Noise Reduction in High-throughput Gene Perturbation Screens - August 20th, 2010

by Danni Yu, Ph.D student, Department of Statistics, Purdue University

Time: 3:00pm on August 20, 2010
Place: Conference room 120, IU Innovation Center

Talk Abstract

Perturbation screens are key for understanding the function of complex organisms. When conducted on genome-wide scale, they provide invaluable global insight into the function of individual gene and gene products, and into their physical and genetic interactions. Recent technological advances enable perturbation screens that are increasingly high in throughput. As a result, the screens now play an important role in functional genomics.

Significance of these investigations extends to a broad range of functional biology, as well as to applications in clinical research and drug target identification. An accurate quantification of screened phenotypes is a pre-requisite for successful investigations. However, the phenotypes are frequently distorted by stochastic variation, and their interpretation can be uncertain. Since large-scale experiments take weeks, and sometimes months, a variety of experimental characteristics (e.g. instruments, labors, reagents etc) can change during this time. Therefore, experimental artifacts need to be distinguished from the systematic deviations due to the perturbations. This talk will introduce several existing normalization methods and a proposed normalization procedure that enables accurate detection of hits in complex large-scale screens.

About Danni Yu

Danni Yu is a PhD candidate in the department of Statistics at Purdue University, working under the direction of Dr. Olga Vitek (Dept. of Statistics and Computer Science) and Dr. David Salt (Dept. of Horticulture). She received a master degree in Mathematical Statistics in 2008 from the Department of Statistics at Purdue. Danni's research focuses on Statistical Bioinformatics, in particular on statistical analysis and interpretation of high-throughput gene perturbation screens.