Design and Implementation of a Computational Platform and a Parallelized Interaction Analysis for Large Scale Genomics Data in Multiple Sclerosis

Sammanfattning: The multiple sclerosis (MS) genetics research group led by professor Jan Hillert at Karolinska Institutet, focuses on investigating the aetiology of the disease. Samples have been collected routinely from patients visiting the clinic for decades. From these samples, large amounts of genetics data is being generated. The traditional methods of analyzing the data is becoming increasingly inefficient as data sets grow larger. New approaches are needed to perform the analyses. This thesis gives an introduction to the relevant genetics and discusses possible approaches for enabling more efficient execution of legacy analysis tools, as well as improving a gene-environment and gene-gene interaction analysis. Different computational paradigms are presented followed by the implementation of a computational platform to support the researchers’ existing, and possibly future, analysis needs. The improved interaction analysis application is then implemented and executed in a virtual instance of this platform. The performance of the analysis application is then evaluated with respect to the original reference application.