Engineering and computer science.

AN AUTOMATED SEGMENTATION METHOD FOR MICROARRAY IMAGE ANALYSIS. Wei-Bang Chen and Chengcui Zhang, Dept. of Computer and Information Sciences, University of Alabama at Birmingham, AL 35294. Wen-Lin Liu, Dept. of Management, Marketing, and Industrial Distribution, University of Alabama at Birmingham, AL 35294

Biologists use microarray to detect gene expression level in parallel for understanding the relationship of genes. Gridding and spot segmentation are two critical steps in microarray gene expression data analysis. However, the problems of noise contamination and donut-shaped spots often make signal extraction process a labor-intensive task. In this study, we propose a three-step approach for automatic gridding and spot segmentation. The approach starts with a background removal and noise eliminating step. The background removal process uses a local-global thresholding technique to pre-label and identify pixels that are most likely background pixels and then remove them from the foreground. The noise eliminating process uses a voting method based on spatial connectivity to eliminate the majority of noise. The second step applies a fully unsupervised method to extract blocks and grids from the pre-labeled and cleaned data. The third step applies a simple, progressive spot segmentation method to deal with inner holes and noise in spots. We tested its performance on real microarray images against widely used software GenePix. Our results show that the proposed method deals effectively with poor conditioned microarray images in both gridding and spot segmentation.

A SIMPLE MACHINE PROOF OF EULER'S FORMULA. Jun Zhang. Dept. of Math & CS, Troy University, Troy, AL 36082

We are living in an information age, such that everything is digitalized. The encoding of functions and the automatic proof of functions are very important in both theories and applications. This paper will discuss a simple method of calculation for Taylor expansion coefficients, which can be easily implemented in a computer algebra system like Maple. As an example, it can be applied to prove Euler's formula automatically.

NEEDS ASSESSMENT FOR SCIENTIFIC VISUALIZATION OF MULTIVARIATE, HIGH-DIMENSIONAL MICROARRAY DATA. Vetria L. Byrd, Department of Computer and Information Sciences, Univ. of Ala., Birmingham, Birmingham, AL 35294. Tarynn M. Witten, Center for the Study of Biological Complexity, Virginia Commonwealth Univ., Richmond, VA 232840.

Microarrays are a high throughput technology that allows for the analysis of hundreds to thousands of genes simultaneously. We hypothesized there is an unmet need for more advanced visualization tools that are not currently available. A survey research instrument was designed, and administered from two universities. The protocol was approved by the Institutional Review Boards at Virginia Commonwealth University, and The University of Alabama at Birmingham. Participants were chosen from life and biomedical/health scientists, research institutes, biotech companies and other researchers who do microarray data analysis. Participation in the survey was voluntary. Basic non-identifiable demographics were requested. The survey was designed to ascertain from participants what microarray technologies, microarray tools as well as what graphical methods are commonly used for representing microarray data. Responses from the survey indicated there is a need for an integrative visualization tool for microarray analysis. Respondents indicated they were not satisfied with the ability of the current software tools to visualize their data; however, they are unable to describe new tools and features that could assist them. This suggests that a more comprehensive, thorough study should be implemented and structured to reach a more extensive microarray user base. Responses from this research effort identified a number of open problems that remain to be addressed in the microarray process. Support of this project was funded through grant EEC0234104 from NSF/NIH Bioinformatics and Bioengineering Summer Institute Program at the Virginia Commonwealth University, Center for the Study of Biological. IRB approval: UAB (#X050822007), VCU (#5065).

CLONE DETECTION WITH MICROSOFT PHOENIX. Robert Tairas, Department of Computer and Information Sciences, University of Alabama at Birmingham, Birmingham, AL 35294.

A code clone represents a sequence of statements that are duplicated in multiple locations of a program. Clones often arise in source code as a result of multiple cut/paste operations on the source, or due to the emergence of crosscutting concerns. Programs containing code clones can yield problems, especially during the maintenance phase. When an update is needed on the original copy of a code section, all clones of that code section must also be found so that they can be updated accordingly. If done manually, the detection of clones can be a slow and tedious activity that is also error prone. A tool that can automatically detect clones offers a significant advantage during software evolution. With such a tool, clones can be found and updated in less time. Moreover, these clones can be considered for restructuring or refactoring, which could yield better performance and modularity in the program.

This presentation will describe an investigation into an automatic clone detection technique that finds function level clones in a program. Microsoft's new Phoenix framework is used to host the clone detection process. Our process of detecting clones utilizes abstract syntax trees and suffix trees. The nodes of an abstract syntax tree are used to generate a suffix tree and we use the same methods on the suffix tree that have been successfully applied to biological sequence matching to search for duplicate sections of code.