Computational Biology of Spatiotemporal Expressions in Drosophila
Embryogenesis
Shuiwang Ji
                                                                  Arizona State
                                                                  University

Date: May 25, 2010 (Tuesday)
Place: E & C S BUILDING Second Floor Room 2120
TIME: 10:00 (DONUTS) 10:10 (TALK)


Embryonic development is achieved by the spatial and temporal control of gene
expression. The in situ images capture the spatiotemporal gene expression
dynamics and are important resources for studying development. Currently,
more than a hundred thousand in situ images are available for Drosophila
melanogaster. This wealth of data has provided unprecedented opportunities for
scientists to address questions in animal development with high confidence
and precision. However, the current practice of manually annotating these
biological images with stage- and tissue-specific expressions does not
scale with the continuously expanding collection of images, hindering
the pace of biological discoveries. 

In this talk, I will first describe our work on determining the stage- and
tissue-specific gene expressions using computational approaches. I will then
describe our work on modeling co-expression networks based on the in situ
information. One unified theme among all these tasks is the sparsity in
biological systems. I will show in this talk how such properties can be
exploited in computational systems. Finally, I will discuss how our results can
be used to study the genomic regulatory code as I outline my future research
agenda.


Biography:
Shuiwang Ji is a Ph.D. candidate in Computer Science and Engineering at
Arizona State University. His research interests are computational biology,
machine learning, and computer vision. His doctoral work is associated
with the FlyExpress project ({http://www.flyexpress.net/}[HYPERLINK:
http://www.flyexpress.net/]) in the Center for Evolutionary Medicine and
Informatics at ASU. He received the Outstanding Ph.D. Student Award in
Computer Science at ASU. He was a research intern at NEC Labs and was one of
the designers of the human action recognition system that achieved the best
performance in the worldwide TRECVID evaluation in 2009. He has served on
the program committees of ICML, SIGKDD, and ICASSP, and he has been reviewer
for a dozen of journals.