, for a sequential pattern-mining algorithm, responsiveness should be achieved for all of these factors. For that purpose, we propose a candidate-driven pattern-growth sequential pattern-mining algorithm called FSPM (Fast Sequential Pattern Mining). A useful property of FSPM is that the sequential patterns concerning a user-specified item can be mined directly. Extensive experimental results show that, in most cases FSPM outperforms existing algorithms. An analytical performance study shows that it is the inherent potentiality of FSPM that makes it more effective.

View full abstract

Self-similar traffic patterns have been observed in many measurements of Internet traffic. Self-similarity is very detrimental to the performance of packet networks and recent research has focused on understanding and reducing its amount in Internet traffic. TCP Reno has been identified as being one of the primary sources of self-similarity. We explore the potential of another version of TCP in this paper to reduce the degree of self-similarity in aggregated TCP traffic. We decompose both TCP Reno and TCP Vegas to demonstrate and explain their underlying mechanisms, and separately measure what effects congestion-avoidance and timeouts/exponential backoff mechanisms have on the self-similarity in aggregated TCP flows. We reveal how TCP Vegas reduces the degree of self-similarity and eventually completely eliminates it from aggregated TCP flows at low levels of packet loss. However, at high levels of packet loss we show that TCP Vegas is detrimental, because it increases the degree of aggregated TCP-flow self-similarity.

View full abstract

Copyright protection is a major issue in online content-distribution services and many key-management schemes have been proposed for protecting content. Key-distribution processes impose large burdens even though the communications bandwidth itself is restricted in the distribution of mobile content provided to millions of users. Mobile devices also have low computational capacities. Thus, a new scheme of key management, where the load on the key-distribution server is optimal and loads on clients are practical, is required for services. Tree-based schemes aim at reducing the load on the server and do not take reducing the load on clients into account. The load on clients is minimized in a star-based scheme, on the other hand, while the load on the server increases in proportion to the number of clients. These structures are far from being scalable. We first discuss a relaxation of conventional security requirements for key-management schemes in this paper and define new requirements to improve the efficiency of the schemes. We next propose the τ-gradual key-management scheme. Our scheme satisfies the new security requirements and loads on the server, and it has far fewer clients than conventional schemes. It uses an intermediate configuration between that of a star- and a tree-structure that allows us to continuously change it by controlling the number of clients in a group, m_max. The scheme can be classified as τ-star-based, τ-tree-based, or τ-intermediate depending on the parameter, m_max. We then present a quantitative evaluation of the load on the server and clients using all our schemes based on practical assumptions. The load on the server and that on clients involves a trade-off with the τ-intermediate scheme. We can construct an optimal key-management structure according to system requirements using our schemes, while maintaining security. We describe a concrete strategy for setting parameter m_max. Finally, we present general parameter settings by which loads on both the server and clients using the τ-intermediate scheme are lower than those using the τ-tree-based scheme.

View full abstract

In on-demand ad-hoc routing protocols, control packets are flooded into a network during path discovery. Especially in dense ad-hoc networks, these protocols generate a large number of broadcast packets that cause contention, packet collisions and battery power wastage in mobile nodes. We propose an efficient route establishment method that adaptively lowers re-broadcasting overhead based on the number of adjacent nodes and the number of routes that these nodes accommodate. Through simulation, we demonstrate that our proposed method is especially efficient in densely populated areas. It achieves efficiency by lowering the number of control packets for path discovery without causing a drop in the path discovery success ratio. Also, by taking path concentration into account, our method further improves packet delivery. We provide useful guidelines based on simulation results.

View full abstract

Finding an edit script between data has played an important role in data retrieval and data transformation. So far many methods for finding an edit script between two XML documents have been proposed, but few studies on finding an edit script between an XML document and a DTD have been made. In this paper, we first present a polynomial-time algorithm for finding an edit script between an XML document and a DTD, which is optimum under some restrictions on operations. We next prove the correctness of the algorithm.

View full abstract

An active database is a database system that can react to internal, as well as external, database events. The reactive behavior of an active database is determined by a predefined set of active database rules, along with a rule processing strategy. A common problem associated with active database systems is the possible non-termination of the active database rules. Previous work on the analysis of the conditions required for the termination of active database rules has only considered limited rule processing strategies. This paper proposes an approach for automatically detecting the non-termination of active database rules using a model checking technique. With this approach, a general framework for modeling active database systems is first proposed. This framework is useful for analyzing the behavior of rules with different rule processing strategies and for allowing the adoption of different contexts and different execution coupling modes for the active database rules. Based on the proposed modeling framework, the termination property of active database rules with various rule processing strategies is next checked using SPIN, a model checking tool. Through experimental results, we demonstrated the feasibility of using this method.

View full abstract

As the recent advances in the wireless technologies and mobile terminals, mobile users equipped with mobile devices are able to access wireless services through 3G cellular network, WiFi hotspot, or WiMAX link, as well as through satellite digital broadcast or terrestrial digital broadcast. By effectively taking advantage of these complementary communication modes, we explore a new hybrid data delivery model, i.e., Hybrid Wireless Broadcast (HWB) model to benefit from the optimal combination of the push-based and pull-based broadcast and on-demand point-to-point wireless communication. The HWB model can provide a flexible and complementary information service in different bandwidths and service ranges, and greatly improve the responsibility, scalability, and efficiency of the system. The results of simulation study show the proposed HWB approach achieves a significant improvement in system performance.

View full abstract

With the advent of mobile technology, broadcasting data over one or more wireless channels has been considered an excellent way to efficiently disseminate data to a large number of mobile users. In such a wireless broadcast environment, the minimization of both access and tuning times is an important problem because mobile devices usually have limited battery power. In this paper, we propose an index allocation method for data access over multiple channels. Our method first derives external index information from the scheduled data, and then allocates it over the channels, which have shorter broadcast cycles and hotter data items. Moreover, local exponential indexes with different parameters are built within each channel for local data search. Experiments are performed to compare the effectiveness of our approach with an existing approach. The results show that our method outperforms the existing method by 30% in average access time and by 16% in average tuning time when the data skew is high and data item size is much bigger than index node size.

View full abstract