Hide/Show Apps

Analyses and modeling of ovarian cancer microarray data

Download
2019
Karakelle, Barış S
Ovarian cancer is one of the common cancer types among other oncological diseases. The major causes of this cancer can be listed as age, obesity, hormone therapy, material inheritance and contraceptive pills. Due to its generality and importance, many researches have been conducted from distinct labs about this illness and its plausible causes have been intensively investigated either inmicroarray studies, where just part of the related genes are detected, or in thepairwise correlation analyses between the disease and selected symptoms via contingency tables. Hereby, in this study, we use an ovarian cancer microarray dataset and describe gene interactions in these data via two different modelling approaches, namely, Gaussian graphical model as a parametric model and artificial neural network as a nonparametric model. From these analyses, we evaluate certain findings biologically and then, compare the performance of the model accuracies in distinct accuracies measures by controlling the true network structures of selected genes. By this way, we aim to assess the performance of these two fundamental models by using this specific oncogene data.