Applicazione della spettrometria di massa MALDI-TOF in microbiologia clinica

Riassunto La spettrometria di massa (MS) nata negli anni ’70 trova oggi, grazie alla tecnologia Matrix-Assisted Laser Desorption Ionization-Time of Flight (MALDI-TOF), importanti applicazioni in diversi settori: biotecnologico (per la caratterizzazione ed il controllo di qualità di proteine ricombinanti ed altre macromolecole), medico–clinico (per la diagnosi di laboratorio di malattie e per lo sviluppo di nuovi trattamenti terapeutici mirati), alimentare ed ambientale. Negli ultimi anni, questa tecnologia è diventata un potente strumento anche per la diagnosi di laboratorio in microbiologia clinica, rivoluzionando il flusso di lavoro per una rapida identificazione di batteri e funghi, sostituendo l’identificazione fenotipica convenzionale basata su saggi biochimici. Attualmente mediante MALDI-TOF MS sono possibili due diversi approcci per la caratterizzazione dei microrganismi: (1) confronto degli spettri (“mass spectra”) con banche dati contenenti profili di riferimento (“database fingerprints”) e (2) “matching” di bio-marcatori con banche dati proteomiche (“proteome database”). Recentemente, la tecnologia MALDI-TOF, oltre alla sua applicazione classica nell’identificazione di microrganismi, è stata utilizzata per individuare, indirettamente, meccanismi di resistenza agli antibiotici. Primo scopo di questo studio è stato verificare e dimostrare l’efficacia identificativa della metodica MALDI-TOF MS mediante approccio di comparazione degli spettri di differenti microrganismi di interesse medico per i quali l’identificazione risultava impossibile a causa della completa assenza o presenza limitata, di spettri di riferimento all’interno della banca dati commerciale associata allo strumento. In particolare, tale scopo è stato raggiunto per i batteri appartenenti a spirochete del genere Borrelia e Leptospira, a miceti filamentosi (dermatofiti) e protozoi (Trichomonas vaginalis). Secondo scopo di questo studio è stato valutare il secondo approccio identificativo basato sulla ricerca di specifici marcatori per differenziare parassiti intestinali di interesse medico per i quali non è disponibile una banca dati commerciale di riferimento e la sua creazione risulterebbe particolarmente difficile e complessa, a causa della complessità del materiale biologico di partenza analizzato e del terreno di coltura nei quali questi protozoi sono isolati. Terzo ed ultimo scopo di questo studio è stata la valutazione dell’applicabilità della spettrometria di massa con tecnologia MALDI-TOF per lo studio delle resistenze batteriche ai carbapenemi. In particolare, è stato messo a punto un saggio di idrolisi dei carbapenemi rilevata mediante MALDI-TOF MS in grado di determinare indirettamente la produzione di carbapenemasi in Enterobacteriaceae. L’efficacia identificativa della metodica MALDI-TOF mediante l’approccio di comparazione degli spettri è stata dimostrata in primo luogo per batteri appartenenti al genere Borrelia. La banca dati commerciale dello strumento MALDI-TOF MS in uso presso il nostro laboratorio includeva solo 3 spettri di riferimento appartenenti alle specie B. burgdorferi ss, B. spielmani e B. garinii. L’implementazione del “database” con specie diverse da quelle già presenti ha permesso di colmare le lacune identificative dovute alla mancanza di spettri di riferimento di alcune tra le specie di Borrelia più diffuse in Europa (B. afzelii) e nel mondo (come ad esempio B. hermsii, e B. japonica). Inoltre l’implementazione con spettri derivanti da ceppi di riferimento di specie già presenti nel “database” ha ulteriormente migliorato l’efficacia identificativa del sistema. Come atteso, il ceppo di isolamento clinico di B. lusitaniae (specie non presente nel “database”) è stato identificato solo a livello di genere corroborando, grazie all’assenza di mis-identificazione, la robustezza della “nuova” banca dati. I risultati ottenuti analizzando i profili proteici di ceppi di Borrelia spp. di isolamento clinico, dopo integrazione del “database” commerciale, indicano che la tecnologia MALDI-TOF potrebbe essere utilizzata come rapida, economica ed affidabile alternativa ai metodi attualmente utilizzati per identificare ceppi appartenenti a questo genere. Analogamente, per il genere Leptospira dopo la creazione ex-novo della banca dati “home-made”, costruita con i 20 spettri derivati dai 20 ceppi di riferimento utilizzati, è stata ottenuta una corretta identificazione a livello di specie degli stessi ceppi ri-analizzati in un esperimento indipendente condotto in doppio cieco. Il dendrogramma costruito con i 20 MSP-Spectra implementati nella banca dati è formato da due rami principali: il primo formato dalla specie non patogena L. biflexa e dalla specie a patogenicità intermedia L. fainei ed il secondo che raggruppa insieme le specie patogene L. interrogans, L. kirschneri, L. noguchii e L. borgpetersenii. Il secondo gruppo è ulteriormente suddiviso in due rami, contenenti rispettivamente L. borgpetersenii in uno e L. interrogans, L. kirschneri e L. noguchii nell’altro. Quest’ultimo, a sua volta, è suddiviso in due rami ulteriori: il primo comprendente la sola specie L. noguchii, il secondo le specie L. interrogans e L. kirshneri non separabili tra loro. Inoltre, il dendrogramma costruito con gli MSP-Spectra dei ceppi appartenenti ai generi Borrelia e Leptospira acquisiti in questo studio, e appartenenti al genere Brachyspira (implementati in un lavoro precedentemente condotto) mostra tre gruppi principali separati tra loro, uno per ogni genere, escludendo possibili mis-identificazioni tra i 3 differenti generi di spirochete. Un’analisi più approfondita dei profili proteici ottenuti dall’analisi ha mostrato piccole differenze per ceppi della stessa specie probabilmente dovute ai diversi pattern proteici dei distinti sierotipi, come confermato dalla successiva analisi statistica, che ha evidenziato picchi sierotipo-specifici. È stato, infatti, possibile mediante la creazione di un modello statistico dedicato ottenere un “pattern” di picchi discriminanti in grado di differenziare a livello di sierotipo sia i ceppi di L. interrogans sia i ceppi di L. borgpetersenii saggiati, rispettivamente. Tuttavia, non possiamo concludere che i picchi discriminanti da noi riportati siano universalmente in grado di identificare il sierotipo dei ceppi di L. interrogans ed L. borgpetersenii; i picchi trovati, infatti, sono il risultato di un’analisi condotta su uno specifico pannello di sierotipi. È stato quindi dimostrato che attuando piccoli cambiamenti nei parametri standardizzati come l’utilizzo di un modello statistico e di un programma dedicato applicato nella routine diagnostica è possibile utilizzare la spettrometria di massa MALDI-TOF per una rapida ed economica identificazione anche a livello di sierotipo. Questo può significativamente migliorare gli approcci correntemente utilizzati per monitorare l’insorgenza di focolai epidemici e per la sorveglianza degli agenti patogeni. Analogamente a quanto dimostrato per Borrelia e Leptospira, l’implementazione della banca dati dello spettrometro di massa con spettri di riferimento di miceti filamentosi (dermatofiti) si è rilevata di particolare importanza non solo per l’identificazione di tutte le specie circolanti nella nostra area ma anche per l’identificazione di specie la cui frequenza nel nostro Paese è in aumento a causa dei flussi migratori dalla zone endemiche (M. audouinii, T. violaceum e T. sudanense). Inoltre, l’aggiornamento del “database” ha consentito di superare la mis-identificazione dei ceppi appartenenti al complesso T. mentagrophytes (T. interdigitale e T. mentagrophytes) con T. tonsurans, riscontrata prima dell’implementazione della banca dati commerciale. Il dendrogramma ottenuto dai 24 spettri implementati appartenenti a 13 specie di dermatofiti ha rivelato raggruppamenti che riflettono quelli costruiti su base filogenetica. Sulla base dei risultati ottenuti mediante sequenziamento della porzione della regione ITS del genoma fungino non è stato possibile distinguere T. interdigitale e T. mentagrophytes, conseguentemente anche gli spettri di queste due specie presentavano picchi dello stesso peso molecoalre. Da sottolineare che il dendrogramma costruito con i 12 profili proteici già inclusi nel database commerciale e con i 24 inseriti nel nuovo database non riproduce l’albero filogenetico per alcune specie del genere Tricophyton: gli spettri MSP già presenti nel database e quelli aggiunti delle specie T. interdigitale e T. mentagrophytes raggruppano separatamente. Questo potrebbe spiegare le mis-identificazioni di T. interdigitale e T. mentagrophytes con T. tonsurans ottenute prima dell’implementazione del database. L’efficacia del sistema identificativo MALDI-TOF è stata anche dimostrata per microrganismi diversi da batteri e funghi per i quali la metodica originale è stata sviluppata. Sebbene tale sistema identificativo sia stato applicato con successo a Trichomonas vaginalis è stato necessario apportare modifiche nei parametri standard previsti per l’identificazione di batteri e funghi. Le interferenze riscontrate tra i profili proteici ottenuti per i due terreni utilizzati per la coltura di questo protozoo e per i ceppi di T. vaginalis hanno, infatti, reso necessario l’utilizzo di nuovi parametri per la creazione degli spettri di riferimento (MSP-Spectra). L’importanza dello sviluppo del nuovo metodo risiede nel fatto che è possibile identificare sulla base del profilo proteico (e non sulla base di singoli marcatori) microorganismi cresciuti su terreni complessi che potrebbero presentare picchi nell'intervallo di peso molecolare utilizzato a scopo identificativo: metaboliti, pigmenti e nutrienti presenti nel terreno possono interferire con il processo di cristallizzazione e portare ad un basso punteggio identificativo. Per T. vaginalis, in particolare, la “sottrazione” di picchi dovuti a molecole riconducibili al terreno di crescita utilizzato, è stata ottenuta escludendo dall'identificazione l'intervallo di peso molecolare compreso tra 3-6 kDa, permettendo la corretta identificazione di ceppi di isolamento clinico sulla base del profilo proteico. Tuttavia, l’elevata concentrazione di parassita richiesta (105 trofozoiti/ml) per una corretta identificazione, difficilmente ottenibile in vivo, ha impedito l’identificazione di ceppi di T. vaginalis direttamente in campioni clinici. L’approccio identificativo mediante individuazione di specifici marcatori proteici (secondo approccio identificativo) è stato provato ed adottato in questo studio per l’identificazione e la differenziazione di ceppi di Entamoeba histolytica (ameba patogena) ed Entamoeba dispar (ameba non patogena), specie morfologiacamente identiche e distinguibili solo mediante saggi molecolari (PCR) aventi come bersaglio il DNA-18S, che codifica per l’RNA della subunità ribosomiale minore. Lo sviluppo di tale applicazione ha consentito di superare l’impossibilità della creazione di una banca dati dedicata, a causa della complessità del materiale fecale di partenza e del terreno di coltura impiagato per l’isolamento, e di identificare 5 picchi proteici in grado di differenziare E. histolytica da E. dispar. In particolare, l’analisi statistica ha mostrato 2 picchi specifici per E. histolytica e 3 picchi specifici per E. dispar. L’assenza dei 5 picchi discriminanti trovati per E. histolytica e E. dispar nei profili dei 3 differenti terreni di coltura utilizzati in questo studio (terreno axenico LYI-S-2 e terreno di Robinson con e senza E. coli) permettono di considerare questi picchi buoni marcatori in grado di differenziare le due specie. La corrispondenza dei picchi con il PM di due specifiche proteine di E. histolytica depositate in letteratura (Amoebapore A e un “unknown putative protein” di E. histolytica ceppo di riferimento HM-1:IMSS-A) conferma la specificità dei picchi di E. histolytica identificati mediante analisi MALDI-TOF MS. Lo stesso riscontro non è stato possibile per i picchi di E. dispar in quanto nessuna proteina del PM di interesse è presente in GenBank. Tuttavia, va ricordato che non tutte le proteine E. dispar sono state ad oggi caratterizzate e depositate in letteratura. I 5 marcatori hanno permesso di differenziare 12 dei 13 ceppi isolati da campioni di feci e cresciuti in terreno di Robinson confermando i risultati ottenuti mediante saggio di Real-Time PCR. Per un solo ceppo di isolamento clinico di E. histolytica l’identificazione, confermata mediante sequenziamento della porzione 18S-rDNA, non è stata ottenuta mediante sistema MALDI-TOF MS in quanto non sono stati trovati né i picchi corrispondenti a E. histolytica né i picchi corrispondenti a E. dispar. Per questo ceppo è possibile ipotizzare la presenza di mutazioni geno/fenotipiche a livello delle proteine individuate come marcatori specifici per E. histolytica. Per confermare questa ipotesi sarebbe necessario analizzare un numero maggiore di ceppi di isolamento clinico con analogo profilo proteico. L’analisi condotta a diversi tempi di incubazione del campione di feci positivo per E. histolytica ed E. dipar ha mostrato il ritrovamento dei 5 picchi discriminanti solo dopo 12 ore dall’inoculo del campione nel terreno iniziale di Robinson. Questo risultato suggerisce la possibile applicazione del sistema MALDI-TOF MS per identificare ceppi di isolamento clinico di E. histolytica ed E. dipar nonostante la presenza di materiale fecale che materialmente può disturbare e rendere difficile l’interpretazione dello spettro ottenuto mediante analisi MALDI-TOF MS. Infine in questo studio è stata valutata l’applicabilità della tecnologia MALDI-TOF MS come saggio fenotipico rapido per la determinazione di ceppi produttori di carbapenemasi, verificando l'avvenuta idrolisi del meropenem (carbapeneme di riferimento utilizzato in questo studio) a contatto con i ceppi di riferimento e ceppi di isolamento clinico potenzialmente produttori di carbapenemasi dopo la messa a punto di un protocollo analitico dedicato. Il saggio di idrolisi del meropenem mediante MALDI-TOF MS ha dimostrato la presenza o l’assenza indiretta di carbapenemasi nei 3 ceppi di riferimento e nei 1219 (1185 Enterobacteriaceae e 34 non-Enterobacteriaceae) ceppi di isolamento clinico inclusi nello studio. Nessuna interferenza è stata riscontrata per i ceppi di Enterobacteriaceae variamente resistenti ai tre carbapenemi ma risultati non produttori di carbapenemasi mediante i saggi fenotipici comunemente impiegati nella diagnostica routinaria di laboratorio: nessuna idrolisi del farmaco è stata infatti osservata al saggio di idrolisi mediante MALDI-TOF MS. In un solo caso (ceppo di K. pneumoniae N°1135) è stato ottenuto un profilo anomalo in quanto presenti sia i picchi del farmaco intatto che quelli del farmaco idrolizzato. Per questo ceppo resistente ai tre carbapenemi saggiati, negativo ai saggi fenotipici per la presenza di carbapenemasi, è stata dimostrata la presenza del gene blaKPC mediante Real-Time PCR. Per questo ceppo si può ipotizzare la presenza di mutazioni a carico del gene blaKPC che sebbene non interferiscano con il suo rilevamento mediante PCR (Real-Time PCR positiva), potrebbero condizionare l’attività della proteina prodotta (Saggio di Hodge modificato e Test di Sinergia negativi) riducendone la funzionalità come dimostrato, mediante analisi MALDI-TOF MS, dalla presenza dei picchi relativi sia all’idrolisi del farmaco sia dei picchi relativi al farmaco intatto. Questa ipotesi dovrebbe essere confermata mediante sequenziamento del gene blaKPC e successiva analisi strutturale della sequenza amminoacidica deducibile. L’utilizzo della tecnologia MALDI-TOF MS per la verifica dell’avvenuta idrolisi del maropenem è risultato un saggio fenotipico indiretto in grado di distinguere, al pari del test di Hodge modificato impiegato comunemente nella routine diagnostica in microbiologia, un ceppo produttore di carbapenemasi da un ceppo non produttore sia per scopi diagnostici che per la sorveglianza epidemiologica. L’impiego del MALDI-TOF MS ha mostrato, infatti, diversi vantaggi rispetto ai metodi convenzionali (Saggio di Hodge modificato e Test di Sinergia) impiegati nella routine diagnostica di laboratorio i quali richiedono personale esperto per l’interpretazione del risultato e lunghi tempi di esecuzione e di conseguenza di refertazione. La semplicità e la facilità richieste per la preparazione dei campioni e l’immediata acquisizione dei dati rendono questa tecnica un metodo accurato e rapido. Inoltre, il metodo risulta conveniente dal punto di vista economico, con un costo totale stimato di 1,00 euro per ceppo analizzato. Tutte queste considerazioni pongono questa metodologia in posizione centrale in ambito microbiologico anche nel caso del rilevamento di ceppi produttori di carbapenemasi. Indipendentemente dall’approccio identificativo utilizzato, comparato con i metodi convenzionali il MALDI-TOF MS conferisce in molti casi un guadagno in termini di tempo di lavoro tecnico (procedura pre-analititca per la preparazione dei campioni) e di tempo di ottenimento dei risultati (procedura analitica automatizzata). Questo risparmio di tempo si accentua quando sono analizzati in contemporanea un maggior numero di isolati. Inoltre, la semplicità e la facilità richieste per la preparazione dei campioni e l’immediata acquisizione dei dati rendono questo un metodo di identificazione accurato e rapido risultando più conveniente anche dal punto di vista economico, con un costo totale di 0,50 euro (materiale consumabile) per ceppo analizzato. I risultati ottenuti dimostrano che la spettrometria di massa MALDI-TOF sta diventando uno strumento importante in microbiologia clinica e sperimentale, data l’elevata efficacia identificativa, grazie alla disponibilità sia di nuove banche dati commerciali sia di aggiornamenti delle stesse da parte di diversi utenti, e la possibilità di rilevare con successo anche se in modo indiretto le antibiotico-resistenze.

The behaviour of the rare earth elements during estuarine mixing-revisited

Complete rare earth element (except Eu) and Y concentrations from the estuarine mixing zone (salinity =0.2 to 33) of Elimbah Creek, Queensland, Australia, were measured by quadrupole ICP-MS without preconcentration. High sampling density in the low salinity regime along with high quality data allow accurate tracing of the development of the typical marine rare earth element anomalies as well as Y/Ho fractionation. Over the entire estuary, the rare earth elements are strongly removed relative to a freshwater endmember (60-80% removal). This large overall removal occurs despite a strong remineralisation peak (190% for La, 130% for Y relative to the freshwater endmember) in the mid-salinity zone. Removal and remineralisation are accompanied by fractionation of the original (freshwater) rare earth element pattern, resulting in light rare earth element depletion. Estuarine fractionation generates a large positive La anomaly and a superchondritic Y/Ho ratio. Conversely, we observe no evidence to support the generation of the negative Ce anomaly in the estuary. With the exception of Ce, the typical marine rare earth element features can thus be attributed to estuarine mixing processes. The persistence of these features in hydrogenous sediments for at least 3.71 Ga highlights the importance of estuarine processes for marine chemistry on geological timescales. (c) 2005 Elsevier B.V. All rights reserved.

Investigation of Mantle Dynamics from Platinum Group Elements and Rhenium-Osmium Isotope Systematics of Mantle Xenoliths from Oahu, Hawaii

Intraplate volcanism that has created the Hawaiian-Emperor seamount chain is generally thought to be formed by a deep-seated mantle plume. While the idea of a Hawaiian plume has not met with substantial opposition, whether or not the Hawaiian plume shows any geochemical signal of receiving materials from the Earth’s Outer Core and how the plume may or may not be reacting with the overriding lithosphere remain debatable issues. In an effort to understand how the Hawaiian plume works I report on the first in-situ sulfides and bulk rock Platinum Group Element (PGE) concentrations, together with Os isotope ratios on well-characterized garnet pyroxenite xenoliths from the island of Oahu in Hawaii. The sulfides are Fe-Ni Monosulfide Solid Solution and show fractionated PGE patterns. Based on the major elements, Platinum Group Elements and experimental data I interpret the Hawaiian sulfides as an immiscible melt that separated from a melt similar to the Honolulu Volcanics (HV) alkali lavas at a pressure-temperature condition of 1530 ± 100OC and 3.1±0.6 GPa., i.e. near the base or slightly below the Pacific lithosphere. The 187Os/188Os ratios of the bulk rock vary from subchondritic to suprachondritic (0.123-0.164); and the 187Os/188Os ratio strongly correlates with major element, High Field Strength Element (HFSE), Rare Earth Element (REE) and PGE abundances. These correlations strongly suggest that PGE concentrations and Os isotope ratios reflect primary mantle processes. I interpret these correlations as the result of melt-mantle reaction at the base of the lithosphere: I suggest that the parental melt that crystallized the pyroxenites selectively picked up radiogenic Os from the grain boundary sulfides, while percolating through the Pacific lithosphere. Thus the sampled pyroxenites essentially represent crystallized melts from different stages of this melt-mantle reaction process at the base of the lithosphere. I further show that the relatively low Pt/Re ratios of the Hawaiian sulfides and the bulk rock pyroxenites suggest that, upon ageing, such pyroxenites plus their sulfides cannot generate the coupled 186Os- 187Os isotope enrichments observed in Hawaiian lavas. Therefore, recycling of mantle sulfides of pyroxenitic parentage is unlikely to explain the enriched Pt-Re-Os isotope systematics of plume-derived lavas.

Derivation of lake mixing and stratification indices from high-resolution lake buoy data

Lake Analyzer is a numerical code coupled with supporting visualization tools for determining indices of mixing and stratification that are critical to the biogeochemical cycles of lakes and reservoirs. Stability indices, including Lake Number, Wedderburn Number, Schmidt Stability, and thermocline depth are calculated according to established literature definitions and returned to the user in a time series format. The program was created for the analysis of high-frequency data collected from instrumented lake buoys, in support of the emerging field of aquatic sensor network science. Available outputs for the Lake Analyzer program are: water temperature (error-checked and/or down-sampled), wind speed (error-checked and/or down-sampled), metalimnion extent (top and bottom), thermocline depth, friction velocity, Lake Number, Wedderburn Number, Schmidt Stability, mode-1 vertical seiche period, and Brunt-Väisälä buoyancy frequency. Secondary outputs for several of these indices delineate the parent thermocline depth (seasonal thermocline) from the shallower secondary or diurnal thermocline. Lake Analyzer provides a program suite and best practices for the comparison of mixing and stratification indices in lakes across gradients of climate, hydro-physiography, and time, and enables a more detailed understanding of the resulting biogeochemical transformations at different spatial and temporal scales.

High-performance parallel systems for data-intensive computing

Thesis (Ph.D.)--University of Washington, 2016-08

Property cash flow studies : focusing on model consistency and data accuracy

The high degree of variability and inconsistency in cash flow study usage by property professionals demands improvement in knowledge and processes. Until recently limited research was being undertaken on the use of cash flow studies in property valuations but the growing acceptance of this approach for major investment valuations has resulted in renewed interest in this topic. Studies on valuation variations identify data accuracy, model consistency and bias as major concerns. In cash flow studies there are practical problems with the input data and the consistency of the models. This study will refer to the recent literature and identify the major factors in model inconsistency and data selection. A detailed case study will be used to examine the effects of changes in structure and inputs. The key variable inputs will be identified and proposals developed to improve the selection process for these key variables. The variables will be selected with the aid of sensitivity studies and alternative ways of quantifying the key variables explained. The paper recommends, with reservations, the use of probability profiles of the variables and the incorporation of this data in simulation exercises. The use of Monte Carlo simulation is demonstrated and the factors influencing the structure of the probability distributions of the key variables are outline. This study relates to ongoing research into functional performance of commercial property within an Australian Cooperative Research Centre.

Application of data mining techniques in the prediction of coronary artery disease : use of anaesthesia time-series and patient risk factor data

The high morbidity and mortality associated with atherosclerotic coronary vascular disease (CVD) and its complications are being lessened by the increased knowledge of risk factors, effective preventative measures and proven therapeutic interventions. However, significant CVD morbidity remains and sudden cardiac death continues to be a presenting feature for some subsequently diagnosed with CVD. Coronary vascular disease is also the leading cause of anaesthesia related complications. Stress electrocardiography/exercise testing is predictive of 10 year risk of CVD events and the cardiovascular variables used to score this test are monitored peri-operatively. Similar physiological time-series datasets are being subjected to data mining methods for the prediction of medical diagnoses and outcomes. This study aims to find predictors of CVD using anaesthesia time-series data and patient risk factor data. Several pre-processing and predictive data mining methods are applied to this data. Physiological time-series data related to anaesthetic procedures are subjected to pre-processing methods for removal of outliers, calculation of moving averages as well as data summarisation and data abstraction methods. Feature selection methods of both wrapper and filter types are applied to derived physiological time-series variable sets alone and to the same variables combined with risk factor variables. The ability of these methods to identify subsets of highly correlated but non-redundant variables is assessed. The major dataset is derived from the entire anaesthesia population and subsets of this population are considered to be at increased anaesthesia risk based on their need for more intensive monitoring (invasive haemodynamic monitoring and additional ECG leads). Because of the unbalanced class distribution in the data, majority class under-sampling and Kappa statistic together with misclassification rate and area under the ROC curve (AUC) are used for evaluation of models generated using different prediction algorithms. The performance based on models derived from feature reduced datasets reveal the filter method, Cfs subset evaluation, to be most consistently effective although Consistency derived subsets tended to slightly increased accuracy but markedly increased complexity. The use of misclassification rate (MR) for model performance evaluation is influenced by class distribution. This could be eliminated by consideration of the AUC or Kappa statistic as well by evaluation of subsets with under-sampled majority class. The noise and outlier removal pre-processing methods produced models with MR ranging from 10.69 to 12.62 with the lowest value being for data from which both outliers and noise were removed (MR 10.69). For the raw time-series dataset, MR is 12.34. Feature selection results in reduction in MR to 9.8 to 10.16 with time segmented summary data (dataset F) MR being 9.8 and raw time-series summary data (dataset A) being 9.92. However, for all time-series only based datasets, the complexity is high. For most pre-processing methods, Cfs could identify a subset of correlated and non-redundant variables from the time-series alone datasets but models derived from these subsets are of one leaf only. MR values are consistent with class distribution in the subset folds evaluated in the n-cross validation method. For models based on Cfs selected time-series derived and risk factor (RF) variables, the MR ranges from 8.83 to 10.36 with dataset RF_A (raw time-series data and RF) being 8.85 and dataset RF_F (time segmented time-series variables and RF) being 9.09. The models based on counts of outliers and counts of data points outside normal range (Dataset RF_E) and derived variables based on time series transformed using Symbolic Aggregate Approximation (SAX) with associated time-series pattern cluster membership (Dataset RF_ G) perform the least well with MR of 10.25 and 10.36 respectively. For coronary vascular disease prediction, nearest neighbour (NNge) and the support vector machine based method, SMO, have the highest MR of 10.1 and 10.28 while logistic regression (LR) and the decision tree (DT) method, J48, have MR of 8.85 and 9.0 respectively. DT rules are most comprehensible and clinically relevant. The predictive accuracy increase achieved by addition of risk factor variables to time-series variable based models is significant. The addition of time-series derived variables to models based on risk factor variables alone is associated with a trend to improved performance. Data mining of feature reduced, anaesthesia time-series variables together with risk factor variables can produce compact and moderately accurate models able to predict coronary vascular disease. Decision tree analysis of time-series data combined with risk factor variables yields rules which are more accurate than models based on time-series data alone. The limited additional value provided by electrocardiographic variables when compared to use of risk factors alone is similar to recent suggestions that exercise electrocardiography (exECG) under standardised conditions has limited additional diagnostic value over risk factor analysis and symptom pattern. The effect of the pre-processing used in this study had limited effect when time-series variables and risk factor variables are used as model input. In the absence of risk factor input, the use of time-series variables after outlier removal and time series variables based on physiological variable values’ being outside the accepted normal range is associated with some improvement in model performance.

High resolution tiled display walls

Scalable high-resolution tiled display walls are becoming increasingly important to decision makers and researchers because high pixel counts in combination with large screen areas facilitate content rich, simultaneous display of computer-generated visualization information and high-definition video data from multiple sources. This tutorial is designed to cater for new users as well as researchers who are currently operating tiled display walls or 'OptiPortals'. We will discuss the current and future applications of display wall technology and explore opportunities for participants to collaborate and contribute in a growing community. Multiple tutorial streams will cover both hands-on practical development, as well as policy and method design for embedding these technologies into the research process. Attendees will be able to gain an understanding of how to get started with developing similar systems themselves, in addition to becoming familiar with typical applications and large-scale visualisation techniques. Presentations in this tutorial will describe current implementations of tiled display walls that highlight the effective usage of screen real-estate with various visualization datasets, including collaborative applications such as visualcasting, classroom learning and video conferencing. A feature presentation for this tutorial will be given by Jurgen Schulze from Calit2 at the University of California, San Diego. Jurgen is an expert in scientific visualization in virtual environments, human-computer interaction, real-time volume rendering, and graphics algorithms on programmable graphics hardware.

New variational Bayesian approaches for statistical data mining : with applications to profiling and differentiating habitual consumption behaviour of customers in the wireless telecommunication industry

This thesis investigates profiling and differentiating customers through the use of statistical data mining techniques. The business application of our work centres on examining individuals’ seldomly studied yet critical consumption behaviour over an extensive time period within the context of the wireless telecommunication industry; consumption behaviour (as oppose to purchasing behaviour) is behaviour that has been performed so frequently that it become habitual and involves minimal intentions or decision making. Key variables investigated are the activity initialised timestamp and cell tower location as well as the activity type and usage quantity (e.g., voice call with duration in seconds); and the research focuses are on customers’ spatial and temporal usage behaviour. The main methodological emphasis is on the development of clustering models based on Gaussian mixture models (GMMs) which are fitted with the use of the recently developed variational Bayesian (VB) method. VB is an efficient deterministic alternative to the popular but computationally demandingMarkov chainMonte Carlo (MCMC) methods. The standard VBGMMalgorithm is extended by allowing component splitting such that it is robust to initial parameter choices and can automatically and efficiently determine the number of components. The new algorithm we propose allows more effective modelling of individuals’ highly heterogeneous and spiky spatial usage behaviour, or more generally human mobility patterns; the term spiky describes data patterns with large areas of low probability mixed with small areas of high probability. Customers are then characterised and segmented based on the fitted GMM which corresponds to how each of them uses the products/services spatially in their daily lives; this is essentially their likely lifestyle and occupational traits. Other significant research contributions include fitting GMMs using VB to circular data i.e., the temporal usage behaviour, and developing clustering algorithms suitable for high dimensional data based on the use of VB-GMM.

User behaviour modelling in a multi-dimensional environment for personalization and recommendation

Handling information overload online, from the user's point of view is a big challenge, especially when the number of websites is growing rapidly due to growth in e-commerce and other related activities. Personalization based on user needs is the key to solving the problem of information overload. Personalization methods help in identifying relevant information, which may be liked by a user. User profile and object profile are the important elements of a personalization system. When creating user and object profiles, most of the existing methods adopt two-dimensional similarity methods based on vector or matrix models in order to find inter-user and inter-object similarity. Moreover, for recommending similar objects to users, personalization systems use the users-users, items-items and users-items similarity measures. In most cases similarity measures such as Euclidian, Manhattan, cosine and many others based on vector or matrix methods are used to find the similarities. Web logs are high-dimensional datasets, consisting of multiple users, multiple searches with many attributes to each. Two-dimensional data analysis methods may often overlook latent relationships that may exist between users and items. In contrast to other studies, this thesis utilises tensors, the high-dimensional data models, to build user and object profiles and to find the inter-relationships between users-users and users-items. To create an improved personalized Web system, this thesis proposes to build three types of profiles: individual user, group users and object profiles utilising decomposition factors of tensor data models. A hybrid recommendation approach utilising group profiles (forming the basis of a collaborative filtering method) and object profiles (forming the basis of a content-based method) in conjunction with individual user profiles (forming the basis of a model based approach) is proposed for making effective recommendations. A tensor-based clustering method is proposed that utilises the outcomes of popular tensor decomposition techniques such as PARAFAC, Tucker and HOSVD to group similar instances. An individual user profile, showing the user's highest interest, is represented by the top dimension values, extracted from the component matrix obtained after tensor decomposition. A group profile, showing similar users and their highest interest, is built by clustering similar users based on tensor decomposed values. A group profile is represented by the top association rules (containing various unique object combinations) that are derived from the searches made by the users of the cluster. An object profile is created to represent similar objects clustered on the basis of their similarity of features. Depending on the category of a user (known, anonymous or frequent visitor to the website), any of the profiles or their combinations is used for making personalized recommendations. A ranking algorithm is also proposed that utilizes the personalized information to order and rank the recommendations. The proposed methodology is evaluated on data collected from a real life car website. Empirical analysis confirms the effectiveness of recommendations made by the proposed approach over other collaborative filtering and content-based recommendation approaches based on two-dimensional data analysis methods.

A comparison of two commercially available motion capture systems for gait analysis : high end vs low-cost

In 1999 Richards compared the accuracy of commercially available motion capture systems commonly used in biomechanics. Richards identified that in static tests the optical motion capture systems generally produced RMS errors of less than 1.0 mm. During dynamic tests, the RMS error increased to up to 4.2 mm in some systems. In the last 12 years motion capture systems have continued to evolve and now include high-resolution CCD or CMOS image sensors, wireless communication, and high full frame sampling frequencies. In addition to hardware advances, there have also been a number of advances in software, which includes improved calibration and tracking algorithms, real time data streaming, and the introduction of the c3d standard. These advances have allowed the system manufactures to maintain a high retail price in the name of advancement. In areas such as gait analysis and ergonomics many of the advanced features such as high resolution image sensors and high sampling frequencies are not required due to the nature of the task often investigated. Recently Natural Point introduced low cost cameras, which on face value appear to be suitable as at very least a high quality teaching tool in biomechanics and possibly even a research tool when coupled with the correct calibration and tracking software. The aim of the study was therefore to compare both the linear accuracy and quality of angular kinematics from a typical high end motion capture system and a low cost system during a simple task.

Mining a data reasoning model for personalized text classification

It is a big challenge to acquire correct user profiles for personalized text classification since users may be unsure in providing their interests. Traditional approaches to user profiling adopt machine learning (ML) to automatically discover classification knowledge from explicit user feedback in describing personal interests. However, the accuracy of ML-based methods cannot be significantly improved in many cases due to the term independence assumption and uncertainties associated with them. This paper presents a novel relevance feedback approach for personalized text classification. It basically applies data mining to discover knowledge from relevant and non-relevant text and constraints specific knowledge by reasoning rules to eliminate some conflicting information. We also developed a Dempster-Shafer (DS) approach as the means to utilise the specific knowledge to build high-quality data models for classification. The experimental results conducted on Reuters Corpus Volume 1 and TREC topics support that the proposed technique achieves encouraging performance in comparing with the state-of-the-art relevance feedback models.

Experimental studies of wireless communication and GNSS kinematic positioning performance in high-mobility vehicle environments

In order to support intelligent transportation system (ITS) road safety applications such as collision avoidance, lane departure warnings and lane keeping, Global Navigation Satellite Systems (GNSS) based vehicle positioning system has to provide lane-level (0.5 to 1 m) or even in-lane-level (0.1 to 0.3 m) accurate and reliable positioning information to vehicle users. However, current vehicle navigation systems equipped with a single frequency GPS receiver can only provide road-level accuracy at 5-10 meters. The positioning accuracy can be improved to sub-meter or higher with the augmented GNSS techniques such as Real Time Kinematic (RTK) and Precise Point Positioning (PPP) which have been traditionally used in land surveying and or in slowly moving environment. In these techniques, GNSS corrections data generated from a local or regional or global network of GNSS ground stations are broadcast to the users via various communication data links, mostly 3G cellular networks and communication satellites. This research aimed to investigate the precise positioning system performances when operating in the high mobility environments. This involves evaluation of the performances of both RTK and PPP techniques using: i) the state-of-art dual frequency GPS receiver; and ii) low-cost single frequency GNSS receiver. Additionally, this research evaluates the effectiveness of several operational strategies in reducing the load on data communication networks due to correction data transmission, which may be problematic for the future wide-area ITS services deployment. These strategies include the use of different data transmission protocols, different correction data format standards, and correction data transmission at the less-frequent interval. A series of field experiments were designed and conducted for each research task. Firstly, the performances of RTK and PPP techniques were evaluated in both static and kinematic (highway with speed exceed 80km) experiments. RTK solutions achieved the RMS precision of 0.09 to 0.2 meter accuracy in static and 0.2 to 0.3 meter in kinematic tests, while PPP reported 0.5 to 1.5 meters in static and 1 to 1.8 meter in kinematic tests by using the RTKlib software. These RMS precision values could be further improved if the better RTK and PPP algorithms are adopted. The tests results also showed that RTK may be more suitable in the lane-level accuracy vehicle positioning. The professional grade (dual frequency) and mass-market grade (single frequency) GNSS receivers were tested for their performance using RTK in static and kinematic modes. The analysis has shown that mass-market grade receivers provide the good solution continuity, although the overall positioning accuracy is worse than the professional grade receivers. In an attempt to reduce the load on data communication network, we firstly evaluate the use of different correction data format standards, namely RTCM version 2.x and RTCM version 3.0 format. A 24 hours transmission test was conducted to compare the network throughput. The results have shown that 66% of network throughput reduction can be achieved by using the newer RTCM version 3.0, comparing to the older RTCM version 2.x format. Secondly, experiments were conducted to examine the use of two data transmission protocols, TCP and UDP, for correction data transmission through the Telstra 3G cellular network. The performance of each transmission method was analysed in terms of packet transmission latency, packet dropout, packet throughput, packet retransmission rate etc. The overall network throughput and latency of UDP data transmission are 76.5% and 83.6% of TCP data transmission, while the overall accuracy of positioning solutions remains in the same level. Additionally, due to the nature of UDP transmission, it is also found that 0.17% of UDP packets were lost during the kinematic tests, but this loss doesn't lead to significant reduction of the quality of positioning results. The experimental results from the static and the kinematic field tests have also shown that the mobile network communication may be blocked for a couple of seconds, but the positioning solutions can be kept at the required accuracy level by setting of the Age of Differential. Finally, we investigate the effects of using less-frequent correction data (transmitted at 1, 5, 10, 15, 20, 30 and 60 seconds interval) on the precise positioning system. As the time interval increasing, the percentage of ambiguity fixed solutions gradually decreases, while the positioning error increases from 0.1 to 0.5 meter. The results showed the position accuracy could still be kept at the in-lane-level (0.1 to 0.3 m) when using up to 20 seconds interval correction data transmission.

Entrepreneurship education in non-business schools : best practice for Australian contexts of knowledge and innovation communities (Final Report)

In the current era of global economic instability, business and industry have already identified a widening gap between graduate skills and employability. An important element of this is the lack of entrepreneurial skills in graduates. This Teaching Fellowship investigated two sides of a story about entrepreneurial skills and their teaching. Senior players in the innovation commercialisation industry, a high profile entrepreneurial sector, were surveyed to gauge their needs and experiences of graduates they employ. International contexts of entrepreneurship education were investigated to explore how their teaching programs impart the skills of entrepreneurship. Such knowledge is an essential for the design of education programs that can deliver the entrepreneurial skills deemed important by industry for future sustainability. Two programs of entrepreneurship education are being implemented at QUT that draw on the best practice exemplars investigated during this Fellowship. The QUT Innovation Space (QIS) focuses on capturing the innovation and creativity of students, staff and others. The QIS is a physical and virtual meeting and networking space; a connected community enhancing the engagement of participants. The Q_Hatchery is still embryonic; but it is intended to be an innovation community that brings together nascent entrepreneurial businesses to collaborate, train and support each other. There is a niche between concept product and business incubator where an experiential learning environment for otherwise isolated ‘garage-at-home’ businesses could improve success rates. The QIS and the Q_Hatchery serve as living research laboratories to trial the concepts emerging from the skills survey. The survey of skills requirements of the innovation commercialisation industry has produced a large and high quality data set still being explored. Work experience as an employability factor has already emerged as an industry requirement that provides employee maturity. Exploratory factor analysis of the skills topics surveyed has led to a process-based conceptual model for teaching and learning higher-order entrepreneurial skills. Two foundational skills domains (Knowledge, Awareness) are proposed as prerequisites which allow individuals with a suite of early stage entrepreneurial and behavioural skills (Pre-leadership) to further leverage their careers into a leadership role in industry with development of skills around higher order elements of entrepreneurship, management in new business ventures and progressing winning technologies to market. The next stage of the analysis is to test the proposed model through structured equation modelling. Another factor that emerged quickly from the survey analysis broadens the generic concept of team skills currently voiced in Australian policy documents discussing the employability agenda. While there was recognition of the role of sharing, creating and using knowledge in a team-based interdisciplinary context, the adoption and adaptation of behaviours and attitudes of other team members of different disciplinary backgrounds (interprofessionalism) featured as an issue. Most undergraduates are taught and undertake teamwork in silos and, thus, seldom experience a true real-world interdisciplinary environment. Enhancing the entrepreneurial capacity of Australian industry is essential for the economic health of the country and can only be achieved by addressing the lack of entrepreneurial skills in graduates from the higher education system. This Fellowship has attempted to address this deficiency by identifying the skills requirements and providing frameworks for their teaching.

Analysis of acoustic emission data for accurate damage assessment for structural health monitoring applications

Structural health monitoring (SHM) refers to the procedure used to assess the condition of structures so that their performance can be monitored and any damage can be detected early. Early detection of damage and appropriate retrofitting will aid in preventing failure of the structure and save money spent on maintenance or replacement and ensure the structure operates safely and efficiently during its whole intended life. Though visual inspection and other techniques such as vibration based ones are available for SHM of structures such as bridges, the use of acoustic emission (AE) technique is an attractive option and is increasing in use. AE waves are high frequency stress waves generated by rapid release of energy from localised sources within a material, such as crack initiation and growth. AE technique involves recording these waves by means of sensors attached on the surface and then analysing the signals to extract information about the nature of the source. High sensitivity to crack growth, ability to locate source, passive nature (no need to supply energy from outside, but energy from damage source itself is utilised) and possibility to perform real time monitoring (detecting crack as it occurs or grows) are some of the attractive features of AE technique. In spite of these advantages, challenges still exist in using AE technique for monitoring applications, especially in the area of analysis of recorded AE data, as large volumes of data are usually generated during monitoring. The need for effective data analysis can be linked with three main aims of monitoring: (a) accurately locating the source of damage; (b) identifying and discriminating signals from different sources of acoustic emission and (c) quantifying the level of damage of AE source for severity assessment. In AE technique, the location of the emission source is usually calculated using the times of arrival and velocities of the AE signals recorded by a number of sensors. But complications arise as AE waves can travel in a structure in a number of different modes that have different velocities and frequencies. Hence, to accurately locate a source it is necessary to identify the modes recorded by the sensors. This study has proposed and tested the use of time-frequency analysis tools such as short time Fourier transform to identify the modes and the use of the velocities of these modes to achieve very accurate results. Further, this study has explored the possibility of reducing the number of sensors needed for data capture by using the velocities of modes captured by a single sensor for source localization. A major problem in practical use of AE technique is the presence of sources of AE other than crack related, such as rubbing and impacts between different components of a structure. These spurious AE signals often mask the signals from the crack activity; hence discrimination of signals to identify the sources is very important. This work developed a model that uses different signal processing tools such as cross-correlation, magnitude squared coherence and energy distribution in different frequency bands as well as modal analysis (comparing amplitudes of identified modes) for accurately differentiating signals from different simulated AE sources. Quantification tools to assess the severity of the damage sources are highly desirable in practical applications. Though different damage quantification methods have been proposed in AE technique, not all have achieved universal approval or have been approved as suitable for all situations. The b-value analysis, which involves the study of distribution of amplitudes of AE signals, and its modified form (known as improved b-value analysis), was investigated for suitability for damage quantification purposes in ductile materials such as steel. This was found to give encouraging results for analysis of data from laboratory, thereby extending the possibility of its use for real life structures. By addressing these primary issues, it is believed that this thesis has helped improve the effectiveness of AE technique for structural health monitoring of civil infrastructures such as bridges.