Decentralised Data Fusion with Parzen Density Estimates

Decentralised sensor networks typically consist of multiple processing nodes supporting one or more sensors. These nodes are interconnected via wireless communication. Practical applications of Decentralised Data Fusion have generally been restricted to using Gaussian based approaches such as the Kalman or Information Filter This paper proposes the use of Parzen window estimates as an alternate representation to perform Decentralised Data Fusion. It is required that the common information between two nodes be removed from any received estimates before local data fusion may occur Otherwise, estimates may become overconfident due to data incest. A closed form approximation to the division of two estimates is described to enable conservative assimilation of incoming information to a node in a decentralised data fusion network. A simple example of tracking a moving particle with Parzen density estimates is shown to demonstrate how this algorithm allows conservative assimilation of network information.

Rich probabilistic representations for bearing only decentralised data fusion

The aim of this paper is to demonstrate the validity of using Gaussian mixture models (GMM) for representing probabilistic distributions in a decentralised data fusion (DDF) framework. GMMs are a powerful and compact stochastic representation allowing efficient communication of feature properties in large scale decentralised sensor networks. It will be shown that GMMs provide a basis for analytical solutions to the update and prediction operations for general Bayesian filtering. Furthermore, a variant on the Covariance Intersect algorithm for Gaussian mixtures will be presented ensuring a conservative update for the fusion of correlated information between two nodes in the network. In addition, purely visual sensory data will be used to show that decentralised data fusion and tracking of non-Gaussian states observed by multiple autonomous vehicles is feasible.

Applying incremental EM to Bayesian classifiers in the learning of hyperspectral remote sensing data

In this paper, we apply the incremental EM method to Bayesian Network Classifiers to learn and interpret hyperspectral sensor data in robotic planetary missions. Hyperspectral image spectroscopy is an emerging technique for geological investigations from airborne or orbital sensors. Many spacecraft carry spectroscopic equipment as wavelengths outside the visible light in the electromagnetic spectrum give much greater information about an object. The algorithm used is an extension to the standard Expectation Maximisation (EM). The incremental method allows us to learn and interpret the data as they become available. Two Bayesian network classifiers were tested: the Naive Bayes, and the Tree-Augmented-Naive Bayes structures. Our preliminary experiments show that incremental learning with unlabelled data can improve the accuracy of the classifier.

Design of a model-generated repository as a service for USDL

SAP and its research partners have been developing a lan- guage for describing details of Services from various view- points called the Unified Service Description Language (USDL). At the time of writing, version 3.0 describes technical implementation aspects of services, as well as stakeholders, pricing, lifecycle, and availability. Work is also underway to address other business and legal aspects of services. This language is designed to be used in service portfolio management, with a repository of service descriptions being available to various stakeholders in an organisation to allow for service prioritisation, development, deployment and lifecycle management. The structure of the USDL metadata is specified using an object-oriented metamodel that conforms to UML, MOF and EMF Ecore. As such it is amenable to code gener-ation for implementations of repositories that store service description instances. Although Web services toolkits can be used to make these programming language objects available as a set of Web services, the practicalities of writing dis- tributed clients against over one hundred class definitions, containing several hundred attributes, will make for very large WSDL interfaces and highly inefficient “chatty” implementations. This paper gives the high-level design for a completely model-generated repository for any version of USDL (or any other data-only metamodel), which uses the Eclipse Modelling Framework’s Java code generation, along with several open source plugins to create a robust, transactional repository running in a Java application with a relational datastore. However, the repository exposes a generated WSDL interface at a coarse granularity, suitable for distributed client code and user-interface creation. It uses heuristics to drive code generation to bridge between the Web service and EMF granularities.

hSSB1 rapidly binds at the sites of DNA double-strand breaks and is required for the efficient recruitment of the MRN complex

hSSB1 is a newly discovered single-stranded DNA (ssDNA)-binding protein that is essential for efficient DNA double-strand break signalling through ATM. However, the mechanism by which hSSB1 functions to allow efficient signalling is unknown. Here, we show that hSSB1 is recruited rapidly to sites of double-strand DNA breaks (DSBs) in all interphase cells (G1, S and G2) independently of, CtIP, MDC1 and the MRN complex (Rad50, Mre11, NBS1). However expansion of hSSB1 from the DSB site requires the function of MRN. Strikingly, silencing of hSSB1 prevents foci formation as well as recruitment of MRN to sites of DSBs and leads to a subsequent defect in resection of DSBs as evident by defective RPA and ssDNA generation. Our data suggests that hSSB1 functions upstream of MRN to promote its recruitment at DSBs and is required for efficient resection of DSBs. These findings, together with previous work establish essential roles of hSSB1 in controlling ATM activation and activity, and subsequent DSB resection and homologous recombination (HR).

Model and prototype for querying multiple linked scientific datasets

Scientists need to transfer semantically similar queries across multiple heterogeneous linked datasets. These queries may require data from different locations and the results are not simple to combine due to differences between datasets. A query model was developed to make it simple to distribute queries across different datasets using RDF as the result format. The query model, based on the concept of publicly recognised namespaces for parts of each scientific dataset, was implemented with a configuration that includes a large number of current biological and chemical datasets. The configuration is flexible, providing the ability to transparently use both private and public datasets in any query. A prototype implementation of the model was used to resolve queries for the Bio2RDF website, including both Bio2RDF datasets and other datasets that do not follow the Bio2RDF URI conventions.

Phosphorylation of Exo1 modulates homologous recombination repair of DNA double-strand breaks

DNA double-strand break (DSB) repair via the homologous recombination pathway is a multi-stage process, which results in repair of the DSB without loss of genetic information or fidelity. One essential step in this process is the generation of extended single-stranded DNA (ssDNA) regions at the break site. This ssDNA serves to induce cell cycle checkpoints and is required for Rad51 mediated strand invasion of the sister chromatid. Here, we show that human Exonuclease 1 (Exo1) is required for the normal repair of DSBs by HR. Cells depleted of Exo1 show chromosomal instability and hypersensitivity to ionising radiation (IR) exposure. We find that Exo1 accumulates rapidly at DSBs and is required for the recruitment of RPA and Rad51 to sites of DSBs, suggesting a role for Exo1 in ssDNA generation. Interestingly, the phosphorylation of Exo1 by ATM appears to regulate the activity of Exo1 following resection, allowing optimal Rad51 loading and the completion of HR repair. These data establish a role for Exo1 in resection of DSBs in human cells, highlighting the critical requirement of Exo1 for DSB repair via HR and thus the maintenance of genomic stability.

Developing research data management services at QUT

QUT Library and the High Performance Computing and Research Support (HPC) Team have been collaborating on developing and delivering a range of research support services, including those designed to assist researchers to manage their data. QUT’s Management of Research Data policy has been available since 2010 and is complemented by the Data Management Guidelines and Checklist. QUT has partnered with the Australian Research Data Service (ANDS) on a number of projects including Seeding the Commons, Metadata Hub (with Griffith University) and the Data Capture program. The HPC Team has also been developing the QUT Research Data Repository based on the Architecta Mediaflux system and have run several pilots with faculties. Library and HPC staff have been trained in the principles of research data management and are providing a range of research data management seminars and workshops for researchers and HDR students.

An analysis of the variability in Queensland WiM data

The Queensland Department of Main Roads uses Weigh-in-Motion (WiM) devices to covertly monitor (at highway speed) axle mass, axle configurations and speed of heavy vehicles on the road network. Such data is critical for the planning and design of the road network. Some of the data appears excessively variable. The current work considers the nature, magnitude and possible causes of WiM data variability. Over fifty possible causes of variation in WiM data have been identified in the literature. Data exploration has highlighted five basic types of variability specifically: ----- • cycling, both diurnal and annual;----- • consistent but unreasonable data;----- • data jumps;----- • variations between data from opposite sides of the one road; and ----- • non-systematic variations.----- This work is part of wider research into procedures to eliminate or mitigate the influence of WiM data variability.

In vitro and in vivo examination of the cell surface glycoprotein CDCP1

A number of reports have demonstrated the importance of the CUB domaincontaining protein 1 (CDCP1) in facilitating cancer progression in animal models and the potential of this protein as a prognostic marker in several malignancies. CDCP1 facilitates metastasis formation in animal models by negatively regulating anoikis, a type of apoptosis triggered by the loss of attachment signalling from cell-cell contacts or cell-extra cellular matrix (ECM) contacts. Due to the important role CDCP1 plays in cancer progression in model systems, it is considered a potential drug target to prevent the metastatic spread of cancers. CDCP1 is a highly glycosylated 836 amino acid cell surface protein. It has structural features potentially facilitating protein-protein interactions including 14 N-glycosylation sites, three CUB-like domains, 20 cysteine residues likely to be involved in disulfide bond formation and five intracellular tyrosine residues. CDCP1 interacts with a variety of proteins including Src family kinases (SFKs) and protein kinase C ä (PKCä). Efforts to understand the mechanisms regulating these interactions have largely focussed on three CDCP1 tyrosine residues Y734, Y743 and Y762. CDCP1-Y734 is the site where SFKs phosphorylate and bind to CDCP1 and mediate subsequent phosphorylation of CDCP1-Y743 and -Y762 which leads to binding of PKCä at CDCP1-Y762. The resulting trimeric protein complex of SFK•CDCP1•PKCä has been proposed to mediate an anti-apoptotic cell phenotype in vitro, and to promote metastasis in vivo. The effect of mutation of the three tyrosines on interactions of CDCP1 with SFKs and PKCä and the consequences on cell phenotype in vitro and in vivo have not been examined. CDCP1 has a predicted molecular weight of ~90 kDa but is usually detected as a protein which migrates at ~135 kDa by Western blot analysis due to its high degree of glycosylation. A low molecular weight form of CDCP1 (LMWCDCP1) of ~70 kDa has been found in a variety of cancer cell lines. The mechanisms leading to the generation of LMW-CDCP1 in vivo are not well understood but an involvement of proteases in this process has been proposed. Serine proteases including plasmin and trypsin are able to proteolytically process CDCP1. In addition, the recombinant protease domain of the serine protease matriptase is also able to cleave the recombinant extracellular portion of CDCP1. Whether matriptase is able to proteolytically process CDCP1 on the cell surface has not been examined. Importantly, proteolytic processing of CDCP1 by trypsin leads to phosphorylation of its cell surface-retained portion which suggests that this event leads to initiation of an intracellular signalling cascade. This project aimed to further examine the biology of CDCP1 with a main of focus on exploring the roles played by CDCP1 tyrosine residues. To achieve this HeLa cells stably expressing CDCP1 or the CDCP1 tyrosine mutants Y734F, Y743F and Y762F were generated. These cell lines were used to examine: • The roles of the tyrosine residues Y734, Y743 and Y762 in mediating interactions of CDCP1 with binding proteins and to examine the effect of the stable expression on HeLa cell morphology. • The ability of the serine protease matriptase to proteolytically process cell surface CDCP1 and to examine the consequences of this event on HeLa cell phenotype and cell signalling in vitro. • The importance of these residues in processes associated with cancer progression in vitro including adhesion, proliferation and migration. • The role of these residues on metastatic phenotype in vivo and the ability of a function-blocking anti-CDCP1 antibody to inhibit metastasis in the chicken embryo chorioallantoic membrane (CAM) assay. Interestingly, biochemical experiments carried out in this study revealed that mutation of certain CDCP1 tyrosine residues impacts on interactions of this protein with binding proteins. For example, binding of SFKs as well as PKCä to CDCP1 was markedly decreased in HeLa-CDCP1-Y734F cells, and binding of PKCä was also reduced in HeLa-CDCP1-Y762F cells. In contrast, HeLa-CDCP1-Y743F cells did not display altered interactions with CDCP1 binding proteins. Importantly, observed differences in interactions of CDCP1 with binding partners impacted on basal phosphorylation of CDCP1. It was found that HeLa-CDCP1, HeLa-CDCP1-Y743F and -Y762F displayed strong basal levels of CDCP1 phosphorylation. In contrast, HeLa-CDCP1-Y734F cells did not display CDCP1 phosphorylation but exhibited constitutive phosphorylation of focal adhesion kinase (FAK) at tyrosine 861. Significantly, subsequent investigations to examine this observation suggested that CDCP1-Y734 and FAK-Y861 are competitive substrates for SFK-mediated phosphorylation. It appeared that SFK-mediated phosphorylation of CDCP1- Y734 and FAK-Y861 is an equilibrium which shifts depending on the level of CDCP1 expression in HeLa cells. This suggests that the level of CDCP1 expression may act as a regulatory mechanism allowing cells to switch from a FAK-Y861 mediated pathway to a CDCP1-Y734 mediated pathway. This is the first time that a link between SFKs, CDCP1 and FAK has been demonstrated. One of the most interesting observations from this work was that CDCP1 altered HeLa cell morphology causing an elongated and fibroblastic-like appearance. Importantly, this morphological change depended on CDCP1- Y734. In addition, it was observed that this change in cell morphology was accompanied by increased phosphorylation of SFK-Y416. This suggests that interactions of SFKs with CDCP1-Y734 increases SFK activity since SFKY416 is critical in regulating kinase activity of these proteins. The essential role of SFKs in mediating CDCP1-induced HeLa cell morphological changes was demonstrated using the SFK-selective inhibitor SU6656. This inhibitor caused reversion of HeLa-CDCP1 cell morphology to an epithelial appearance characteristic of HeLa-vector cells. Significantly, in vitro studies revealed that certain CDCP1-mediated cell phenotypes are mediated by cellular pathways dependent on CDCP1 tyrosine residues whereas others are independent of these sites. For example, CDCP1 expression caused a marked increase in HeLa cell motility that was independent of CDCP1 tyrosine residues. In contrast, CDCP1- induced decrease in HeLa cell proliferation was most prominent in HeLa- CDCP1-Y762F cells, potentially indicating a role for this site in regulating proliferation in HeLa cells. Another cellular event which was identified to require phosphorylation of a particular CDCP1 tyrosine residue is adhesion to fibronectin. It was observed that the CDCP1-mediated strong decrease in adhesion to fibronectin is mostly restored in HeLa-CDCP1-Y743F cells. This suggests a possible role for CDCP1-Y743 in causing a CDCP1-mediated decrease in adhesion. Data from in vivo experiments indicated that HeLa-CDCP1-Y734F cells are more metastic than HeLa-CDCP1 cells in vivo. This indicates that interaction of CDCP1 with SFKs and PKCä may not be required for CDCP1-mediated metastasis formation of HeLa cells in vivo. The metastatic phenotype of these cells may be caused by signalling involving FAK since HeLa-CDCP1- Y734F cells are the only CDCP1 expressing cells displaying constitutive phosphorylation of FAK-Y861. HeLa-CDCP1-Y762F cells displayed a very low metastatic ability which suggests that this CDCP1 tyrosine residue is important in mediating a pro-metastatic phenotype in HeLa cells. More detailed exploration of cellular events occurring downstream of CDCP1-Y734 and -Y762 may provide important insights into the mechanisms altering the metastatic ability of CDCP1 expressing HeLa cells. Complementing the in vivo studies, anti-CDCP1 antibodies were employed to assess whether these antibodies are able to inhibit metastasis of CDCP1 and CDCP1 tyrosine mutants expressing HeLa cells. It was found that HeLa- CDCP1-Y734F cells were the only cell line which was markedly reduced in the ability to metastasise. In contrast, the ability of HeLa-CDCP1, HeLa- CDCP1-Y743F and -Y762F cells to metastasise in vivo was not inhibited. These data suggest a possible role of interactions of CDCP1 with SFKs, occurring at CDCP1-Y734, in preventing an anti-metastatic effect of anti- CDCP1 antibodies in vivo. The proposal that SFKs may play a role in regulating anti-metastatic effects of anti-CDCP1 antibodies was supported by another experiment where differences between HeLa-CDCP1 cells and CDCP1 expressing HeLa cells (HeLa-CDCP1-S) from collaborators at the Scripps Research Institute were examined. It was found that HeLa-CDCP1-S cells express different SFKs than CDCP1 expressing HeLa cells generated for this study. This is important since HeLa-CDCP1-S cells can be inhibited in their metastatic ability using anti-CDCP1 antibodies in vivo. Importantly, these data suggest that further examinations of the roles of SFKs in facilitating anti-metastatic effects of anti-CDCP1 antibodies may give insights into how CDCP1 can be blocked to prevent metastasis in vivo. This project also explored the ability of the serine protease matriptase to proteolytically process cell surface localised CDCP1 because it is unknown whether matriptase can cleave cell surface CDCP1 as it has been reported for other proteases such as trypsin and plasmin. Furthermore, the consequences of matriptase-mediated proteolysis on cell phenotype in vitro and cell signalling were examined since recent reports suggested that proteolysis of CDCP1 leads to its phosphorylation and may initiate cell signalling and consequently alter cell phenotype. It was found that matriptase is able to proteolytically process cell surface CDCP1 at low nanomolar concentrations which suggests that cleavage of CDCP1 by matriptase may facilitate the generation of LWM-CDCP1 in vivo. To examine whether matriptase-mediated proteolysis induced cell signalling anti-phospho Erk 1/2 Western blot analysis was performed as this pathway has previously been examined to study signalling in response to proteolytic processing of cell surface proteins. It was found that matriptase-mediated proteolysis in CDCP1 expressing HeLa cells initiated intracellular signalling via Erk 1/2. Interestingly, this increase in phosphorylation of Erk 1/2 was also observed in HeLa-vector cells. This suggested that initiation of cell signalling via Erk 1/2 phosphorylation as a result of matriptase-mediated proteolysis occurs by pathways independent of CDCP1. Subsequent investigations measuring the flux of free calcium ions and by using a protease-activated receptor 2 (PAR2) agonist peptide confirmed this hypothesis. These data suggested that matriptase-mediated proteolysis results in cell signalling via a pathway induced by the activation of PAR2 rather than by CDCP1. This indicates that induction of cell signalling in HeLa cells as a consequence of matriptase-mediated proteolysis occurs via signalling pathways which do not involve phosphorylation of Erk 1/2. Consequently, it appears that future attempts should focus on the examination of cellular pathways other than Erk 1/2 to elucidate cell signalling initiated by matriptase-mediated proteolytic processing of CDCP1. The data presented in this thesis has explored in vitro and in vivo aspects of the biology of CDCP1. The observations summarised above will permit the design of future studies to more precisely determine the role of CDCP1 and its binding partners in processes relevant to cancer progression. This may contribute to further defining CDCP1 as a target for cancer treatment.

Linking ambulance, emergency department and hospital admissions data : understanding the emergency journey

Objective: to assess the accuracy of data linkage across the spectrum of emergency care in the absence of a unique patient identifier, and to use the linked data to examine service delivery outcomes in an emergency department setting. Design: automated data linkage and manual data linkage were compared to determine their relative accuracy. Data were extracted from three separate health information systems: ambulance, ED and hospital inpatients, then linked to provide information about the emergency journey of each patient. The linking was done manually through physical review of records and automatically using a data linking tool (Health Data Integration) developed by the CSIRO. Match rate and quality of the linking were compared. Setting: 10, 835 patient presentations to a large, regional teaching hospital ED over a two month period (August-September 2007). Results: comparison of the manual and automated linkage outcomes for each pair of linked datasets demonstrated a sensitivity of between 95% and 99%; a specificity of between 75% and 99%; and a positive predictive value of between 88% and 95%. Conclusions: Our results indicate that automated linking provides a sound basis for health service analysis, even in the absence of a unique patient identifier. The use of an automated linking tool yields accurate data suitable for planning and service delivery purposes and enables the data to be linked regularly to examine service delivery outcomes.

Using data mining on road asset management data in analysing road crashes

Road safety is a major concern worldwide. Road safety will improve as road conditions and their effects on crashes are continually investigated. This paper proposes to use the capability of data mining to include the greater set of road variables for all available crashes with skid resistance values across the Queensland state main road network in order to understand the relationships among crash, traffic and road variables. This paper presents a data mining based methodology for the road asset management data to find out the various road properties that contribute unduly to crashes. The models demonstrate high levels of accuracy in predicting crashes in roads when various road properties are included. This paper presents the findings of these models to show the relationships among skid resistance, crashes, crash characteristics and other road characteristics such as seal type, seal age, road type, texture depth, lane count, pavement width, rutting, speed limit, traffic rates intersections, traffic signage and road design and so on.

Road crash proneness prediction using data mining

Developing safe and sustainable road systems is a common goal in all countries. Applications to assist with road asset management and crash minimization are sought universally. This paper presents a data mining methodology using decision trees for modeling the crash proneness of road segments using available road and crash attributes. The models quantify the concept of crash proneness and demonstrate that road segments with only a few crashes have more in common with non-crash roads than roads with higher crash counts. This paper also examines ways of dealing with highly unbalanced data sets encountered in the study.

Identifying differences in wet and dry road crashes using data mining

It is commonly accepted that wet roads have higher risk of crash than dry roads; however, providing evidence to support this assumption presents some difficulty. This paper presents a data mining case study in which predictive data mining is applied to model the skid resistance and crash relationship to search for discernable differences in the probability of wet and dry road segments having crashes based on skid resistance. The models identify an increased probability of wet road segments having crashes for mid-range skid resistance values.

Comprehensive Australian Study of Entrepreneurial Emergence (CAUSEE): design, data collection and descriptive results

The Comprehensive Australian Study of Entrepreneurial Emergence (CAUSEE) is a research programme that aims to uncover the factors that initiate, hinder and facilitate the process of emergence of new economic activities and organizations. It is widely acknowledged that entrepreneurship is one of the most important forces shaping changes in a country’s economic landscape (Baumol 1968; Birch 1987; Acs 1999). An understanding of the process by which new economic activity and business entities emerge is vital (Gartner 1993; Sarasvathy 2001). An important development in the study of ‘nascent entrepreneurs’ and ‘firms in gestation’ was the Panel Study of Entrepreneurial Dynamics (PSED) (Gartner et al. 2004) and its extensions in Argentina, Canada, Greece, the Netherlands, Norway and Sweden. Yet while PSED I is an important first step towards systematically studying new venture emergence, it represents just the beginning of a stream of nascent venture studies – most notably PSED II is currently being undertaken in the US (2005– 10) (Reynolds and Curtin 2008).