Yoseph Barash

Department of Genetics Perelman School of Medicine

Department of Computer and Information Science

 

"Probabilistic Models for Alternative Splicing"

 

Abstract

The bio-medical field is now exploding with vast amounts of data that require new computational methods to process and analyze it. In this talk I will describe how probabilistic models and machine learning algorithms are used to derive new knowledge in the bio-medical field. We will concentrate on one important process in gene expression, termed alternative splicing. No biological background is required for the talk.

Why alternative splicing?
Alternative splicing enables us to create up to thousands of different messages (mRNA) from a single gene DNA code. Alternative splicing occurs in more than 90% of human genes, and splicing defects are associated with many human diseases.

Why Machine Learning?
In the talk I will describe how probabilistic models allow us to (a) quantify changes in splicing across tissues from hundreds of thousands of noisy expression measurements, and then (b) infer what are the regulatory elements that control these changes.
 

 

Refreshments will be served on the

2nd Floor Mezzanine Level

outside Wu & Chen

immediately following the talk.

____________________________________________________________________________