2016 - 19th Annual Steven Galovich Memorial Student Symposium

Presentation Title

Identifying interactions within a Biological system

Student Presenter(s) and Advisor

Nicholas Escanilla

Location

Meyer Auditorium

Abstract

A key focus in systems biology is identifying and studying interactions between the main components of a system. This task can be particularly challenging when the components interact in complex, referred to as “non-linear”, ways. A quintessential example of a non-linear interaction is the exclusive-or (XOR) of two components. In this presentation I will discuss work that combines a non-linear support vector machine with a novel feature selection algorithm to identify the most relevant components of a given system. The novel algorithm is evaluated on synthetic data in which the ground truth components are known and follow XOR relationships. The algorithm is then applied to germline genomic data comprised of single-nucleotide polymorphisms (SNPs) from a region of the human genome believed to contain interactions that modulate breast cancer risk.

Presentation Type

Individual Presentation

Start Date

4-5-2016 9:00 AM

End Date

4-5-2016 10:15 AM

Panel

Genomics, Biostatistics, and Predictive Medicine

Panel Moderator

Cathy Benton

Field of Study for Presentation

Computer Science, Biology, Mathematics

No downloadable materials are available for this event.

Share

Import Event to Google Calendar

COinS
 
Apr 5th, 9:00 AM Apr 5th, 10:15 AM

Identifying interactions within a Biological system

Meyer Auditorium

A key focus in systems biology is identifying and studying interactions between the main components of a system. This task can be particularly challenging when the components interact in complex, referred to as “non-linear”, ways. A quintessential example of a non-linear interaction is the exclusive-or (XOR) of two components. In this presentation I will discuss work that combines a non-linear support vector machine with a novel feature selection algorithm to identify the most relevant components of a given system. The novel algorithm is evaluated on synthetic data in which the ground truth components are known and follow XOR relationships. The algorithm is then applied to germline genomic data comprised of single-nucleotide polymorphisms (SNPs) from a region of the human genome believed to contain interactions that modulate breast cancer risk.