Retrieval, alignment, and clustering of computational models based on semantic annotations

In this poster, we present our experience with techniques for the management of bio-models (i.e. XML-encoded computational models of biological systems) in repositories. These techniques support users to find, access, adapt and then re-submit models - tasks which are essential for the reuse of models in systems biology. Here, we focus on model retrieval, the ranking of search results and concepts to track the history of a model through version control. These concepts have so far not been sufficiently addressed in existing systems.To validate our techniques, we have been implemented a framework called Sombi. The underlying database called mDB is tailored towards a fine-grained storage of the model and its annotations. The design is applicable to different representation formats but, to begin with, has been tested with SBML models (Köhn et al., 2009).We furthermore present a retrieval algorithm for bio-model repositories (Henkel et al., 2010). The algorithm has been implemented and validated with mDB and the BioModels Database (www.ebi.ac.uk/biomodels-main/). The ranked retrieval enhances the result set by incorporating expert knowledge in the annotations. In contrast to existing systems, our approach provides a ranked result which is justified and explains the ordering of retrieved models to the users. The mDB design furthermore enables the application of structured queries on the stored model data.Finally, the poster proposes a solution to model version control. We introduce a library called BiVeS for tracking a model’s history in a diff tree; i.e. a tree-like storage of changes between model versions. These changes are additionally classified into modifications to XML structure, biology and mathematics. This information helps to understand existing models.All approaches are independent of the SBML format and can therefore be applied to other XML based exchange formats. All code is freely available on Sourceforge from http://bives.sourceforge.net/ and http://sombi.sourceforge.net/.

S-adenosylmethionine (SAMe) therapy in liver disease: A review of current evidence and clinical utility

The opportunistic intracellular bacterial pathogen Listeria monocytogenes has in 30 years emerged as an exceptional bacterial model system in infection biology. Research on this bacterium has provided considerable insight into how pathogenic bacteria adapt to mammalian hosts, invade eukaryotic cells, move intracellularly, interfere with host cell functions and disseminate within tissues. It also contributed to unveil features of normal host cell pathways and unsuspected functions of previously known cellular proteins. This review provides an updated overview of our knowledge on this pathogen. In many examples, findings on L. monocytogenes provided the basis for new concepts in bacterial regulation, cell biology and infection processes.

Model repositories such as BioModels Database provide computational models of biological systems for the scientific community. These models contain rich semantic annotations that link model entities to concepts in well-established bio-ontologies such as Gene Ontology. Consequently, thematically similar models are likely to share similar annotations. Based on this assumption, we argue that semantic annotations are a suitable tool to characterize sets of models. These characteristics can then help to classify models, to identify additional features for model retrieval tasks, or to enable the comparison of sets of models. In this paper, we present four methods for annotation-based feature extraction from model sets. All methods have been used with four different model sets in SBML format and taken from BioModels Database. To characterize each of these sets, we analyzed and extracted concepts from three frequently used ontologies for SBML models, namely Gene Ontology, ChEBI and SBO. We find that three of the four tested methods are suitable to determine characteristic features for model sets. The selected features vary depending on the underlying model set, and they are also specific to the chosen model set. We show that the identified features map on concepts that are higher up in the hierarchy of the ontologies than the concepts used for model annotations. Our analysis also reveals that the information content of concepts in ontologies and their usage for model annotation do not correlate.KeywordsGene OntologyFeature ExtractionModel RetrievalSemantic AnnotationModel AnnotationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Speaker recognition is decided by a similarity measure (D) calculated from comparing selected feature vectors among an input speech signal sequence of feature vectors (A) and a selected sequence (B) of reference vectors selected from a plurality of pre-stored reference sequences. Prior to comparison of the input and reference vector sequences, the two sequences are time normalized to align corresponding feature vectors. A significant sound specifying signal (V) including a time sequence of elementary signals is generated in synchronism with one of the input and reference sequences and indicates which feature vectors in that one of the input and reference sequences are considered to represent significant sound. The similarity measure (D) is then calculated in accordance with the comparison of those feature vectors in the one sequence which are indicated by the significant sound specifying signal as representing significant sound and the corresponding feature vectors of the other sequence.

In a content-based image retrieval (CBIR) system, indexing feature vectors and the similarity measure between feature vectors are two key factors for retrieval performance. We present a new CBIR system with statistical-model based image feature extraction in the wavelet domain and a Kullback divergence based similarity measure. A two component Gaussian mixture model (GMM) in the wavelet domain is employed and the model parameters are used to form features for image indexing. A new Kullback divergence based similarity measure is then presented for image retrieval. The experimental results demonstrate that the similarity measure based on the Kullback divergence is more effective than conventional similarity measures, such as the city-block distance and the Euclidean distance. It is shown that the new CBIR system, with the combination of the GMM and the new Kullback divergence based similarity measure, outperforms most other methods in retrieval performance for texture images, while keeping a comparable level of computational complexity.

This article proposed a novel human identification method based on retinal images. The proposed system composed of two main parts, feature extraction component and decision-making component. In feature extraction component, first blood vessels extracted and then they have been thinned by a morphological algorithm. Then, two feature vectors are constructed for each image, by utilizing angular and radial partitioning. In previous studies, Manhattan distance has been used as similarity measure between images. In this article, a fuzzy system with Manhattan distances of two feature vectors as input and similarity measure as output has been added to decision-making component. Simulations show that this system is about 99.75% accurate which make it superior to a great extent versus previous studies. In addition to high accuracy rate, rotation invariance and low computational overhead are other advantages of the proposed systems that make it ideal for real-time systems.

An image retrieval method is presented based on shape similarity measure for multimedia and imaging database systems. Inthe proposed algorithm, the spatial and spectral properties of images are combined using the Radon transform, bispectra, andprincipal components analysis. For each model image in the database, the original 2-D image data are reduced to a set of l-Dprojections via the Radon transform, and then a feature vector is calculated from the bispectra of the resultant l-D functions.The principal component analysis is applied to further reduce the dimension of the feature vector so that it can be storedalong with the original image in the database at a small cost of memory. The derived feature vector is considered as the indexor key of the corresponding image, which uniquely identifies the image independent of rotation, translation, and scaling(RTS). For image retrieval, the data feature vector is computed for a query image, and matched against the feature vectors ofall the model images in the database using the Tanimoto similarity measure. The closely matching images are brought out asthe searching results.The proposed technique has been tested on a large image database. The experimental results show that the retrieval accuracyis very high even for query images with low signal-to-noise ratio (SNR).Keywords: Radon transform, bispectra, principal component analysis, RTS invariant shape features, image retrieval,Tanimoto similarity measure, retrieval accuracy

Software development web resources and their artefacts, e.g., method names, tags and programming concepts, are important for software development and maintenance. Semantic mining and extraction of these multi-form software-related artefacts can facilitate in-depth comprehension of other people’s code as well as technical discussions on online Q&A forums such as Stack Overflow, thereby improving the efficiency of knowledge learning and sharing in the software community. However, the fact that online resources are updated on a second-to-second basis hinders people from manually extracting or constructing such knowledge, not to mention the vast amount of online resources. In this thesis, we tackle the problem of knowledge retrieval and discovery in software domain by exploiting the large amount of available online resources with three different operations: retrieve, link, and generate. Firstly, we represent two heterogeneous forms of the online resources: tags and web resources. We propose a neural language model based approach to jointly learn semantic representations of tags and web resources extracted from Q&A websites. Instead of mining textual content of discussions or relying on statistics such as the co-occurrence of tags and web resources, in our proposed approach, low-dimensional vector representations of tags and web resources are mapped to a joint vector space using word embedding techniques, where the semantically related ones would be close to each other. This allows us to measure the relatedness between a web resource and a technical term with high efficiency, and turns a complicated clustering or recommendation task into a trivial K-nearest neighbour search in the embedding space. Our experiments with Stack Overflow data set show that the learned representations work well in capturing the semantic relatedness of tags and web resources, even when they appear in seemingly different contexts. Secondly, we extend the task of global searching for semantic-relatedness pairs of tags and web resources to a context-sensitive setting, which is automatically linking software artefacts mentioned in Q&A threads with web resources. As an easy-to-use mechanism for knowledge sharing, referencing URLs of external web resources has a broad application in Q&A websites. However, there lack of effective methods to manage and reuse already-shared web resources. To fill in the gap, we formulate this issue as a URL-sense disambiguation problem and propose novel metrics to resolve the linking ambiguity. The proposed metrics consider both the global popularity and the local context relatedness of the URL candidates. To evaluate our metrics, we also build up a large knowledge base of already-shared official document URLs based on the posts on Stack Overflow, and present results from systematic evaluations in which the balanced combination of global popularity and local context relatedness is shown capable of significantly improving the performance of resource linking. Finally, we target at extending the scope of the study beyond pre-existing contents, i.e. artifact names, tags or web resources that are already created. Finding good names for language constructs is important for code understanding and maintenance. However, this is not a trivial task. Good method names need to be functionally descriptive and at the same time follow the naming conventions, which include common rules for lexical choices as well as the language-specific conventions. To help developers, especially novices, discover the underlying knowledge in naming, we propose a neural generation network to directly generate conventional method names from natural language description of the code. Different from rule-based or unsupervised methods, we formulate the task of generating high-quality method names from natural language as a sequence to sequence problem and propose using the encoder-decoder model to learn the naming conventions and automatically generate method names. To improve the generation performance, the attention and copying mechanisms are incorporated as well. We conduct comparison experiments with the state-of-the-art baseline models and the results show that the proposed method can achieve significant improvement in method name generation. Besides, we also explore the naming convention transferability across different projects and demonstrate that it is promising to transfer the knowledge learned from massive open source data to specific projects. In summary, our work focuses on modelling and leveraging the semantics of software related artefacts to enhance developers’ experience in knowledge sharing and discovery. The proposed methods provide novel insights into semantics relatedness or similarity measurement of different forms of software related artefacts, and can also be extended to other applications such as web resource dissemination.

Global view of bionetwork dynamics: adaptive landscape.

The classifications of uncertain data turned into one of the dreary procedures in the data mining domain. The uncertain data have tuples with distinctive probability distribution, which helps to find similar class of tuples. When we consider an uncertain data, the feature vector will not be a single valued but a function. In this paper, we proposed fuzzy entropy and similarity measure to characterize the uncertain data through binary decision tree algorithm. Fuzzy entropy is used to find the best split point for the decision tree to handle the uncertain data. Similarity measure is used to make the better decision for the uncertain data with high accuracy. Initially, fuzzy entropy for each feature vector is calculated to select the best feature vector. Then, best split is selected from the selected feature vector. With the help of trained uncertain data, the binary tree starts to grow. Once the split point is selected, then the constructed decision tree is evaluated by the testing phase of uncertain data. The testing data are subjected to the trained decision tree to obtain the classified data. The experimental analyses are made to evaluate the performance of the proposed FUDT approach. Proposed FUDT algorithm is compared with the existing classification algorithm UDT in terms of accuracy and running time. The experimental analysis finalizes that our FUDT algorithm outperforms the existing UDT algorithm.

Anoxybacillus sp. SK 3-4 is a Gram-positive, rod-shaped bacterium and a member of family Bacillaceae. We had previously reported that the strain is an aluminum resistant thermophilic bacterium. This is the first report to provide a detailed analysis of the global transcriptional response of Anoxybacillus when the cells were exposed to 600 mg L-1 of aluminum. The transcriptome was sequenced using Illumina MiSeq sequencer. Total of 708 genes were differentially expressed (fold change >2.00) with 316 genes were up-regulated while 347 genes were down-regulated, in comparing to control with no aluminum added in the culture. Based on Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, the majority of genes encoding for cell metabolism such as glycolysis, sulfur metabolism, cysteine and methionine metabolism were up-regulated; while most of the gene associated with tricarboxylic acid cycle (TCA cycle) and valine, leucine and isoleucine metabolism were down-regulated. In addition, a significant number of the genes encoding ABC transporters, metal ions transporters, and some stress response proteins were also differentially expressed following aluminum exposure. The findings provide further insight and help us to understand on the resistance of Anoxybacillus sp. SK 3-4 toward aluminium.

In this study, we used combined transcriptomics and metabolomics to analyze the H. mutabilis cultivar's genetic and physiological mechanisms during three flower color transition periods (from white to pink, then from pink to red) within the span of one day. As a result, 186 genes were found to be significantly increased with the deepening of the H. mutabilis flower color; these genes were mainly involved in the expression of peroxidase 30, zinc finger protein, phosphate transporter PHO1, etc. In contrast, 298 genes were significantly downregulated with the deepening of H. mutabilis flower color, including those involved in the expression of probable O-methyltransferase 3, copper binding protein 9, and heat stress transcription factor A-6b. Some genes showed differential expression strategies as the flower color gradually darkened. We further detected 19 metabolites that gradually increased with the deepening of the H. mutabilis flower color, including L-isoleucine, palmitic acid, L-methionine, and (+)-7-isonitrobenzene. The content of the metabolite hexadecanedioate decreased with the deepening of the H. mutabilis flower color. Combined transcriptomics and metabolomics revealed that the metabolic pathways, including those related to anthocyanin biosynthesis, cysteine and methionine metabolism, and sulfur metabolism, appear to be closely related to H. mutabilis flower color transition. This study served as the first report on the genetic and physiological mechanisms of short-term H. mutabilis flower color transition and will promote the molecular breeding of ornamental cultivars of H. mutabilis.

With the huge amount and diversity of multimedia resources available on the web, it is becoming necessary to satisfy user’s information and knowledge needs in the mobile computing era. These resources need to be described and correlated systematically to allow applications and web services to access and reuse them properly. Although popular web search engines are very efficient in retrieving any type of information they are not designed to search different types of multimedia information at the same time. Traditional information retrieval techniques should be enhanced by semantic techniques, which analyze web resources information and metadata and use similarity calculations to find correlations between any two resources on the web. Web resources metadata are the attribute values that characterize a multimedia file and give a description about its content. In this research we propose a method that measures the semantic similarity between multimedia resources in the web by using resources’ metadata attributes. CCS concepts: Information Systems---Information Retrieval---Retrieval Models and Ranking ---Similarity Measures.

Large repository of music that can be accessed or downloaded over the Internet, provides a new way of trading or sharing. However, the technologies for features based Music Information Retrieval (MIR), which is a multidisciplinary field of research, are not well established. Existing MIR techniques and products suffer from either limited capabilities or poor performance. In this paper, we proposed a data model that describes the music information using both Music Definition Language (MDL) and Music Manipulation Language (MML), and supports extensible hybrid methods for music classification and similarity measurement. With proposed musical data model, we further developed a hybrid method that combines both contour and rhythm features, and employed an Artificial Neural Network (Unsupervised Kohonen Self-Organized Map) based classification mechanism that maps variations of music pieces to their corresponding originals using a new vector/matrix format defined as MDL. The proposed hybrid method based on a deep-learning mechanism and a new similarity measurement method has been introduced to fulfil analysis on the music classification and their similarity scores. The test results demonstrate that an accuracy of around 70% in the experiments has been achieved.