Parallel computing on RNA-Sequencing processes

• Main Author: Daniel João Lopes Moreira
• Publication Date: 2016
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: https://hdl.handle.net/10216/85786
• topic: Engenharia electrotécnica, electrónica e informática; Electrical engineering, Electronic engineering, Information engineering

Medicine is an area which have been always under a constant evolution with an increasing number of researches being performed every day. With all the innovations on this topic the concept of personalized medicine have emerged and is starting to be increasingly known. Personalized medicine is a medical model that proposes the customization of health care, through changes or variations in either medical decisions, practices or even medications. This model was created as an attempt to provide patients with individualized optimal treatment in every occasion, as treatments certainly do not affect every single person in the same way. For this model to be, possibly, part of our future, medical care must start to be individually customized based on the personal characteristics of every long-suffering. The only way in which this can be achieved is by knowing before-hand the treatments the set of patients' unique characteristics or in other words by knowing every person's unique variation of the human genome, so that the best treatment for that specific patient can be selected. Moreover another possibility that can come from the analysis of a group of patients' genome is the acquisition of crucial knowledge for the development of better treatments for a specific disease. However, the process of obtaining the individual's variation of genome is a very complex and time consuming process and that's the reason why quite recently some tools have been being designed and developed to help and execute this process. The main goal of this Dissertation is to, based on iRAP which is one of the existing tools, optimize it by mostly taking advantage of multicore architecture, which is nowadays available in every CPU and GPU. In an initial stage, the already existing tool, have been carefully looked at and profiled, registering execution times, making possible the clear identification of the possible points of execution where it can be improved.