Integration techniques for modern bioinformatics workflows

Integration techniques for modern bioinformatics workflows

Recent advances in technology have allowed the mass profiling of individual genetic profiles. Although this procedure is fast and relatively inexpensive, it generates huge amounts of data that describe many stages of the genetic regulation. Interpreting the data can shed light on human physiology and, most importantly, can help improve the prognosis and treatment of certain diseases. Transforming these data into meaningful information has been a major challenge in the life sciences. To achieve this transformation, we used the synergies between various scientific fields, including information technology (IT), biology, medicine and genetics. Efforts to build this synergy have given rise to a relatively new field of science called bioinformatics.

Bioinformatics offers the scope to combine existing computational solutions into re-usable components, also called workflows. In this thesis, I present a collection of techniques to improve the quality of modern bioinformatic workflows. These techniques go beyond the classic objectives of workflows, which are to efficiently automate analyses and to make optimal use of the underlying computational infrastructure. As an extension to these objectives, I present techniques to help inexperienced users deploy such workflows, enhance their connectivity, exploit web technologies and increase their fault-tolerance. I also show how one can encourage users to engage in a community-based support system and learn to maintain workflows through the concept of crowdsourcing. Overall, these techniques can be used to increase the reproducibility of the conducted research and bring genetics closer to clinical practice.

Dissertation: http://hdl.handle.net/11370/b2bc9a72-cc56-428e-b54d-5d2c234699a0