Dissertationen zum Thema „Predictive functional tests“

Les cancers ovariens constituent la deuxième cause de décès par cancers gynécologiques dans le monde, principalement en raison d'un diagnostic tardif associé au développement de résistances à la chimiothérapie. Environ la moitié de ces cancers présentent des altérations dans la recombinaison homologue (RH), ce qui les rend sensibles aux inhibiteurs des protéines PARP (PARPi), impliquées dans la réparation de l'ADN. Cependant, identifier les patientes répondeuses à la chimiothérapie et sélectionner celles éligibles aux PARPi demeure un défi pour les cliniciens. Dans ce contexte, l'utilisation de tumoroïdes pour des tests fonctionnels prédictifs représente une approche prometteuse pour orienter les choix thérapeutiques en première ligne et au-delà. L'objectif de cette thèse est d'étudier la faisabilité de tests fonctionnels basés sur des tumoroïdes afin d'évaluer leur applicabilité potentielle en médecine de précision. L’établissement d'un panel de tumoroïdes issus de divers sous-types histologiques ovariens a permis de démontrer que ces modèles récapitulent les caractéristiques histologiques et moléculaires de leurs tumeurs d'origine. Suite aux test fonctionnels d’exposition directe des tumoroïdes à des traitements de 1ère et 2nde ligne, nous avons pu montrer que ces modèles présentent des réponses hétérogènes aux traitements et notamment que les modèles tumoroïdes identifiés par le test prédictif comme sensibles au carboplatine provenaient principalement de patientes répondeuses. Parallèlement, nous avons étudié les résultats d'un test fonctionnel évaluant le statut RH, le test RECAP et avons démontré que ce test était complémentaire à la méthode actuelle de détermination du statut RH, qui repose sur des techniques de séquençage NGS. Bien que des investigations à plus grande échelle soient nécessaires pour confirmer le potentiel des tumoroïdes, ces résultats fournissent des arguments supplémentaires en faveur de l'utilisation des tumoroïdes ovariens dans un contexte de médecine de précision
Ovarian cancers are the second leading cause of death from gynecological cancers worldwide, primarily due to late diagnosis combined with the development of resistance to chemotherapy. Approximately half of these cancers exhibit alterations in homologous recombination (HR), making them sensitive to PARP protein inhibitors (PARPi), which are involved in DNA repair. However, identifying patients who respond to chemotherapy and selecting those eligible for PARPi remains a challenge for clinicians. In this context, the use of patient-derived tumor organoids (PDTO) for predictive functional testing represents a promising approach to guide therapeutic choices in first-line treatment and beyond. The aim of this thesis is to study the feasibility of functional tests based on PDTO to evaluate their potential applicability in precision medicine. Establishing a panel of PDTO derived from various ovarian histological subtypes has demonstrated that these models recapitulate the histological and molecular characteristics of their tumors of origin. Following direct exposure functional tests of the tumor organoids to first- and second-line treatments, we showed that these models exhibit heterogeneous responses to treatments, and particularly that PDTO identified by the predictive test as sensitive to carboplatin mainly originated from responding patients. Additionally, we investigated the results of a functional test assessing HR status, the RECAP test, and demonstrated that this test is complementary to the current method for determining HR status, which relies on NGS sequencing techniques. Although larger-scale investigations are needed to confirm the potential of tumor organoids, these results provide further support for the use of ovarian tumor organoids in the context of precision medicine

BACKGROUND: Chronic ankle instability (CAI) patients have varying levels of mechanical and sensorimotor impairments that may lead to disparate functional movement patterns. Current literature on landing biomechanics in a CAI population, however, considers all patients as a homogeneous group. In our prior work, we identified 6 subgroups of movement patterns using lower extremity kinematics during a landing/cutting task and that showed promise in furthering understanding of movement patterns in a laboratory-based environment. To increase the utility of this methodology in clinical settings, there is a need to find easily administered clinical tests that can help identify multiple subgroups of movement patterns in a CAI population. The purpose of the present study was to identify clinical tests that would help identify frontal and sagittal kinematic movement pattern subgroups during a landing/cutting task. We hypothesized that clinical tests would help predict group assignment; which CAI patient is assigned to frontal and sagittal kinematic movement pattern subgroups, respectively. METHODS: We recruited 100 CAI patients from a university population. We used three-dimensional instrumented motion analysis to capture ankle, knee and hip kinematics as subjects performed a single-leg maximal jump landing/cutting task. We used sagittal and frontal joint angle waveforms to group CAI patients. We then used 12 demographic and clinical measures to predict these subgroups of CAI. These consisted of gender, Star Excursion Balance Test-Anterior (SEBT-ANT), Biodex static balance, figure 8 hop, triple crossover hop, dorsiflexion range of motion (DFROM), number of failed trials, body mass index, a score of Foot and Ankle Ability Measure-Activities of Daily Living (FAAM-ADL), a score of FAAM-Sports, number of "yes" responses on Modified Ankle Instability Index, and number of previous ankle sprains. First, we used functional principal component analysis to create representative curves for each CAI patient and plane from the 3 lower extremity joint angles. We then used these curves as inputs to a predictor-dependent product partition model to cluster each CAI patient to unique subgroups. Finally, we used a multinomial prediction model to examine the accuracy of predicting group membership from demographic and clinical metrics. RESULTS: The predictor-dependent product partition model identified 4 frontal and 5 sagittal movement pattern subgroups. Six predictors (e.g., gender, SEBT-ANT, figure 8 hop, triple crossover hop, DFROM, and FAAM-ADL) predicted group membership with 55.7% accuracy for frontal subgroups. Ten predictors (minus Biodex static balance and number of previous ankle sprains) predicted group membership with 59% accuracy for sagittal subgroups. CONCLUSION: Novel statistical analyses allowed us to predict group membership for multiple frontal and sagittal kinematic movement patterns during landing/cutting using a series of clinical predictors. However, due to relatively lower accuracy (56–59% accuracy), the clinical utility of the current prediction model may be limited. Future work should consider including other clinical predictors to maximize prediction accuracy for identifying multiple kinematic movement patterns during a landing/cutting task.

In recent years, the main end point of immunosuppressive therapy after liver transplantation has moved from the prevention of acute cellular rejection (ACR) toward the preservation of long-term graft function and prevention of immunosuppression-related side effects. This approach requires an optimal management of immunosuppressive therapy according to patient risk factors. However, the concentration of immunosuppressive drugs in the serum of patients, which is generally used as a surrogate for the level of immunosuppression, does not provide information about the magnitude of suppression of the immune system. Therefore a reliable marker for the development of ACR, or to predict patients who could tolerate reduced immunosuppression, would be crucial for improving post-transplant management of liver transplanted patients. The aims of the studies presented in this thesis were: 1) to assess the incidence of ACR after liver transplantation, to identify potential risk factors for ACR, and to evaluate the impact of ACR and its histological severity on outcomes; 2) to evaluate the role of liver function tests and blood eosinophil count as potential biomarkers for ACR after liver transplantation, with special attention on prediction of histologically proven moderate and severe ACR; 3) to evaluate the expression of specific immunological markers for ACR in patients before and after liver transplantation. The results of the studies showed that patient and graft survival at 1, 5 and 10 years after liver transplantation were not different with respect to presence or absence of ACR. Only untreated moderate/severe ACR was associated with increased death/graft loss using adjusted Cox regression analysis, whereas mild ACR, whether treated or not, had no effect. With regards to the evaluation of potential markers of ACR, despite peripheral eosinophilia was not sufficiently predictive of moderate/severe ACR, the delta in eosinophil count between the first and second biopsies was the only independent predictor of histological improvement, irrespective of whether bolus steroids were used. Lastly, we demonstrated that the increased expression of C28 and C38 on both CD4+ and CD8+ T cells and the increased levels of IL-17. These alterations of immune system could be used routinely in clinical practice to assess the immune status of liver transplanted patients and to properly manage immunosuppressive therapy
Lo scopo principale della terapia immunosoppressiva dopo trapianto di fegato è passato dalla prevenzione del rigetto acuto alla preservazione della funzionalità a lungo termine dell’organo trapiantato e alla prevenzione degli effetti collaterali dovuti alla terapia immunosoppressiva. Per perseguire tale scopo è necessaria una gestione ottimale della terapia immunosoppresiva stessa. Tuttavia, la misurazione dei livelli ematici dei farmaci immunosoppressori, generalmente utilizzati come surrogato dei livelli di immunosoppressione, non fornisce informazioni relative alla reale intensità della soppressione del sistema immunitario. Pertanto l’individuazione di marcatori biologici di rigetto acuto e/o di tolleranza risulta fondamentale per poter migliorare la gestione della terapia immunosoppressiva dopo-trapianto di fegato. Gli scopi degli studi riportati in questa tesi sono: 1) determinare l’incidenza e gli eventuali fattori di rischio di rigetto acuto dopo trapianto di fegato, valutare in che l’influenza del rigetto acuto e della sua severità istologica sulla sopravvivenza dell’organo e del paziente dopo trapianto di fegato; 2) valutare il ruolo degli indici di funzionalità epatica e della conta eosinofilica ematica come potenziali marcatori biologici di rigetto acuto dopo trapianto di fegato, in particolare di grado moderato/severo; 3) valutare, prima e dopo trapianto di fegato l’espressione di specifici marcatori immunologici di rigetto acuto. I risultati degli studi condotti hanno evidenziato come pazienti con diagnosi di rigetto acuto alla biopsia di protocollo presentino una sopravvivenza di organo e paziente, a 1, 5 e 10 anni dal trapianto di fegato, del tutto sovrapponibile a quella di pazienti senza evidenza istologica di rigetto acuto alla biopsia di protocollo. L’insorgenza di rigetto acuto di grado moderato/severo non sottoposto a trattamento farmacologico è tuttavia associata ad aumentata incidenza di decesso o perdita dell’organo post-trapianto. Nel valutare potenziali marcatori biologici di rigetto acuto, abbiamo dimostrato che nonostante la conta eosinofilica periferica non sia sufficientemente predittiva per lo sviluppo di rigetto acuto post-trapianto, la differenza nella conta eosinofilica tra la prima e la seconda biopsia epatica può essere considerato un fattore predittivo di miglioramento istologico, indipendentemente dall’utilizzo o meno di terapia con boli steroidei. Non è stata invece evidenziata alcuna associazione tra l’alterazione degli indici di funzionalità epatica e l’insorgenza di rigetto acuto. Infine, è stato dimostrato che l’insorgenza di rigetto acuto risulta associata ad aumentata espressione di CD28 e CD38 sia sui linfociti T CD4+ che CD8+ e ad un aumento dei livelli di IL-17. Tali alterazioni del sistema immunitario potrebbero essere utilizzate nella pratica clinica per valutare lo stato di soppressione del sistema immunitario in pazienti sottoposti a trapianto di fegato con il fine ultimo di una gestione ottimale e personalizzata della terapia immunooppressiva

In this dissertation, we introduce three projects in machine learning and reliability applications after the general introductions in Chapter 1. The first project concentrates on the multivariate sensory data, the second project is related to the bivariate recurrent process, and the third project introduces thermal index (TI) estimation in accelerated destructive degradation test (ADDT) data, in which an R package is developed. All three projects are related to and can be used to solve certain reliability problems. Specifically, in Chapter 2, we introduce a clustering method for multivariate functional data. In order to cluster the customized events extracted from multivariate functional data, we apply the functional principal component analysis (FPCA), and use a model based clustering method on a transformed matrix. A penalty term is imposed on the likelihood so that variable selection is performed automatically. In Chapter 3, we propose a covariate-adjusted model to predict next event in a bivariate recurrent event system. Inspired by geyser eruptions in Yellowstone National Park, we consider two event types and model their event gap time relationship. External systematic conditions are taken account into the model with covariates. The proposed covariate adjusted recurrent process (CARP) model is applied to the Yellowstone National Park geyser data. In Chapter 4, we compare estimation methods for TI. In ADDT, TI is an important index indicating the reliability of materials, when the accelerating variable is temperature. Three methods are introduced in TI estimations, which are least-squares method, parametric model and semi-parametric model. An R package is implemented for all three methods. Applications of R functions are introduced in Chapter 5 with publicly available ADDT datasets. Chapter 6 includes conclusions and areas for future works.
Doctor of Philosophy

Les données issues de la recherche en neurosciences permettent de considérer la dépression comme une affection invalidante générale, caractérisée par des déficits neurocognitifs et comportementaux, au delà des symptômes dépressifs cliniques définis dans les classifications nosographiques. Ces déficits coexistent à la phase aiguë d’un épisode dépressif caractérisé (EDC) et interfèrent dans la prise de décision et réalisation d’un comportement orienté vers un but, et la sensation d’effort associée. Ils semblent persister en période de rémission clinique, altérant la qualité de la réponse thérapeutique et fonctionnelle et aggravant à terme le pronostic du trouble. L’objectif de ce travail est d’identifier des marqueurs neurocognitifs objectivement mesurables en pratique clinique, et d’étudier leur association au pronostic d’un EDC, afin de mieux prédire les probabilités de rémission et d’optimiser les stratégies de prescription thérapeutique des patients. L’altération des processus neurocognitifs liés à la récompense constitue un premier marqueur de vulnérabilité du trouble dépressif : dans une étude explorant la production d’un effort moteur dans le but d’obtenir une récompense, les patients déprimés présentaient un déficit de production d’effort, à la différence des sujets sains. Ce trouble de la motivation par incitation - processus sous tendu en imagerie fonctionnelle chez le sujet sain par l’activation de circuits cortico-striataux ventraux -pourrait constituer une dimension spécifique de la maladie dépressive. Participant à l’altération des processus de prise de décision et d’action, ce déficit motivationnel est associé, et possiblement secondaire, à des déficits plus spécifiquement cognitifs que nous avons ensuite étudiés. Dans une étude explorant plusieurs fonctions cognitives chez des patients déprimés au sein d’une large cohorte, la présence d’un ralentissement psychomoteur séquellaire après 6 à 8 semaines de traitement – chez des patients pourtant en rémission clinique -, était positivement et de manière indépendante, significativement corrélée au nombre d’épisodes dépressifs passés, constituant ainsi un marqueur d’une sévérité « cumulative » du trouble dépressif. Enfin, dans une revue de la littérature sur le caractère progressivement évolutif des déficits cognitifs dans le trouble dépressif unipolaire, nous avons discuté l’existence d’un effet « neurotoxique » cérébral de l’accumulation d’EDC, à l’origine de troubles neurocognitifs et de conséquences sur le cours évolutif de la maladie (risque majoré de rémission clinique et/ou fonctionnelle partielle, de récurrence, d’évolution démentielle). Un des principaux intérêts de l’identification de marqueurs de vulnérabilité cliniques et cognitifs est de mettre en évidence leur rôle prédictif du cours évolutif d’un épisode -ou d’un trouble- dépressif. Dans une étude menée sur une cohorte de plus de 500 patients déprimés, une variable attentionnelle (d2 test d’attention) était capable de prédire l’évolution ultérieure vers la rémission complète (clinique et fonctionnelle) de façon significative, linéaire, et indépendante des autres variables et de représenter un marqueur-trait de la dépression, aisément utilisable en pratique clinique. D’autres marqueurs cognitifs (tels que les fonctions exécutives) ont montré une valeur prédictive élevée de la réponse thérapeutique, avec une précision proche de celle de marqueurs d’imagerie ou électrophysiologie, selon les résultats d’une méta-analyse récente, justifiant leur emploi dans le suivi des patients
Hese deficits coexist in the acute phase of a depressive episode and interfere with decision-making and goal-directed behaviors, and the associated feeling of effort. They appear to persist in periods of clinical remission, decreasing the quality of the therapeutic and functional response and lately worsening the prognosis of the disorder. The aim of this work is to identify objectively measurable neurocognitive markers in clinical practice, and to study their association with the prognosis of a depressive episode, in order to better predict remission and potentially to optimize therapeutic prescribing strategies for patients accordingly. The impairment of neurocognitive processes related to reward constitutes a first vulnerability marker for major depressive disorder (MDD): in a study assessing the production of motor effort in order to obtain a reward, depressed patients had a deficit in production of effort, unlike healthy subjects. Such deficit in incentive motivation - a process underpinned by the activation of ventral cortico-striatal circuits in healthy subjects - may constitute a specific dimension of MDD. It participates in the decision-making and action processes impairments and is associated with – and possibly a consequence of- more specifically cognitive deficits. In a study assessing several cognitive functions in a large cohort of depressed patients, the persistence of psychomotor retardation after 6 to 8 weeks of treatment - in patients considered as being in clinical remission - was positively and independently correlated with the number of past depressive episodes, thus constituting a marker of "cumulative" marker of past depressive episodes. Finally, in a literature review on the progressive evolution of cognitive deficits in MDD, we discussed the existence of a “neurotoxic” effect of the lifetime accumulation of depressive episodes on neurocognitive deficits and its consequences on disease prognosis (increased risk of incomplete functional/clinical remission, relapses, evolution towards dementia). One of the main interest in identifying clinical and cognitive markers of vulnerability is to highlight their capacity to predict the course of a depressive episode-or disorder. In a study based on a cohort of more than 500 depressed patients, a measurement of attention (d2 attention test) was able to significantly and independently predict the subsequent course towards complete remission (clinical and functional) and to constitute a trait -marker of depression, easy to use in clinical practice. Other cognitive markers (such as executive functions) have shown high predictive values for therapeutic response, comparable to those provided by imaging or electrophysiology markers, according to the results of a recent meta-analysis, that emphasizes the interest of using them in patient’s follow-up. Finally, in order to better assess the prognosis of depressive disorder, we have shown that Seasonal Affective Disorder (SAD) diagnosis criteria - which nevertheless represents a specific depressive disorder with well-known physiopathology substrates (construction validity) - had a low predictive validity, prompting to consider this disorder as a temporary expression of a mood disorder, rather than a specific disorder. The identification of clinical tools measuring motivational and cognitive deficits in clinical routine and predicting the course of a depressive episode or disorder represents a major challenge in the improvement of personalized therapeutic management and the long-term prognosis in depressed patients

Anterior cruciate ligament (ACL) ruptures are one of the most common knee ligament injuries suffered by both male and female athletes. These injuries are severe in nature and also have long-term impacts on activities of daily living. Significant research has been conducted utilizing a drop landing task to attempt to better understand the mechanics behind the injury and to help identify at-risk athletes for targeted intervention. However, there have not been any published standards for the height of the drop landing activity, and previous researchers have also raised some concerns about the ability of a drop landing task to replicate the landing mechanics of a sport-specific task. To examine possible differences in performance based on specific landing tasks, the first study compared the landing mechanics of male and female high school athletes in three different landing conditions (drop landing, DL; adjusted height drop landing, AHDL; and a vertical jump task, VJL) (Chapter 3). Thirty-seven (37) athletes completed bilateral landings in the three conditions, and their kinetic and kinematic landing mechanics were compared across conditions. For the male participants, maximum knee flexion during landing was greater in AHDL condition as compared to the DL and VJL conditions. Both male and female participants demonstrated greater hip adduction at impact and overall maximum value in the VJL condition as compared to the two drop landings. As drop landing tasks have been used to identify at-risk athletes, it was important to examine the three different tasks’ ability to predict lower extremity ligamentous injuries, and whether those 3D motion analysis predictors were more precise than a quick clinical symmetry screening tool (Chapter 4). One-hundred-and-sixty-five (165) athletes completed the clinical symmetry screen, and a subgroup of thirty-seven (37) athletes completed the 3D motion analysis. All of these participants were surveyed for lower extremity ligamentous injuries over the course of a season. Due to a small number of reported injuries, none of the injury predictor models based on 3D motion analysis landing mechanics or the clinical symmetry screening tool were able to produce accurate predictor models of injury. Knee abduction moment has been shown to be one of the strongest predictors of ACL injuries, and due to the collection of bilateral kinetics for a previous study (Chapter 3), there was a need to examine differences in KAM between the three different landing tasks (Chapter 5). Ten (10) recreational athletes completed bilateral landings in the three conditions, with foot placement relative to force plates to enable KAM calculation. The participants did not demonstrate any difference in KAM between the three landing conditions; however, a test for constant variance showed that the AHDL resulted in significantly less variance in KAM than DL or VJL. The results of these studies suggest that while easy to standardize, a set height drop landing task does not produce identical landing mechanics to those from an adjusted height drop landing task or a vertical jump task. Further research is needed to create or justify standardized landing tasks for researchers to utilize that produce consistent results that best duplicate the landing mechanics athletes performed during sporting activities. While the landing mechanics demonstrated in the three tasks and the results from the clinical screening were not able to predict injuries, future studies should examine quick clinical screening tools to identify athletes at a high risk of injury.

Hronična opstruktivna bolest pluća je jedna od najčešćih hroničnih bolesti pluća i važan uzrok morbiditeta i mortaliteta u svetu. Egzacerbacije predstavljaju značajan događaj u toku bolesti, jer imaju negativan uticaj na mortalitet, kvalitet života, opadanje plućne funkcije i povećanje troškova lečenja. Cilj rada je utvrditi nezavisne faktore rizika za egzacerbaciju i smrtni ishod tokom jednogodišnjeg praćenja obolelih od hronične opstruktivne bolesti pluća i kreiranje prediktivnog modela za neželjeni događaj. U ispitivanje je uključeno 200 pacijenata sa potvrđenom dijagnozom hronične opstruktivne bolesti pluća, koji su lečeni prema preporukama smernice Globalne inicijative za hroničnu opstruktivnu bolest pluća. Pacijenti su praćeni godinu dana, evaluirani na kontrolnim pregledima i beležen je broj egzacerbacija na osnovu vanrednih poseta i eventualni smrtni ishod. Statističkom obradom podataka utvrđeni su nezavisni prediktori egzacerbacije (starost > 65 godina, test procene hronične opstruktivne bolesti pluća > 9, modifikovana skala dispneje > 2, saturacija hemoglobina kiseonikom ≤ 93%) i smrtnog ishoda (starost >65 godina, potreba za primenom antiagregacione terapije, brzina maksimalnog ekspiratornog protoka na 25% vitalnog kapaciteta ≤ 1,16 l, modifikovana skala dispneje >2, puls > 89). Od navedenih nezavisnih faktora su kreirani modeli za predikciju neželjenih događaja. Unutrašnjom validacijom modela dokazana je dobra prediktivna vrednost oba matematička modela, bez statistički značajne razlike opserviranog i očekivanog procenta pojave egzacerbacije i smrtnog ishoda tokom praćenja pacijenata obolelih od HOBP.
Chronic obstructive pulmonary disease is one of the most common chronic lung diseases and is an important cause of morbidity and mortality in the world. Exacerbations are an important event in the course of the disease, as they have a negative impact on mortality, quality of life, lung function decline and increased costs of treatment. The aim of study is to identify risk factors for exacerbation or death during the one-year follow-up of patients with chronic obstructive pulmonary disease, and creation of predictive models for exacerbation and mortality during the follow-up period. The study included 200 patients with a confirmed diagnosis of chronic obstructive pulmonary disease who have had the therapy according to the Global initiative for chronic obstructive airway diseases guidelines. Patients were followed for one year, evaluated the number of exacerbations on the basis of emergency visits and eventual death. With statistical data processing there were identified independent predictors of exacerbations (age > 65 years, COPD Assessment Test > 9, modified Medical Research Council scale >2, oxygen saturation ≤ 93%) and death (age > 65 years, the need for application of antiplatelet therapy, the rate of maximum expiratory flow at 25% of vital capacity ≤ 1,16 l, modified Medical Research Council scale >2, heart rate > 89th). Of these independent factors was created a models for the prediction of adverse events during the one-year mark of COPD patients. Internal validation showed good predictive value of both models. No difference between the observed and the expected percentage of occurrence of exacerbations or death during the the follow-up period.

A análise de dados contendo informações sequenciais é um problema de crescente interesse devido à grande quantidade de informação que é gerada, entre outros, em processos de monitoramento. As séries temporais são um dos tipos mais comuns de dados sequenciais e consistem em observações ao longo do tempo. O algoritmo k-Nearest Neighbor - Time Series Prediction kNN-TSP é um método de previsão de dados temporais. A principal vantagem do algoritmo é a sua simplicidade, e a sua aplicabilidade na análise de séries temporais não-lineares e na previsão de comportamentos sazonais. Entretanto, ainda que ele frequentemente encontre as melhores previsões para séries temporais parcialmente periódicas, várias questões relacionadas com a determinação de seus parâmetros continuam em aberto. Este trabalho, foca-se em dois desses parâmetros, relacionados com a seleção de vizinhos mais próximos e a função de previsão. Para isso, é proposta uma abordagem simples para selecionar vizinhos mais próximos que considera a similaridade e a distância temporal de modo a selecionar os padrões mais similares e mais recentes. Também é proposta uma função de previsão que tem a propriedade de manter bom desempenho na presença de padrões em níveis diferentes da série temporal. Esses parâmetros foram avaliados empiricamente utilizando várias séries temporais, inclusive caóticas, bem como séries temporais reais referentes a variáveis ambientais do reservatório de Itaipu, disponibilizadas pela Itaipu Binacional. Três variáveis limnológicas fortemente correlacionadas são consideradas nos experimentos de previsão: temperatura da água, temperatura do ar e oxigênio dissolvido. Uma análise de correlação é realizada para verificar se os dados previstos mantem a correlação das variáveis. Os resultados mostram que, o critério de seleção de vizinhos próximos e a função de previsão, propostos neste trabalho, são promissores
Treating data that contains sequential information is an important problem that arises during the data mining process. Time series constitute a popular class of sequential data, where records are indexed by time. The k-Nearest Neighbor - Time Series Prediction kNN-TSP method is an approximator for time series prediction problems. The main advantage of this approximator is its simplicity, and is often used in nonlinear time series analysis for prediction of seasonal time series. Although kNN-TSP often finds the best fit for nearly periodic time series forecasting, some problems related to how to determine its parameters still remain. In this work, we focus in two of these parameters: the determination of the nearest neighbours and the prediction function. To this end, we propose a simple approach to select the nearest neighbours, where time is indirectly taken into account by the similarity measure, and a prediction function which is not disturbed in the presence of patterns at different levels of the time series. Both parameters were empirically evaluated on several artificial time series, including chaotic time series, as well as on a real time series related to several environmental variables from the Itaipu reservoir, made available by Itaipu Binacional. Three of the most correlated limnological variables were considered in the experiments carried out on the real time series: water temperature, air temperature and dissolved oxygen. Analyses of correlation were also accomplished to verify if the predicted variables values maintain similar correlation as the original ones. Results show that both proposals, the one related to the determination of the nearest neighbours as well as the one related to the prediction function, are promising

Apresentamos duas abordagens para obter uma densidade de probabilidades para o preço futuro de um ativo: uma densidade preditiva, baseada em um modelo Bayesiano para série de tempo e uma densidade implícita, baseada na fórmula de precificação de opções de Black & Scholes. Considerando o modelo de Black & Scholes, derivamos as condições necessárias para obter a densidade implícita do preço do ativo na data de vencimento. Baseando-­se nas densidades de previsão, comparamos o modelo implícito com a abordagem histórica do modelo Bayesiano. A partir destas densidades, calculamos probabilidades de ordem e tomamos decisões de vender/comprar um ativo. Como exemplo, apresentamos como utilizar estas distribuições para construir uma fórmula de precificação.
We present two different approaches to obtain a probability density function for the stocks future price: a predictive distribution, based on a Bayesian time series model, and the implied distribution, based on Black & Scholes option pricing formula. Considering the Black & Scholes model, we derive the necessary conditions to obtain the implied distribution of the stock price on the exercise date. Based on predictive densities, we compare the market implied model (Black & Scholes) with a historical based approach (Bayesian time series model). After obtaining the density functions, it is simple to evaluate probabilities of one being bigger than the other and to make a decision of selling/buying a stock. Also, as an example, we present how to use these distributions to build an option pricing formula.

Povratak na posao nakon operacije lumbalne diskus hernije determinisan je funkcionalnim stanjem, prisustvom i stepenom tegoba od strane lumbosakralne kičme, zahtevima na radnom mestu bolesnika ali i psihosocijalnim faktorima, koji pri oceni radne sposobnosti zaposlnih zahtevaju individualni pristup. Grupa pacijenata koja se neće vratiti na posao može biti identifikovana putem prognostičkog modela. Cilj ove studije je definisanje prognostičkog modela za povratak na posao bolesnika operisanih zbog lumbalne diskus hernije kao i identifikacija najznačajnijih faktora rizika odgovornih za loš ishod operativnog lečenja, posmatrano kroz prizmu povratka na posao. Istraživanje je prospektivna studija koja je obuhvatila ukupno 200 ispitanika, koji su operisani zbog lumbalne diskus hernije na jednom nivou i praćeni su u vremenskom period do 12 meseci nakon operativnog lečenja. U statističku ananlizu je ušlo 153 bolesnika, koji su ispunili kriterijume selekcije ispitanika studije. Nakon određivanja značaja posmatranih bioloških, profesionalnih i psihosocijalnih faktora rizika za povratak na posao, kreirani su i evaluirani prognostički modeli bazirani na svim i na odabranim atributima desetostrukom kros-validacijom: stablo odlučivanja (DT), model višeslojnih perceptrona (MLP) i model potpornih vektora (SVM). Za predviđanje povratka na posao najveću tačnost, specifičnost i senzitivnost za odabrane atribute postiže model potpornih - podržavajućih vektora (SVM). Najbolju intuitivnu i praktičnu vrednost za predviđanje povratka na posao pruža model stabla odluka (DT). Identifikacijom najznačajnijih faktora rizika za nepovoljan ishod povratka na posao omogućeno je preventivno delovanje na iste, u cilju smanjenja broja pacijenata sa umanjenjem radne sposobnosti i invaliditeta.
Return to work after lumbar discectomy is determinated by functional status, presence and degree of discomfort in the lumbosacral spine, the requirements in the workplace of patients and psychosocial factors that in the assessment of working capabilities require an individual approach. Groups of patients which don’t return to work after surgery could be identified by predictive model. The aim of this study is to define prognostic model to return to work patients after lumbar discectomy, as well as the identification major risk factors responsible for the poor outcome of operative treatment viewed through the prism of returning to work. This prospective study included a total of 200 patients, who underwent surgery for lumbar disc herniation on one level and were followed up in period of 12 months following surgery. The statistical analysis included 153 patients who fulfilled all selection criteria of the study subjects. After determining significance of the observed biological, professional and psychosocial risk factors for return to work, prognostic models were designed and evaluated based on all and selected attributes by tenfold cross-validation: decision tree (DT) model of multilayer perception (MLP) model and support vector (SVM). For the prediction of return to work best accuracy, specificity and sensitivity for selected attributes, is achieved by supporting vector model (SVM). The decision tree model (DT) provides the best intuitive and practical value for predicting return to work. By identifying the most important risk factors for adverse outcome for return to work it is made possible for preventive actions, to reduce the number of patients with reduced work ability and disability.

A confiabilidade de um sistema de componentes depende da confiabilidade de cada componente. Assim, a estimação da função de confiabilidade de cada componente do sistema é de interesse. No entanto, esta não é uma tarefa fácil, pois quando o sistema falha, o tempo de falha de um dado componente pode não ser observado, isto é, um problema de dados censurados. Neste trabalho, propomos modelos Bayesianos paramétricos para estimação das funções de confiabilidade de componentes e sistemas em quatro diferentes cenários. Inicialmente, um modelo Weibull é proposto para estimar a distribuição do tempo de vida de um componente de interesse envolvido em sistemas coerentes não reparáveis, quando estão disponíveis o tempo de falha do sistema e o estado do componente no momento da falha do sistema. Não é imposta a suposição de que os tempos de vida dos componentes sejam identicamente distribuídos, mas a suposição de independência entre os tempos até a falha dos componentes é necessária, conforme teorema anunciado e devidamente demonstrado. Em situações com causa de falha mascarada, os estados dos componentes no momento da falha do sistema não são observados e, neste cenário, um modelo Weibull com variáveis latentes no processo de estimação é proposto. Os dois modelos anteriormente descritos propõem estimar marginalmente as funções de confiabilidade dos componentes quando não são disponíveis ou necessárias as informações dos demais componentes e, por consequência, a suposição de independência entre os tempos de vida dos componentes é necessária. Com o intuito de não impor esta suposição, o modelo Weibull multivariado de Hougaard é proposto para a estimação das funções de confiabilidade de componentes envolvidos em sistemas coerentes não reparáveis. Por fim, um modelo Weibull para a estimação da função de confiabilidade de componentes de um sistema em série reparável com causa de falha mascarada é proposto. Para cada cenário considerado, diferentes estudos de simulação são realizados para avaliar os modelos propostos, sempre comparando com a melhor solução encontrada na literatura até então, em que, em geral, os modelos propostos apresentam melhores resultados. Com o intuito de demonstrar a aplicabilidade dos modelos, análises de dados são realizadas com problemas reais não só da área de confiabilidade, mas também da área social.
The reliability of a system of components depends on reliability of each component. Thus, the initial statistical work should be the estimation of the reliability of each component of the system. This is not an easy task because when the system fails, the failure time of a given component can be not observed, that is, a problem of censored data. We propose parametric Bayesian models for reliability functions estimation of systems and components involved in four scenarios. First, a Weibull model is proposed to estimate component failure time distribution from non-repairable coherent systems when there are available the system failure time and the component status at the system failure moment. Furthermore, identically distributed failure times are not a required restriction. An important result is proved: without the assumption that components\' lifetimes are mutually independent, a given set of sub-reliability functions does not identify the corresponding marginal reliability function. In masked cause of failure situations, it is not possible to identify the statuses of the components at the moment of system failure and, in this second scenario, we propose a Bayesian Weibull model by means of latent variables in the estimation process. The two models described above propose to estimate marginally the reliability functions of the components when the information of the other components is not available or necessary and, consequently, the assumption of independence among the components\' failure times is necessary. In order to not impose this assumption, the Hougaard multivariate Weibull model is proposed for the estimation of the components\' reliability functions involved in non-repairable coherent systems. Finally, a Weibull model for the estimation of the reliability functions of components of a repairable series system with masked cause of failure is proposed. For each scenario, different simulation studies are carried out to evaluate the proposed models, always comparing then with the best solution found in the literature until then. In general, the proposed models present better results. In order to demonstrate the applicability of the models, data analysis are performed with real problems not only from the reliability area, but also from social area.

This study examines the validity of using a combination of two psychometric measures, an emotional adjustment measure, and functional assessment measures to predict vocational outcome in a traumatically brain injured population. Patients included 33 males and 11 females, with an average age of 32.3 years, and a stable work history over the past three years prior to injury. All had sustained a traumatic brain injury in the 12 months prior to initial testing, with a mean of 3.8 months since injury. Levels of severity of injury included 24 patients with severe injury, 12 patients with moderate injury, and 8 patients with mild injury. Patients completed the Logical Memory subtest (LM) of the Wechsler Memory Scale-Revised, Paced Auditory Serial Addition Test (PASAT), Beck Depression Inventory (BDI), and Personal Capacities Questionnaire (PCQ). A clinician working closely with the patient also completed the Functional Assessment Inventory (FAI) and the Behavior Checklist (BC) at the time of initial testing. Follow-up testing on available patients (n=16) was completed approximately six months after initial testing. Comparison of the functional assessment measures demonstrated that patients exhibited a decreased awareness of functional limitations relative to clinician's ratings, but identified an increased number of personal strengths. The present study demonstrates the first comparison of FAI and PCQ ratings in a TBI population, as well as the first available field research using the PCQ. Results also indicated that the only significant predictor on earned income after six months was the overall functional limitations score on the PCQ. The only significant difference in patients' test performance at six months with scores at initial testing was seen on the PASAT, which suggested that patients had a significant improvement in their speed of information processing after six months. In addition, comparison of patients from Canadian and American rehabilitation agencies, respectively, revealed no significant differences between patients at either initial testing or at follow-up.
Graduate

Tese de mestrado em Bioinformática, apresentada à Universidade de Lisboa, através da Faculdade de Ciências, 2004
This dissertation presents an application of text mining to automate the identification of the function of large sets of genes from the biomedical literature. It proposes an approach to obtain this knowledge as annotations that associate biologic entities to Gene Ontology terms. The discovered annotations can then be efficiently curated through an interactive interface. This approach was validated by building the APEG (Arabidopsis Pollen Expressed Genes) database system, which integrates information about 147 pollen selectively expressed genes of Arabidopsis thaliana, from various public databases available on the Web. APEG operates with ProFAL, a text mining and automatic database annotation tool. The effectiveness of the automatic annotation of the genes was evaluated by comparing the set of annotations discovered by ProFAL with those obtained by domain experts scanning the same literature. Functional annotations were extracted with an average precision and recall of 61% and 78%, respectively. ProFAL has also identified 21 probable functions for 8 genes, which, to the best of my knowledge, have not been documented. The results show that mining the biomedical literature can effectively increase our knowledge about a set of genes or proteins of interest, leading to more conclusive answers to the underlying biological problems.
Nesta dissertação investiga-se a aplicação da prospecção de textos da área da biomedicina para automatizar a identificação da função de colecções de genes de grande dimensão, sob a forma de anotações que associam entidades biológicas a termos de GO (Gene Ontology, Ontologia Génica). As anotações descobertas podem ser curadas através de uma interface de utilizador interactiva. Esta abordagem foi validada através da construção do sistema APEG (Arabidopsis Pollen Expressed Genes, Genes Expressos em Pólen de Arabidopsis), que integra informação obtida a partir de várias bases de dados públicas disponíveis na Web sobre 147 genes da planta Arabidopsis expressos selectivamente. O APEG integra a ferramenta de prospecção e anotação automática da bases de dados ProFAL. Na avaliação da eficácia da anotação automática dos genes observou-se que as anotações foram extraídas com uma precisão e abrangência médias de 61% e 78%, respectivamente. A ProFAL identicou 21 funções prováveis para 8 genes, que, segundo os dados disponíveis, ainda não foram documentadas. Os resultados mostram que a abordagem proposta permite melhorar o conhecimento relativo a um conjunto de genes ou proteínas em estudo, conduzindo a respostas mais significativas face as questões biológicas subjacentes.

Word prediction has been offered as support for children with spelling difficulties. The literature however has shown wide-ranging results, as the use of word prediction is at the cost of time and fatigue due to increased visual-cognitive demands. Spelling support with word prediction is through word completion, keystroke reduction and the interactive process between spelling and reading. The research project was a cross-over within-subject design using 80 Grade 4 – 6 children with spelling difficulties in a school for special needs. The research task took the form of entering 30 words through an on-screen keyboard, with and without the use of word prediction software. The subjects were divided into four groups, who completed the research task in combinations of one of two equivalent wordlists and the presentation order of the typing method used. The Graded Word Spelling Test, administered before the study began, served to investigate whether there was a relationship between the children’s current spelling knowledge and word prediction efficacy. The results indicated an increase in spelling accuracy with the use of word prediction, but at the cost of time and the tendency to use word approximations, which decreased as grade and age increased. Children’s current spelling knowledge could not serve as an indicator of who would be most likely to benefit from word prediction use. The cross-over design counter-balanced the effects of the inequalities in the two wordlists and the effects of practice and fatigue noted in the presentation order. Further research into the impact that more extensive training and practice would have on word prediction efficacy and the usefulness of word prediction in more functional writing is necessary.
Dissertation (M (Augmentative and Alternative Communication))--University of Pretoria, 2005.
Centre for Augmentative and Alternative Communication (CAAC)
unrestricted

The utilization of in vitro tests with a tiered testing strategy for detection of mild ocular irritants can reduce the use of animals for testing, provide mechanistic data on toxic effects, and reduce the uncertainty associated with dose selection for clinical trials. The first section of this thesis describes how in vitro methods can be used to improve the prediction of the toxicity of chemicals and ophthalmic products. The proper utilization of in vitro methods can accurately predict toxic threshold levels and reduce animal use in product development. Sections two, three and four describe the development of new sensitive in vitro methods for predicting ocular toxicity. Maintaining the barrier function of the cornea is critical for the prevention of the penetration of infections microorganisms and irritating chemicals into the eye. Chapter 2 describes the development of a method for assessing the effects of chemicals on tight junctions using a human corneal epithelial and canine kidney epithelial cell line. In Chapter 3 a method that uses a primary organ culture for assessing single instillation and multiple instillation toxic effects is described. The ScanTox system was shown to be an ideal system to monitor the toxic effects over time as multiple readings can be taken of treated bovine lenses using the nondestructive method of assessing for the lens optical quality. Confirmations of toxic effects were made with the utilization of the viability dye alamarBlue. Chapter 4 describes the development of sensitive in vitro assays for detecting ocular toxicity by measuring the effects of chemicals on the mitochondrial integrity of bovine cornea, bovine lens epithelium and corneal epithelial cells, using fluorescent dyes. The goal of this research was to develop an in vitro test battery that can be used to accurately predict the ocular toxicity of new chemicals and ophthalmic formulations. By comparing the toxicity seen in vivo animals and humans with the toxicity response in these new in vitro methods, it was demonstrated that these in vitro methods can be utilized in a tiered testing strategy in the development of new chemicals and ophthalmic formulations.