Quantification of Serum Protein using 2D Electrophoresis Image

Abstract

Serum Protein Electrophoresis is a widely used test to identify the presence of various Serum proteins such us, Albumin, Alpha, Beta and Gamma. In electrophoresis testing, charged protein molecules are moved in an electric field according to their size and weight. When a unique gel component is added with serum as a color die that generates an image on a glass plate. This electrophoresis image is used to quantify as well as analyze the various proteins using image processing technique. We propose an image processing framework to detect the various protein bands and length, also quantification of each band. The software was fully developed in MATLAB, which consists of three stages. In the first pre-processing stage, un-sharpening filter is used to enhance the frequency components of image. Next, Multidimensional filter is used in second stage for edge correction. Savitzky-Golay filter is used in third stage to smooth out the noisy signal. Finally, quantification of each protein band is done by manually. Obtained total protein and albumin of various Serum samples are compared with traditional test results, where they are very closer or the same.