Parallel implementation of the variant calling process using the UPMEM technology

Organisms have enormous amounts of instructions encoded in their DNA. When DNA sequences are found using sequencing methods, such as New Generation Sequencing, they are stored as very large datasets. Merely knowing the sequence of an organism’s DNA is not enough, however, and there needs to be further processing of the datasets to extract meaningful data about the organism’s cellular structure and function. The way to process DNA sequences in order to understand why certain cancers and other diseases occur is to look at the variance between the DNA sequences of organisms, this is called variant calling. In this project, we are implementing a more rapid and energy-efficient method to perform variant calling by using High Performance Computing techniques. The approach of this method will be the processing-in-memory architectures by using the UPMEM company’s DRAM chips.

Faculty Supervisor:

Lucian Ilie

Student:

Partner:

Université de Rennes 1

Discipline:

Life Sciences

Sector:

Education

University:

Western University

Program: