Liu Delivers Plenary Keynote at ISPACS Conference
ECE Professor K. J. Ray Liu delivered a plenary keynote on February 9, 2009 in Bangkok, Thailand, at the International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS), one of the most well known international symposiums in signal processing, and communication systems. The title of the talk was “Genomic/Proteomic Signal Processing for Cancer Classification and Prediction.”
More information about the conference can be found at: http://www.ispacs2008.org/.
Below, please find an abstract of Dr. Liu's talk:
Genomic/Proteomic Signal Processing for Cancer Classification and Prediction
DNA microarray and proteomic mass spectrum technologies make it possible to simultaneously monitor thousands of genes/protein expression levels and distribution. A topic of great interest is to study the different expression profiles from cancer patients and normal subjects, by classifying them at gene/protein expression levels. Currently, various clustering methods have been proposed in the literature to classify cancer and normal samples based on microarray data, and they are dominantly data-driven approaches. In this talk, an alternative model-driven approach, named ensemble dependence model, is presented aiming at exploring the group dependence relationship of gene clusters. Because of the limited size of current data, it is not feasible to examine the regulation relationship between all genes. Also, both the microarray gene expression and mass spectrum data are noisy. However, if they are clustered in a right way, the noise level in the resulting cluster expression will be reduced, thus the ensemble dependence dynamics of gene clusters will be revealed. Under the framework of hypothesis-testing, we employ genes’ dependence relationship as a feature to model and classify cancer and normal samples. The classification scheme is then applied to several real cancer data sets. It is noted that the method yields very promising performance. Further investigation of the eigen-value pattern of the proposed method allows us to discover that the transition in between cancer and normal patterns suggests that the eigenvalue pattern of the proposed models may have potential to predict the early stage of cancer development. A dependence network was also developed for biomarker identification and selection.
More information on this research is available at the IEEE Signal Processing Magazine website.
February 10, 2009