"Cancer continues to be one of the major causes of deaths worldwide. Advancement in health-care technologies has made it possible to detect it at an early stage with the help of cancer molecular markers. However, there still exists open problems in this area. NGS technology has revolutionized cancer research by making availability of omics data, faster and easier than the traditional sequencing techniques. To unveil molecular mechanisms of complex diseases like cancer, efficient computational methods are needed to deal with this high-throughput and high-dimensional data. In this regard, two major questions can be addressed from this high dimensional biological data. (1) Identification of responsible biomarkers, (2) Grouping biomarkers involved in similar biological processes.
In this regard, various computational techniques have came up as a promising solution to these problems. These techniques combine the knowledge of statistical as well as computational approaches to extract useful information from the omics data. They, however, have some shortcomings and as a result, most of them are although computationally efficient, but do not always produce biologically meaningful results. The current thesis work focuses on overcoming some of these shortcomings and developing different computational techniques that can extract biologically relevant information from omics data. This thesis comprises two studies, they are:

\begin{enumerate}
\item Development of a computational method for identification of responsible biomarkers
\item Development of an algorithm for grouping effective biomarkers potentially involved in same biological processes
\end{enumerate}
This information can be further utilized to develop prognostic, diagnostic as well as therapeutic tools for various cancer types. The effectiveness of proposed approaches is demonstrated over other existing methods on several data sets using standard matrices and enrichment analysis."