- Start Date:
June 4, 2014
- Event Start Time:
10:00 AM
- Event End Time:
11:00 AM
- Organizers:
Gene Fiorini
- Event Location:
DIMACS Seminar room
- Seminar Series:
REU Seminar
- Presenter(s):
Jaroslaw Zola - Rutgers University
- Abstract:
Have you ever wondered how many microorganisms live in an air you breath or a dirt you touch? And what about your gut or skin? Who are these microorganisms? What do they do? These questions are the focus of metagenomics, the study of a population of organisms by fragmenting and sequencing their collective DNA. Metagenomic protocols are essential in discovering composition of environmental samples, and thus are important tool in medicine, environmental studies, and national security. Like many other areas of life sciences metagenomics is a source of many fascinating problems for computer scientists and engineers. In this tutorial, we will learn how some of the classic challenges in metagenomics can be expressed via graph-theorethic concepts and how parallel and distributed computing can be used to scale the resulting computational problems to big data and large parallel computers. We will start with a gentle introduction to the underlying biological concepts, followed by a presentation of an efficient sketching-based algorithm to construct DNA similarity graphs, and a Map-Reduce heuristic to enumerate quasi-cliques. We will conclude the tutorial with a brief description of some interesting open problems in metagenomics. To make the tutorial even more inviting, we will look at belly button as a source of input data.
