超分辨定位成像中的弱光探测方法研究

发布时间：2018-03-18 11:24

本文选题：超分辨定位成像　切入点：弱光探测器　出处：《华中科技大学》2014年博士论文　论文类型：学位论文

【摘要】：超分辨定位成像(Super-resolution localization microscopy,或简称定位成像)技术实现了纳米量级的空间分辨率,成为生命科学领域极具潜力的研究工具之一。定位成像本质上是依赖单分子成像实现的,然而单分子的荧光信号通常在几百至几千个光子,并且要覆盖多个像素,所以单分子微弱信号的探测是实现定位成像的关键技术之一。传统的定位成像通常使用读出速度较慢(10MHz)的弱光探测器--电子倍增电荷耦合器件(Electron multiplying charge coupled device, EMCCD),成像通量较小,限制了该技术在大样本超分辨成像(如亚突触水平神经回路成像)等方面的应用。因此研究适用于定位成像的新型弱光探测方法对提高成像通量是必要的。本文围绕该问题,建立了一套研究弱光探测器的超分辨定位成像能力的方法。具体内容如下： (1)通过表征探测器噪声和光子传递曲线等特性,从弱光探测灵敏度和成像信噪比等方面定量研究不同类型弱光探测器的整体成像性能。结果表明：(i)在弱光探测灵敏度方面,以Andor iXon897为代表的EMCCD相机在电子倍增工作模式下具有最高灵敏度(最小可探测器限为2photon),以Hamamatsu Flash4.0为代表的sCMOS相机的灵敏度稍差(最小可探测限为3photon);(ii)在成像信噪比方面,EMCCD相机在极微弱信号强度下(13photon/pixel)成像信噪比最高,但是随着信号强度的增加,sCMOS相机的成像信噪比优于EMCCD相机；(iii)sCMOS相机具备并行高速数据读出能力,其读出速度(以Hamamatsu Flash4.0为例,400MHz)比EMCCD目机(iXon897为10MHz)快四十倍,但是sCMOS相机的成像均匀性较差。综合以上发现,sCMOS有望为提高定位成像的成像速度和通量提供新的有效途径。 (2)基于点状发光体的重复成像与定位,发展了一种通过定量表征探测器的单分子探测和定位能力的实验方法。结果表明,该方法对定位精度的测量达到了～1nm的精度。进一步利用该方法,我们首次演示了在超分辨定位成像的典型信号范围(50-2000photon/pixel)内,商品化sCMOS相机(Hamamatsu Flash4.0)可以表现出优于EMCCD (Andor iXon897)的成像性能。此发现证实sCMOS相机有能力代替常用的EMCCD相机,将定位成像的成像通量提高～40倍,为亚突触水平神经回路成像提供了新的探测途径。 (3)基于仿真方法,利用单分子成像模型及固定图形噪声模型,系统研究了sCMOS相机的成像不均匀性对单分子定位的影响。结果发现：(i)固定图形噪声几乎对定位精度没有影响,但是会引入一定的定位偏差；(ii)当额外增加像素固定图形噪声一半强度的列固定图形噪声时定位偏差相应的只增加了～20%。对Hamamatsu Flash4.0sCMOS相机而言,其像素固定图形噪声要高于列固定图形噪声～3-5倍。因此,该sCMOS相机的列固定图形噪声对单分子定位的影响很难察觉。该结论消除了研究人员对sCMOS相机的成像不均匀性的顾虑,有利于推动sCMOS相机在定位成像领域中的应用。
[Abstract]:Super-resolution localization microscopy- (or "location imaging") technology has become one of the most promising research tools in the field of life science because of its nanoscale spatial resolution, which is essentially dependent on single molecule imaging. However, the fluorescence signal of a single molecule is usually in the range of hundreds to thousands of photons, and it has to cover multiple pixels. So the detection of single molecular weak signal is one of the key techniques to realize the localization imaging. The traditional localization imaging usually uses the weak light detector (Electron multiplying charge coupled device, EMC CDD), which has a slow readout speed of 10 MHz), and the imaging flux is relatively small. The application of this technique in large sample superresolution imaging (such as subsynaptic horizontal neural loop imaging) is limited. Therefore, it is necessary to study a new weak light detection method suitable for localization imaging to improve the imaging flux. A set of methods to study the super-resolution imaging ability of weak light detectors are established. The main contents are as follows:. 1) by characterizing the characteristics of detector noise and photon transfer curve, we quantitatively study the whole imaging performance of different types of weak light detectors from the aspects of low light detection sensitivity and imaging signal-to-noise ratio. The results show that: 1) in low light detection sensitivity, The EMCCD camera, represented by Andor iXon897, has the highest sensitivity (the minimum detector limit is 2 photontons) in the mode of electron multiplication, and the sensitivity of the sCMOS camera represented by Hamamatsu Flash4.0 is slightly lower (the minimum detectable limit is 3photonine / ii) in imaging signal to noise ratio (SNR). The EMCCD camera has the highest signal-to-noise ratio (SNR) of 13photon / pixel at very weak signal intensity. However, with the increase of signal intensity, the signal-to-noise ratio of EMCCD camera is better than that of EMCCD camera. Its readout speed is 40 times faster than that of EMCCD camera iXon897 (taking Hamamatsu Flash4.0 as an example). But the imaging uniformity of sCMOS camera is poor. Based on the repeated imaging and localization of the spot luminescence, an experimental method for quantificationally characterizing the detection and localization ability of single molecule detector is developed. The results show that, By using this method, we have demonstrated for the first time that we are in the range of 50-2000photon / pixel in the typical signal range of super-resolution positioning imaging. The commercial sCMOS camera Hamamatsu Flash 4.0) can show better imaging performance than EMCCD Andor iXon897. the findings confirm that the sCMOS camera has the ability to replace the conventional EMCCD camera and increase the imaging flux by 40 times. It provides a new approach for the imaging of subsynaptic horizontal neural circuits. (3) based on the simulation method, the effects of imaging inhomogeneity of sCMOS camera on single molecule positioning are systematically studied by using monolayer imaging model and fixed figure noise model. The results show that the fixed figure noise has little effect on the positioning accuracy. But a certain positioning bias will be introduced.) when the column fixed figure noise with half the intensity of the pixel fixed figure noise is added, the positioning deviation will only increase by 20%. For the Hamamatsu Flash4.0sCMOS camera, The pixel fixed figure noise is 3-5 times higher than the column fixed figure noise. Therefore, the effect of the column fixed figure noise on the single molecular positioning of the sCMOS camera is difficult to detect. This conclusion eliminates the researchers' concern about the inhomogeneity of the sCMOS camera's imaging. It is helpful to promote the application of sCMOS camera in the field of positioning imaging.
【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：R318

【共引文献】