【作者】 王立强；

【导师】 陆祖康；

【作者基本信息】 浙江大学 ， 光学工程， 2004， 博士

【摘要】 生物芯片技术是生命科学领域中迅速崛起的一项高新技术，它以玻片、硅片或尼龙等为载体，在其表面高密度地排列大量的生物材料，实现对DNA、蛋白质、细胞以及其他生物组分的准确、快速、并行和大信息量的检测利分析，可广泛应用于药物研究、疾病诊断、基因结构与功能研究等领域。生物芯片采用分子杂交原理进行工作，将待测样品加以荧光染料标记，然后与已知结构的生物芯片进行充分杂交，用荧光分析仪检测发生杂交反应位置处的荧光信号。因此，荧光分析仪是获取生物芯片信息的必备仪器，也是决定生物芯片能否得到广泛应用的关键仪器。 本论文主要涉及激光共聚焦生物芯片荧光分析仪的设计与实现，详细论述了仪器的设计方案、性能指标、实验测试及结果分析，研究如何在降低成本的前提下，开发出具有自主知识产权、性能良好的生物芯片分析仪器，从产业化的角度出发，做了一些有益的探索工作，初步完成了实验样机的研制。样机中，荧光的激发光源为532nm的绿激光和635nm的红激光，采用一个光电倍增管分时实现了Cy3和Cy5两种荧光信号的检测，生物芯片的一维扫描由振镜与远心f-theta物镜实现，另一维扫描由步进电机驱动精密导轨实现。结合具体的设计参数，分析了荧光分析仪的分辨率、信噪比、探测灵敏度及动态范围，讨论了它的重复性与稳定性，研究了f-theta物镜的残留畸变和振镜的响应误差对扫描均匀性的影响，提出了光电倍增管的自动增益控制方法及串扰误差的校正方法。绿光扫描时，荧光分析仪的光学分辨率可达到5μm，Cy3荧光的探测灵敏度为1fluor／μm~2，动态范围大于10~3，完成5μm双光扫描耗时400秒，比大部分现有同类产品的扫描速度快。样机集成有嵌入式操作系统以及显示打印部件，数据处理与分析由DSP完成，荧光图像及诊断结果可以在液晶屏上显示，也可通过热敏打印机给出。整台样机面向医院应用，生物芯片的检测分析自动完成，操作方便。 本论文还研究了生物芯片荧光图像的基本处理技术。设计了一种图像增强算法，可在滤除脉冲噪声的同时，校正荧光图像中的缓慢背景变化；根据图像的行列投影，实现了分子靶点的自动定位；针对分子靶点的不规则结构，研究了基于边缘检测的自适应圆分割及基于区域生长的自适应形状分割技术，较为准确地实现了分子靶点的分割与识别。更多还原

【Abstract】 The microarray technology is a new branch of biology and life science, which emerges at a tremendous pace. It is based on the glass surface, silicon wafer, nylon substrate etc, on which large quantities of biomaterials are immobilized in high density. The microarray technology allows an accurate, rapid and parallel analysis of thousands of genes, proteins, cells and other biomolecules in a single experiment. It is becoming a common tool in many areas of biology and life research, including drug discovery, clinical diagnosis, genome structure and function research etc. Microarray experiments depend on the hybridization reaction. The test samples are tagged with the fluorescent dyes and hybridized with the microarrays. A fluorescent analyzer scans the microarrays and interprets the fluorescent signals, so it is a necessary instrument to obtain the microarray experimental results and a key one which boosts the microarray technology to be used widely.How to design and implement a laser scanning confocal microarray analyzer was introduced in the dissertation, mainly including four aspects: the design scheme, performance analysis, experimental testing and result discussion. The analyzer would be reasonable cost, good performance and own intellectual property. From the view of industrialization, we did some meaningful work and made an experimental prototype. In this prototype, the Cy3 and Cy5 fluorophores were excited by a 532nm green laser and a 635nm red laser respectively. The emitted fluorescent signal was detected using a photomultiplier tube (PMT) sequentially. One dimension scanning of the microarray slide was performed by a telecentric f-theta objective with a moving coil optical scanner; the other dimension was scanned through a stepping motor driving the precision guidance. The features and specifications of the analyzer, such as the resolution, signal-to-noise ratio, detection limit and dynamic range, were analyzed based on the actual parameters. The distortion of the f-theta objective and the frequency response error of the optical scanner were analyzed too, they would affect the field uniformity of the instrument. The reproducibility and cross-talk reduction were discussed and an auto gain control system for the PMT was proposed. When scanning by green laser, the resolution of the analyzer would be 5uim and the detection limit would be 1 fluor/um2 for Cy3. Its dynamic range of detection would be linear over 3 orders of magnitude. When scanning the microarray slide us’ ig two lasers and at 5 microns resolution, it cost 400 seconds, faster than most of the confocal microarray analyzer under the same condition. The prototype was integrated with the embedded operating system, monitor and printer. The data were processed and analyzed by a DSP. The fluorescent image and the experimental results could be displayed on the monitor or printed out by the printer. It was designed according the requirement of the hospital. All the procedures of scanning and analyzing were automated and straightforward.The basic image processing technique was also included in the dissertation. An algorithm for image enhancing was designed. It not only reduced the impulsive noise, but also corrected the background variation in the image. The microarray gridding was performed according image projections. To identify the spots from the background accurately, both the adaptive circle and adaptive shape segmentation algorithm were discussed in detail.更多还原