Oncogene Identification using Filter based Approaches between Various Cancer Types in Lung

Lung cancer accounts for the most cancer related deaths for men as well as for women. The identification of cancer associated genes and the related pathways are essential to provide an important possibility in the prevention of many types of cancer. In this work two filter approaches, namely the information gain and the biomarker identifier (BMI) are used for the identification of different types of small-cell and non-small-cell lung cancer. A new method to determine the BMI thresholds is proposed to prioritize genes (i.e., primary, secondary and tertiary) using a k-means clustering approach. Sets of key genes were identified that can be found in several pathways. It turned out that the modified BMI is well suited for microarray data and therefore BMI is proposed as a powerful tool for the search for new and so far undiscovered genes related to cancer.

Evaluation of the Impact of Dataset Characteristics for Classification Problems in Biological Applications

Availability of high dimensional biological datasets such as from gene expression, proteomic, and metabolic experiments can be leveraged for the diagnosis and prognosis of diseases. Many classification methods in this area have been studied to predict disease states and separate between predefined classes such as patients with a special disease versus healthy controls. However, most of the existing research only focuses on a specific dataset. There is a lack of generic comparison between classifiers, which might provide a guideline for biologists or bioinformaticians to select the proper algorithm for new datasets. In this study, we compare the performance of popular classifiers, which are Support Vector Machine (SVM), Logistic Regression, k-Nearest Neighbor (k-NN), Naive Bayes, Decision Tree, and Random Forest based on mock datasets. We mimic common biological scenarios simulating various proportions of real discriminating biomarkers and different effect sizes thereof. The result shows that SVM performs quite stable and reaches a higher AUC compared to other methods. This may be explained due to the ability of SVM to minimize the probability of error. Moreover, Decision Tree with its good applicability for diagnosis and prognosis shows good performance in our experimental setup. Logistic Regression and Random Forest, however, strongly depend on the ratio of discriminators and perform better when having a higher number of discriminators.