Integration of Heterogeneous Computing Infrastructures for Genome Sequencing Data Analysis
Aulov V., Klimentov A., Mashinistov R., Nedoluzhko A., Novikov A., Poyda A., Sharko F., Tertychnyy I., Teslyuk A.
National Research Centre “Kurchatov Institute”, Moscow, Russia
Annotation. Recent technological achievements in Next Generation Sequencing (NGS) lead to significant increase in amounts of data which must be processed, analyzed and made remotely available to scientists. This, in turn, increased the requirements to the computing platforms used for data processing in terms of RAM and processors’ power. Entirely new approaches in organization of computations are required in order to process data efficiently. Authors of this paper have researched the possibility of adapting the methods and approaches employed in high energy physics for integration of heterogeneous computing resources into unified computing platform. Fully specified data and task management system based on computing resources of National Research Centre “Kurchatov Institute” was developed. We’ve also developed a workflow for processing genome sequencing data with PALEOMIX package and integrated it into the system. Results of the developed system’s approbation on ancient mammoth DNA sequencing task have demonstrated a significant reduction in total computational time of the task.
Key words: distributed computations, supercomputers, genome sequencing.