|
|
 |
|||
 |
 Overview | Research | Education | Resources | Partnerships | Home |
||
 |
|
Adjust Font Size: "Cardiovascular Research in Action:
Moving Basic Discoveries to the Bedside" Who: Francis G. Spinale, M.D., Ph.D. (Professor of Surgery & Physiology; Vice-Chair, Department of Surgery; Director, Cardiovascular Translational Research Center
Medical University of South Carolina )
 "Immune Dysregulation and Oxidant Stress in Diabetic Vasculopathy" Who: Sonia Flores, Ph.D. (Associate Professor, Departments of Medicine & Biochemistry, Webb-Waring Antioxidant Research Institute, University of Colorado Health Sciences Center)
"A systems approach to death-survival signaling in mammalian cells" Who: Peter Sorger, Ph.D. (Professor of Systems Biology, Harvard Medical School) Abstract: Our laboratory studies the mammalian signaling circuits that determine the choice between alternate cell fates of continued proliferation or programmed cell death. We focus on pathways activated by binding of EGF, IGF1 and hepatocyte growth factor (HGF) proliferation-survival factors and the TNF, TRAIL and FAS death ligands to their receptors on human and murine cells. We approach these topics from a systems perspective in which measurement and experimental manipulations with RNAi, drugs and knock-in mutations are combined with high throughput measurement and mathematical modeling at levels of abstraction that vary from purely statistical to explicitly mechanistic. I will describe recent work on the core apoptosis pathways that regulate executioner caspases, proteases that dismantle dying cells. Caspase normally switch from off to on in an all-or-none process that enforces an unambiguous choice between life and death. To understand the operation of this switch in quantitative terms we constructed a mass-action mathematical model of receptor-mediated cell death based on known reaction pathways and then trained the model on data from single cells perturbed by protein depletion, over-expression, or inhibition. Receptor-mediated cell death is characterized by sudden and efficient cleavage of caspase substrates, but only after a remarkably long delay (1 ~12 hr), whose mean duration and variance depend on ligand dose. Caspase regulatory pathways must simultaneously achieve snap-action activation, long and variable delay and high efficiency; it is not sufficient that all processes be fast. Moreover, caspases are potent and lethal proteases that must be held safely in check throughout the life of a cell. Our analysis reveals caspase regulation to involve a multi-step feed-forward circuit whose key feature is a transcritical bifurcation in the reactions controlling mitochondrial membrane permeability, rather than the cusp bifurcation and feedback circuits familiar in gene regulatory pathways. Simulations and experiments reveal unexpected failure modes in caspase control, at least one of which has the potential to promote genomic instability and oncogenic transformation. About the Speaker: Dr. Sorger obtained his BA in Biochemistry at Harvard and Ph.D in Biochemistry from Cambridge University. His post-doctoral training included time at the Medical Research Council in Cambridge as well as UCSF in the labs of Drs. Varmus and Murray. He joined the faculty of MIT in 1994 and has recently joined the faculty of the Department of Systems Biology at the Harvard Medical School.
"The Genetic Networks Underlying Wnt Signaling During Early Mammalian Development" Who: Terry P. Yamaguchi, Ph.D. (Principal Investigator,
Cell Signaling in Vertebrate Development Section,
Cancer and Developmental Biology Lab,
National Cancer Institute,
Frederick, NIH)
|| Thursday, March 22, 2007 || "Vectorial scale-based fuzzy-connected image segmentation" Who: Ying Zhuge, Ph.D. (University of Pennsylvania, Medical Image Processing Group) Abstract: Fuzzy connectedness framework has been effectively utilized in many medical applications. In this framework, a fuzzy topological construct, fuzzy connectedness, characterizes how the spatial elements (spels) of an image hang together to form an object. A local fuzzy relation called affinity is defined on image domain; the strength of affinity between any two spels depends on how close the spels are spatially and how similar their intensity-based properties are in the image. A global fuzzy relation called fuzzy connectedness is defined on the image domain which assigns to every pair of spels a strength of global hanging togetherness. This value is determined by finding the strongest among all possible connecting paths between the two spels in each pair. The strength assigned to a particular path is defined as the weakest affinity between successive pairs of elements along the path. This talk presents a generalization of the theory and algorithms of fuzzy connectedness from scalar images to vectorial images. The vectorial images may represent color images with R-, G-, and B-components, or multispectral images such as T1-, T2-, and PD-weighted images acquired via MR imaging, or vector-valued features estimated from the given scalar or vectorial image. Two different components of affinity, namely homogeneity-based affinity and object-feature-based affinity, are devised in a fully vectorial manner. Local morphological structure called scale is used to determine affinity between any two spels, which makes the method more robust to noise. To obviate the need for a threshold of fuzzy-connected image segmentation, the relative fuzzy connectedness algorithm is utilized to delineate a specified object via a competing strategy among multiple objects. Since vectorial image provide more information than scalar image, our desire to develop fully vectorial scale-based fuzzy connectedness approach was based on the postulate that a fully vectorial formulation leads to more effective segmentation. Several studies will be shown to evaluate the performance of this method based on simulated and clinical MR images. Finally, I will describe an parallel implementation of fuzzy-connected image segmentation by using Message Passing Interface on a cluster of workstations, in order to efficiently segment objects in high resolution, very large image data set. Without need of specially high processing power of CPUs/Memory for desktop workstation, we achieve almost linear speedup for parallelization comparing to the sequential implementation of the same operation.
"cGMP, Calcium and Neurodegenerative Diseases" Who: Vishy Ramamurthy, Ph.D. (Assistant Professor, Department of Ophthalmology, Sensory Neuroscience Research Center, West Virginia University Eye Institute) Refreshments will be served at 3:45 p.m.
|
|
