Study On Microarray Scanner And Automatic Analysis Algorithms

Microarray technology has provided a way to quantitate the simultaneous expression of a large number of genes. As microarray that developed towards miniaturization, high throughput and weak signals needs high sensitive and addressable detection, fluorescent detection has received more and more attention. In this technique, fluorescence intensities, which reflect the degree of biochemical reaction, are detected by imaging the array with a laser and capturing the image with a photomultiplier tube or a CDD camera, resulting in the production of digital images. The images from the reaction arrays constitute the essential raw data for microarray. Therefore, microarray scanner and robust image processing are particularly important and have large impacts on downstream analysis. In this paper, the basic concepts and general application flows of microarray technology are introduced, including the detailed principles of fluorescence detections for microarray. According to these principles, a microarray scanner with sensitive CCD based line scanning technology is designed and made. The performances, including linearity, sensitivity and repeatability, are tested with special methods. This microarray scanner is a high performance detection system at an incredibly affordable price. To analyze microarray image automatically, the algorithm based on projection principle and that based on image statistical information and probe design for automatic gridding from two types of microarray images are introduced. These two algorithms include image filtering, extracting gray information from the image, addressing of spots and adjusting the location coordinates of each individual spot. The test results show that the algorithms can analyze microarray images automatically and accurately. Finally we compare the results from different courses to validate the performances of our microarray scanner and the automatic analysis methods. Our implementations show that the data extracted with the automatic analysis methods from the microarray image manufactured using our microarray scanner is credible.