Glioblastoma multiforme (GBM) is the most common and lethal form of primary brain tumor that mainly affects adults. GBM displays remarkable intra- and inter-tumoral heterogeneity and contains a subpopulation of cells named glioma stem cells that is believed to be responsible for tumor maintenance, progression and recurrence. We have established and characterized a biobank of 48 cell lines derived from GBM patients. The cells were explanted and maintained as adherent cultures in serum-free, defined neural stem cell medium. These GBM cells (GCs) displayed NSC marker expression in vitro, had orthotopic tumor initiating capability in vivo, harboured genomic alterations characteristic of GBM and represented all four TCGA molecular subtypes. Our newly established biobank is also connected with a database (www.hgcc.se) that provides all molecular and clinical data. This resource provides a valuable platform of valid in vitro and in vivo models for basic GBM research and drug discovery. By using RCAS/tv-a mouse models for glioma, we found that GBMs originating from a putative NSC origin caused more tumorigenic GCs that had higher self-renewal abilities than those originating from putative glial precursor cell origin. By transcriptome analysis a mouse cell origin (MCO) gene signature was generated to cluster human GCs and GBM tissue samples and a functional relationship between the differentiation state of the initially transformed cell and the phenotype of GCs was discovered, which provides the basis for a new predictive MCO-based patient classification. LGR5 was found to be highly expressed in the most malignant mouse GC lines of putative NSC origin and also enriched in proneural GBMs characterized by PDGFRA alterations and OLIG2 up-regulation. By overexpressing or depleting LGR5 we discovered that high LGR5 expression in proneural GC lines increased the tumorigenicity, self-renewal and invasive capacities of the cells and could potentiate WNT signalling through its ligand RSPO1. Through transcriptome analysis we identified the candidate genes CCND2, PDGFRA, OLIG2, DKK1 that were found to be regulated by LGR5. In the last study, we found that mouse OPCs could initiate both astrocytic and oligdendroglial gliomas, which indicated that oncogenic signalling is dominant to cell of origin in affecting the histology of gliomas.

Malignant brain tumors are among the deadliest of human cancers. Despit recent advancements in conventional therapies, glioblastomas remain incurable, largel y due to their ability to invade surrounding tissue. Matrix metalloproteinases are thought to contribute to the invaseive phenotype of human gliomas. Absent in normal brain, matrix metalloproteinase-1 (MMP-1) has been shown to be present in gliomas, and in particular in glioblastoma multiforme (GBM). To begin to examine the role of MMP-1 in these tumore, two human glioma cell lines were stably transfected with MMP-1 cDNA. Confirmation of MMP-1 over-expression in these cells was achieved through real-time PCR and Western blot analysis. The functional consequences of MMP-1 over-expression were analyzed using a collagen type-I invasion assay along with clonogenic and ATP viability assays. Data presented demonstrate that MMP-1 over-expressing cells were more invasive in both cell types and interestingly more clonogenic in on of the glioma cell lines, supporting a possible role for MMP-1 in glioma growth and invasion.

This paper describes methods for pricing European and American options. Monte Carlo simulation and control variates methods are employed to price call options. The binomial model is employed to price American put options. Using daily stock data I am able to compare the model price and market price and speculate as to the cause of difference. Lastly, I build a portfolio in an Interactive Brokers paper trading [1] account using the prices I calculate. This project was done a part of the masters capstone course Math 573: Computational Methods of Financial Mathematics.

Ingeniero Civil Industrial / En ambientes competitivos, los consumidores pueden adquirir sus requerimientos de productos y servicios en diversas empresas. En este contexto, las compañías desarrollan estrategias de gestión de clientes, como programas de fidelización, estrategias de cross-selling o up-selling y muchos otros esfuerzos de marketing focalizados, con el objetivo de que los clientes adquieran la mayoría de sus requerimientos con ellos. Sin embargo, estas estrategias se realizan con información incompleta, puesto que desconocen el total de gasto que los consumidores poseen en cada una de las categorías. El objetivo principal de esta memoria es desarrollar una metodología que permita estimar cuánto del gasto mensual de una persona es capturado por medio de la tarjeta de crédito perteneciente a una tienda por departamentos, lo cual se denomina principalidad. Para lograr el objetivo propuesto, se trabaja con una muestra de 161.005 clientes que han realizado alguna compra en el periodo comprendido entre junio de 2010 a mayo de 2012. La estimación de la principalidad se logra utilizando el modelo binomial generalizado (GBM), el cual, en ausencia de datos del comportamiento del cliente fuera de la compañía, supone que el número de transacciones totales que éste realiza se distribuye Poisson. En primera instancia, se aplica el GBM utilizando sólo datos transaccionales, para luego incluir variables demográficas y de negocio, modelo llamado GBM extendido. La estimación realizada con el GBM extendido tiene un pseudo de 66% y un MAPE tanto de ajuste como de validación del orden del 12,5%, mientras que el modelo actual que posee la empresa tiene un error de un 25%. Si bien, tanto el modelo de la empresa como el propuesto en este trabajo entregan una principalidad promedio de un 20%, la metodología propuesta permite una mayor variabilidad en el potencial de gasto junto con asegurar que el gasto total es siempre mayor o igual al realizado con la tarjeta de crédito. Una vez realizada la estimación se realiza una segmentación basada en la principalidad y en el gasto total, donde se observa que sólo un 1,8% de la muestra total, representa a los clientes que cuentan con un alto gasto y con un gran porcentaje de éste realizado con la tarjeta. Por otro lado, al evaluar a que categoría está más vinculado cada cliente, se observa que aquellos asignados a las unidades de consumo que no son cubiertas por el holding al cual pertenece la empresa, como lo son educación, salud y recreación, presentan la mayor principalidad, donde los asignados a educación tienen el mayor valor con un promedio de 47% pero representan sólo el 0.43% de la muestra. Como trabajo futuro se propone obtener información para una muestra de clientes respecto de cuanto gastan mensualmente, con el objeto de evaluar de mejor manera los resultados entregados por la metodología propuesta y poder desarrollar modelos supervisados, para luego comparar los niveles de ajuste. Además, se propone evaluar de manera dinámica las estrategias que serán implementadas para aumentar la principalidad.

No description available.

This thesis addresses the challenge of prognosis, in terms of survival prediction, for patients with Glioblastoma Multiforme brain tumors. Glioblastoma is the most malignant brain tumor, which has a median survival time of no more than a year. Accurate assessment of prognostic factors is critical in deciding amongst different treatment options and in designing stratified clinical trials. This thesis is motivated by two observations. Firstly, clinicians often refer to properties of glioblastoma tumors based on magnetic resonance images when assessing prognosis. However, clinical data, along with histological and most recently, molecular and gene expression data, have been more widely and systematically studied and used in prognosis assessment than image based information. Secondly, patient survival times are often used along with clinical data to conduct population studies on brain tumor patients. Recursive Partitioning Analysis is typically used in these population studies. However, researchers validate and assess the predictive power of these models by measuring the statistical association between survival groups and survival times. In this thesis, we propose a learning approach that uses historical training data to produce a system that predicts patient survival. We introduce a classification model for predicting patient survival class, which uses texture based features extracted from magnetic resonance images as well as other patient properties. Our prognosis approach is novel as it is the first to use image-extracted textural characteristics of glioblastoma scans, in a classification model whose accuracy can be reliably validated by cross validation. We show that our approach is a promising new direction for prognosis in brain tumor patients.

Les Glioblastomes (GBM) sont des tumeurs cérébrales de trés sombre pronostic malgré leur traitement associant résection chirurgicale et chimio/radiothérapie. Ils contiennent notamment une sous-population de cellules souches/initiatrices de GBM (GIC) impliquée dans la radio/chimiorésistance et la récidive de ces tumeurs en étant capable de générer la majorité des cellules tumorales plus différenciées. Des études ont montré que les cellules tumorales pourraient avoir la capacité de se dédifférencier et d'acquérir un phénotype plus proche des GIC en réponse à des stress. Notre hypothèse est qu'une telle plasticité pourrait avoir lieu en réponse aux radiations ionisantes (RI) et participer à la récidive rapide de ces tumeurs après thérapie. En effet, j'ai montré que l'exposition de cultures primaires de cellules de GBM établies à partir de résection de patients à une dose infra-toxique et cliniquement relevante de RI potentialise à long terme la réacquisition de caractéristiques associées au phénotype des GIC (auto-renouvellement, expression de marqueurs souches et tumorigénicité). J'ai identifié au cours de ce processus (1) une surexpression de la protéine anti-apoptotique Survivine; dont l'inhibition pharmacologique bloque la plasticité radio-induite, (2) une reprogrammation métabolique précoce et (3) une enzyme impliquée dans l'acidification du pH extracellulaire, qui semble favoriser le processus de dédifférenciation radio-induite. A terme, le ciblage des mécanismes impliqués dans de ce processus adaptif aux RI pourrait contribuer à développer des stratégies thérapeutiques innovantes pour radiosensibiliser ces tumeurs. / Glioblastomas (GBM) are some highly lethal brain tumors despite a treatment associating surgical resection and radio-chemotherapy. Amongst these tumors, a subpopulation of radio/chemoresistant GBM stem-like/initiating cells (GIC) appears to be involved in the systematic GBM recurrence through the generation of more differenciated tumoral cells. Recent studies showed that tumor cells may have the ability to dedifferentiate and acquire a GIC phenotype in response to microenvironment stresses. We hypothesized that GBM cells could be subjected to a similar dedifferentiation process after ionizing radiations (IR), then supporting the GBM rapid recurrence after radiotherapy. Indeed, I showed that the exposure of several primo-cultures of differentiated GBM cells isolated from patient resections to a subtoxic and clinically relevant IR dose potentiated the long-term reacquisition of GIC properties (self-renewal ability, expression of stemness markers and tumorigenicity). I also identified during this process (1) an up-regulation of the anti-apoptotic protein Survivin whose pharmacological down-regulation led to a blockade of the IR-induced plasticity, (2) the presence of a metabolic shift occurring quickly after IR and (3) an enzymatic target, which appears to be involved in extracellular acidification under IR and could also potentiate the long term dedifferentiation induced by IR. At term, targeting the mechanisms associated with IR-induced plasticity in order to inhibit the IR-induced adaptive processes will likely contribute to develop some innovating pharmacological strategies for an improved radio-sensitization of these brain tumors.

Original analyses of a large vegetation cover dataset from Roosevelt National Forest in northern Colorado were carried out by Blackard (1998) and Blackard and Dean (1998; 2000). They compared the classification accuracies of linear and quadratic discriminant analysis (LDA and QDA) with artificial neural networks (ANN) and obtained an overall classification accuracy of 70.58% for a tuned ANN compared to 58.38% for LDA and 52.76% for QDA. Because there has been tremendous development of machine learning classification methods over the last 35 years in both computer science and statistics, as well as substantial improvements in the speed of computer hardware, I applied five modern machine learning algorithms to the data to determine whether significant improvements in the classification accuracy were possible using one or more of these methods. I found that only a tuned gradient boosting machine had a higher accuracy (71.62%) that the ANN of Blackard and Dean (1998), and the difference in accuracies was only about 1%. Of the other four methods, Random Forests (RF), Support Vector Machines (SVM), Classification Trees (CT), and adaboosted trees (ADA), a tuned SVM and RF had accuracies of 67.17% and 67.57%, respectively. The partition of the data by Blackard and Dean (1998) was unusual in that the training and validation datasets had equal representation of the seven vegetation classes, even though 85% of the data fell into classes 1 and 2. For the second part of my analyses I randomly selected 60% of the data for the training data and 20% for each of the validation data and test data. On this partition of the data a single classification tree achieved an accuracy of 92.63% on the test data and the accuracy of RF is 83.98%. Unsurprisingly, most of the gains in accuracy were in classes 1 and 2, the largest classes which also had the highest misclassification rates under the original partition of the data. By decreasing the size of the training data but maintaining the same relative occurrences of the vegetation classes as in the full dataset I found that even for a training dataset of the same size as that of Blackard and Dean (1998) a single classification tree was more accurate (73.80%) that the ANN of Blackard and Dean (1998) (70.58%). The final part of my thesis was to explore the possibility that combining several of the machine learning classifiers predictions could result in higher predictive accuracies. In the analyses I carried out, the answer seems to be that increased accuracies do not occur with a simple voting of five machine learning classifiers.

No description available.

Glioblastoma is the most common and lethal malignant brain tumor with a survival rate of 14.6 months and a tumor recurrence rate of ninety percent. Two key causes for glioblastomas grim outcome derive from the lack of applicable prognostic markers and effective therapeutic targets. By employing a loss of function RNAi screen in glioblastoma cells we found a list of 20 kinases that can be considered glioblastoma survival kinases. These survival kinases which we term as survival kinase genes, (SKGs) were investigated to find prognostic markers as well as therapeutic targets for glioblastoma. Analyzing these survival kinases in The Cancer Genome Atlas patient database, we found that CDCP1, CDKL5, CSNK1𝜀, IRAK3, LATS2, PRKAA1, STK3, TBRG4, and ULK4 genes could be used as prognostic markers for glioblastoma with or without temozolomide chemotherapeutic treatment as a covariate. For the first time, we found that patients with increased levels of NEK9 and PIK3CB mRNA expression had a higher probability of recurrent tumors. We also discovered that expression of CDCP1, IGF2R, IRAK3, LATS2, PIK3CB, ULK4, or VRK1 in primary glioblastoma tumors was associated with tumor recurrence prognosis. To note, of these recurrent prognostic candidates, PIK3CB expression in recurrent tumor tissue had much higher expression compared to primary tissue. Further investigation in the PI3K pathway showed a strong correlation with recurrence rate, days to recurrence and survival emphasizing the role of PIK3CB in tumor recurrence in glioblastoma. In efforts to find effective therapeutic targets for glioblastoma we used SKGs as potential candidates. We chose the serine/threonine kinase, Casein Kinase 1 Epsilon (CSNK1𝜀) as a target for glioblastoma because multiple shRNAs targeted this gene in our loss of function screen and multiple commercially available inhibitors of this gene are available. Casein kinase 1 epsilon protein and mRNA expression were investigated using computational tools. It was revealed that CSNK1𝜀 expression has higher expression in glioblastoma than normal tissue. To further examine this gene we knocked down (KD) or inhibited CSNK1𝜀 in glioblastoma cells lines and noticed a significant increase in cell death without any significant effect on normal cell lines. KD and inhibition of CSNK1𝜀 in cancer stem cells, a culprit of tumor recurrence, also revealed limited self-renewal and proliferation in cancer stem cells and a significant decrease in cell survival without affecting normal stem cells. Further analysis of downstream effects of CSNK1𝜀 knockdown and inhibition indicate a significant increase in the protein expression of β-catenin (CTNNB1). We found that CSNK1𝜀 KD activated β-catenin, which increased GBM cell death, but can be rescued using CTNNB1 shRNA. Our survival kinase screen, computational analyses, patient database analyses and experimental methods contributed to the discovery of novel prognostic markers and therapeutic targets for glioblastoma. / Ph. D.