Speaker
Description
Cancer is a deadly disease worldwide and an early diagnosis is essential to a successful treatment and the development of effective therapeutic strategies. Previously, approaches for diagnosis are expensive and most often are able to detect the disease when it has already grown. Moreover, analytical and computational methods are complicated by the heterogeneity of cancer cell phenotypes that exist within a single tumor, which drives metastasis, resistance to treatment, and eventually, recurrence.
However, with the recent advent of sequencing technologies (RNA-Seq) which are able to monitor the activities of the cells, revealing molecular aberrations in the cells and gene expression levels, large-scale datasets have been produced. The analysis of these datasets could reveal diseases initiating genes in the cells.
Notably, not all faulty genes lead to cancer therefore, we seek to identify some of these key genes and develop a support vector classifier that will learn from these key genes and use it to predict new instances of the disease.
