Design And Experiment Of Protein Gel Electrophoresis Image Analysis System

A simple, convenient and practical protein gel image analysis system was designedmainly used for detecting protein content, percentage, molecular weight and other parametersof protein gel, for the molecular biology research.The system consists of hardware design part and software design part. The hardware partis mainly composed of light source, camera obscura, loading platform, and digital camera. Inthe camera obscura, light source irradiates the gel and the gel image is collected by digitalcamera to the computer analysis system. The work we done is selection of light source anddigital camera, as well as the design of internal optical path in camera obscura.Software part completed the design of the protein gel image analysis software, includingimage preprocessing, division of lanes and bands, and the bands information extraction. Themain task of image pre-processing is to enhance the gel image, suppress the background andnoise, while enhancing the bands information, in this paper we use the wavelet domainenhancement algorithm, first the image signal is decomposed into a low frequency sub-band ahigh-frequency sub-band by Mallat fast decomposition algorithms, then the low frequencysub-band use B-spline wavelet to achieve the edge enhancement, while the high-frequencysub-band use wavelet domain Garrote threshold to achieve denoising, at last the enhancedimage was completed by wavelet reconstruction. Division of lanes and bands part we use apartitioning algorithm by combining modulus maximum wavelet-based edge detection andwhite spot detection, first use modulus maxima wavelet edge detection to process theenhanced image, and get a clear lane demarcation line, then use the white spot detection bysetting the white point threshold to get all the lanes divided, finally, use white spot detectionagain to divide all the bands of each lane. Bands information extraction part includes theextract of the protein content and the extract of protein molecular weight, extract of theprotein content uses the method of calculating the integrated optical density, protein content isproportional to the value of the integral optical density of the bands, the calculation of netintegral optical density can be calibrated through a standard protein, through the distributionof the optical density of the bands can also get a percentage of protein content in each lane.The extract of protein molecular weight is usually calculated by measuring the mobility, inthis paper we discuss the determination method of starting point and leading edge of bands, receive a more reliable mobility, through the relationship between molecular weight andmobility of the standard molecular weight lane, we can calculate the molecular weight of allbands on the gel image.As the design of this paper has coherence, each part of the software design was carriedout an experiment and results analysis, to ensure safe and reliable of every step, also toguarantee the accuracy of the final analysis. Through the analysis of protein gel imagesamples in this paper, we can see this analysis system has high recognition rate, low leak rate,with accurate divided position, and be able to calculate the integrated optical density,percentage, mobility, molecular weight and other parameters, the accuracy is very high, anddeserve better promotional value.