Guest Editorial: Learning from multiple sources

Selection of genes that are differentially expressed and critical to a particular biological process has been a major challenge in post-array analysis. Recent development in bioinformatics has made various data sources available such as mRNA and miRNA expression profiles, biological pathway and gene annotation, etc. Efficient and effective integration of multiple data sources helps enrich our knowledge about the involved samples and genes for selecting genes bearing significant biological relevance. In this work, we studied a novel problem of multi-source gene selection: given multiple heterogeneous data sources (or data sets), select genes from expression profiles by integrating information from various data sources. We investigated how to effectively employ information contained in multiple data sources to extract an intrinsic global geometric pattern and use it in covariance analysis for gene selection. We designed and conducted experiments to systematically compare the proposed approach with representative methods in terms of statistical and biological significance, and showed the efficacy and potential of the proposed approach with promising findings.

In recent years, huge different biological databases have been stored in various locations. Using distinct data sets from multiple sources results in more reliable data analysis. However, it is so difficult to combine heterogeneous data in one single server. The most obvious reasons include data privacy, large data sizes, costs, and different geographical locations of data sources. In this paper, we present two new algorithms for clustering data from multiple remote data sources using kernel k-means. The first algorithm is the center-based algorithm built on k-means algorithm. The second algorithm uses distributed kernel k-means over multiple data sources. In the distributed scheme, clustering methods are executed only on their local data sources themselves. Partial clustering results are synched between data sources. To evaluate performance of our proposed algorithms, we merged all data from different sources into one large data set to perform kernel k-means. The results showed that our center-based algorithm greatly reduced transmission data between data sources while still yielding acceptable clustering results. Our distributed kernel k-means algorithm achieved even better performance. The clustering results are very close to those generated by kernel k-means on one merged data set.

With the increasing percentage penetration of wind energy, accurate Wind Speed/Power Forecasting (WSF/WPF) has become a crucial tool for power trading and electricity grid operations. The physical and statistical methods used to forecast mostly rely on meteorological information, site specific data and historical data from wind farms. Most of the existing works are limited to utilize single input data source and a prediction variable. However, some gap in using multiple data sources and multiple prediction variables as input has been identified to substantiate and enhance forecast accuracy of these models. In this paper, a multi-source and multivariate surveillance approach has been conducted using time series methods and machine learning algorithms. Firstly, the proposed model uses multiple sources to understand the deciding prediction variables. Secondly, an analogy among time series Autoregressive Integrated Moving Average (ARIMA), Recurrent Neural Networks (RNNs) with Long Short-term Memory units (LSTMs) is used. Although, the prediction models using single data source are economical for forecast providers and wind power producers considering multiple sources can certainly improve the forecasting accuracy thus reducing wind curtailments and imbalance penalties. A comparative approach has been made using data sources obtained for Gulf of Khambhat, a south to north penetration of the Arabian Sea on the western shelf of India. The obtained results show that the forecasting accuracy can be improved by considering multiple sources of input data.

In this paper, we focus on the visualization of heterogeneous semantic networks obtained from multiple data sources. A semantic network comprising a set of entities and relationships is often used for representing knowledge derived from textual data or database records. Although the semantic networks created for the same domain at different data sources may cover a similar set of entities, these networks could also be very different because of naming conventions, coverage, view points, and other reasons. Since digital libraries often contain data from multiple sources, we propose a visualization tool to integrate and analyze the differences among multiple social networks. Through a case study on two terrorism-related semantic networks derived from Wikipedia and Terrorism Knowledge Base (TKB) respectively, the effectiveness of our proposed visualization tool is demonstrated.

Detecting functional modules from protein-protein interaction (PPI) networks is an active research area with many practical applications. However, there is always a critical concern on the false PPI interactions which are derived from the high-throughput experiments and the unsatisfactory results obtained from single PPI network with severe information insufficiency. To address this problem, we propose a Collective Non-negative Matrix Factorization (CoNMF) based soft clustering method which efficiently integrates information of gene ontology (GO), gene expression data and PPI networks. In our method, the three data sources are formed into two graphs with similarity adjacency matrices and these graphs are approximated by a matrix factorization with their common factor which provides the straight-forward interpretation of clustering results. Extensive experiments show that we can improve the module detection performance by integrating multiple biological data sources and that CoNMF yields superior results compared to other multiple data sources fusion methods by identifying a larger number of more precise protein modules with actual biological meaning and certain degree of overlapping.

Entity alignment is the problem of identifying which entities in a data source refer to the same real-world entity in the others. Identifying entities across heterogeneous data sources is paramount to many research fields, such as data cleaning, data integration, information retrieval and machine learning. The aligning process is not only overwhelmingly expensive for large data sources since it involves all tuples from two or more data sources, but also need to handle heterogeneous entity attributes. In this paper, we propose an unsupervised approach, called EnAli, to match entities across two or more heterogeneous data sources. EnAli employs a generative probabilistic model to incorporate the heterogeneous entity attributes via employing exponential family, handle missing values, and also utilize the locality sensitive hashing schema to reduce the candidate tuples and speed up the aligning process. EnAli is highly accurate and efficient even without any ground-truth tuples. We illustrate the performance of EnAli on re-identifying entities from the same data source, as well as aligning entities across three real data sources. Our experimental results manifest that our proposed approach outperforms the comparable baseline.

Stimulated by the arrival of the big data era, various and heterogeneous data sources such as data in social networks, mobile devices and sensor data for users have emerged, mirroring characteristics and preferences of data owners. These data sources are often used to construct user profiles so as to facilitate personalized services like recommendations or personalized data access. In the context of second language learning, learner data involve learning logs, standard test results, and individual learning preferences and styles. Given its attribute of reflecting the characteristics of learners, such data can be exploited to build the learner profiles. However, these data sources possibly include noises or bias, and hence influence the reliability of the correspondingly constructed learner profiles. Consequently, the inaccurate profiles may result in ineffective learning tasks that are generated by e-Learning systems. To tackle this issue, it is significant and critical to evaluate the accuracy of learner profiles. In a response to this call, we propose a novel metric named “profile mean square error” to examine the accuracy of learner profiles founded upon diverse sources. We also demonstrate how to construct various learner profiles though applying different data sources such as learning logs, standard test results, and personal learning preferences in e-Learning systems and pedagogical activities. Moreover, we conduct an experimental study among some second language learners, the results of which illustrate that the most accurate profiles are generated from multiple data sources if they are integrated in a rational way.

ABSTRACT
 ObjectivesThe objective was to deliver a platform to accelerate collaborative research into the genetics of dementia. The project is part of a larger effort to establish a platform for research incorporating components related to epidemiological data, imaging, wearable technology and tissue banking with initiatives to link these different data sets.
 The genomics platform aimed to address a number of challenges encountered by researchers:
 
 Enabling patient-level linkage between genomic and other datasets
 Selection of study cohorts using genotypic and phenotypic characteristics
 Collaborating with other groups to perform large-scale meta analysis
 Supporting repeatable analysis workflows
 
 ApproachThe genomics platform was developed using a mixture of open source, commercial and bespoke components developed for this project. It is designed to securely handle data and scale to cope with increasing data volumes as well as the quantity and complexity of research undertaken.
 It is located in a data centre where researchers can access, explore and analyse the data in a secure environment addressing data security and privacy concerns.
 The platform integrates with a supercomputer allowing complex analysis of data to be undertaken easily using novel code or predefined workflows.
 The platform was designed and developed through close collaboration with a prominent academic research team.
 ResultsWe have created a collaborative genomics informatics platform that provides efficient and intuitive selection of study cohorts from multiple data sources, with potentially varying data formats and types, based on both genotypic and phenotypic characteristics. Researchers can combine cohorts definitions using set operations and publish these definitions allowing other researchers to apply them to other datasets.
 There is an internet accessible portal which allows researchers to share results of their analysis and support meta analysis of this research. A sophisticated search engine allows these results to be found based on genetic information or annotations. This portal accelerates meta-analysis, an intrinsically collaborative endeavour in genomics research.
 There is also close integration with HPC (High Performance Computing) resources for computationally expensive tasks. The platform is underpinned by analytical workflows that not only accelerate and enable reproducible research, but reduce the technical barriers to its use.
 The platform enables patient-level linkage with other non-genomics data sources through integration with other research platforms.
 ConclusionThe developed genomics informatics platform provides a step-change in this type of genetic research, accelerating reproducible collaborative research across multiple disparate organisations and data sources, of varying type and complexity.

Personalized treatments and targeted therapies are the most promising approaches to treat complex diseases, especially for cancer. However, drug resistance is often acquired after treatments. To overcome or reduce drug resistance, treatments using drug combinations have been actively investigated in the literature. Existing methods mainly focus on chemical properties of drugs for potential combination therapies without considering relationships among different diseases. Also, they often do not consider the rich knowledge of drugs and diseases, which can enhance the prediction of drug combinations. This motivates us to develop a new computational method that can predict the beneficial drug combinations. We propose DrugCom, a tensor-based framework for computing drug combinations across different diseases by integrating multiple heterogeneous data sources of drugs and diseases. DrugCom first constructs a primary third-order tensor (i.e., drug×drug ×disease) and several similarity matrices from multiple data sources regarding drugs (e.g., chemical structure) and diseases (e.g., disease phenotype). DrugCom then formulates an objective function, which simultaneously factorizes coupled tensor and matrices to reveal the molecular mechanisms of drug synergy. We adopt the alternating direction method of multipliers algorithm to effectively solve the optimization problem. Extensive experimental studies using real-world datasets demonstrate superior performance of DrugCom.

It is challenging to identify meaningful gene networks because biological interactions are often condition-specific and confounded with external factors. It is necessary to integrate multiple sources of genomic data to facilitate network inference. For example, one can jointly model expression datasets measured from multiple tissues with molecular marker data in so-called genetical genomic studies. In this paper, we propose a joint conditional Gaussian graphical model (JCGGM) that aims for modeling biological processes based on multiple sources of data. This approach is able to integrate multiple sources of information by adopting conditional models combined with joint sparsity regularization. We apply our approach to a real dataset measuring gene expression in four tissues (kidney, liver, heart, and fat) from recombinant inbred rats. Our approach reveals that the liver tissue has the highest level of tissue-specific gene regulations among genes involved in insulin responsive facilitative sugar transporter mediated glucose transport pathway, followed by heart and fat tissues, and this finding can only be attained from our JCGGM approach.

MultiDS-MDA: Integrating multiple data sources into heterogeneous network for predicting novel metabolite-drug associations

The growth of available information sources in the industry requires new efficient methods for data access by end-users whose ability to analyse data is at the core of making business decisions. Data analysis suffer from end-users' ability to efficiently clean and integrate and query data spread across a multitude of different data sources. Current centralised approaches, where an IT expert translates the requirements of domain experts into Extract-Transform-Load (ETL) processes to integrate the data and to apply predefined analytical reporting tools, are too heavy-weight and inflexible. To support interactive and efficient data exploration directly by domain experts, without IT expert involvement, a more flexible and modular information systems architecture is needed. The Optique platform reduces the time and cost of data access dramatically by automating the process of translating an information requirement into the retrieval of the relevant data. The key to this automated translation is Ontology-Based Data Access (OBDA), where the central idea is to use an ontology, a semantically rich conceptual model of the problem domain, and mappings, that describe the relationship between the ontology and the data sources, to allow users to formulate queries and receive answers in a domain-centric and intelligible form, without any knowledge of the implementational details of the data sources. We present the Optique platform as a next generation OBDA system featuring user-friendly query formulation interface(s), automated query translation over multiple sources using novel ontology and mapping management components, and distributed query optimisation and execution exploiting cloud resources for scale-out. At Statoil Exploration, the Optique platform integrates multiple large and complex data sources from different vendors, and allows G&G experts to execute and present detailed information needs in the subsurface domain directly into existing expert client tools and also into web portals for easy information distribution of central assets across the enterprise.

ABSTRACTObjective: Traffic incidents occurring on roadways require the coordinated effort of multiple responder and recovery entities, including communications, law enforcement, fire and rescue, emergency medical services, hazardous materials, transportation agencies, and towing and recovery. The objectives of this study were to (1) identify and characterize transportation incident management (TIM)-related occupational fatalities; (2) assess concordance of surveillance data sources in identifying TIM occupations, driver vs. pedestrian status, and occupational fatality incident location; and (3) determine and compare U.S. occupational fatality rates for TIM industries.Methods: The Kentucky Fatality Assessment and Control Evaluation (FACE) program analyzed 2005–2016 TIM occupational fatality data using multiple data sources: death certificate data, Collision Report Analysis for Safer Highways (CRASH) data, and media reports, among others. Literal text analysis was performed on FACE data, and a multiple linear regression model and SAS proc sgpanel were used to estimate and visualize the U.S. TIM occupational mortality trend lines and confidence bounds.Results: There were 29 TIM fatalities from 2005 to 2015 in Kentucky; 41% of decedents were in the police protection occupation, and 21% each were in the fire protection and motor vehicle towing industries. Over one half of the TIM decedents were performing work activities as pedestrians when they died. Media reports identified the majority of the occupational fatalities as TIM related (28 of 29 TIM-related deaths); the use of death certificates as the sole surveillance data source only identified 17 of the 29 deaths as TIM related, and the use of CRASH data only identified 4 of the 29 deaths as TIM related. Injury scenario text analysis showed that law enforcement vehicle pursuit, towing and recovery vehicle loading, and disabled vehicle response were particular high-risk activities that led to TIM deaths. Using U.S. data, the motor vehicle towing industry had a significantly higher risk for occupational mortality compared to the fire protection and police protection industries.Conclusions: Multiple data sources are needed to comprehensively identify TIM fatalities and to examine the circumstances surrounding TIM fatalities, because no one data source in itself was adequate and undercounted the total number of TIM fatalities. The motor vehicle towing industry, in particular, is at elevated risk for occupational mortality, and targeted mandatory TIM training for the motor vehicle towing industry should be considered. In addition, enhanced law enforcement roadside safety training during vehicle pursuit and apprehension of suspects is recommended.

An integrated dynamic ship risk model based on Bayesian Networks and Evidential Reasoning

Relatively mature integration technologies are available that make application integration tractable from an engineering perspective. However, integrating multiple heterogeneous data sources into applications is still a time-consuming, costly and error-prone engineering task, because existing integration technologies are based on static integration architectures. Such architectures tightly couple a data source to the data integration infrastructure. This paper describes the architecture and evolving prototype implementation of the Data Concierge, which provides a dynamic solution to integrate heterogeneous data sources. The core architecture and mechanisms in the Data Concierge can be used for dynamically attaching to a previously unidentified source of information, without the need for a pre-existing adapter component. In this paper, an ontology based API description mechanism supported by the Data Concierge is described, along with the architecture and prototype tools for describing the metadata necessary to facilitate dynamic integration. The use of an ontology and associated state machine to generate API calls to access data sources is illustrated using an FTP data source.