Abstract: The image analysis of two-dimensional electrophoresis images is a difficult task were authors were not able to find any other work in the literature handling with evolutionary computation in combination with a second order operator for edge detection. In this work, a novel Genetic Algorithm-based protein detection method from two-dimensional electrophoresis gel images is presented. Such a method makes use of a second order operator for edge detection by means of a Genetic Algorithm-based technique. The proposed method is able to detect proteins in two-dimensional gel images, but a reduction in the False Positive ratio is necessary. A manually selection process should be done by the clinicians to reduce this ratio; this represents a bottleneck due to the number of proteins in each imagein order to discriminate real proteins detected. The goal here was to avoid the loss of time caused by the manual revision of proteins detected by the image analysis software packages. To decrease this ratio, binary and real coded Genetic algorithms were probed and BLX-alpha crossover function was chosen. A comparative test with Z3 and Melanie 3.0, two-dimensional electrophoresis image analysis software packages, is done in order to check the accuracy of the proposed method. All images used for these tests are available on the Internet (http://www.umbc.edu/proteome).

Keywords: Image Analysis, Real Coded Genetic Algorithm, Optimization, Electrophoresis, Edge Detection.

ACM Classification Keywords: F.1.1 Models of Computations – Self-modifying machines. I.5.2 Design Methodology – Classifier design and evaluation. I.5.4 Applications – Computer Vision. I.4.6 Segmentation – Edge and feature detection.

Link:

GENETIC BASED SPOT DETECTION METHOD IN TWO-DIMENSIONAL ELECTROPHORESIS IMAGES

CarlosFernandez-Lozano?,Jose A. Seoane, Daniel Rivero, Julian Dorado

http://www.foibg.com/ijitk/ijitk-vol07/ijitk07-01-p08.pdf