Statistical image analysis of photoactivated localization microscopy
Metadata[+] Show full item record
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Fluorescent microscopy is a traditional way of localizing the biological molecules in the living cells. However, the diffraction limit makes it difficult to resolve any molecules whose distance are less than 250 nm. To break the diffraction limit, many techniques have been developed recently to study the cellular structures at super resolution, which can reach 10-20 nm lateral resolution and 30-50 axial resolution. Photoactivated localization microscopy is one super resolution technique that is based on photoswitchable fluorescent proteins and 2D guassian fitting. By randomly photoactivating a small percentage of molecules back to fluorescent state, the centriod of fluorophore can be fitted to 20-30 nm that is beyond the diffraction limit. After thousands cycles of photoactivation, most of fluorescent molecules are photoactivated and localized to super resolution. Image was constructed based on all the fitted image to form the final super resolution image. The limitation of this technique is that you have to fix the cells, which means that the cells are dead. It also takes lots of time to obtain one super resolved image. However, if you can pick up the appropriate biological question, super resolution microscopy still can give you some important and interesting conclusions that can not be provided by other techniques. For example, the structure of telomere can be resolved clearly that is smaller than the diffraction limit. In this study, we used statistical method to analyze the images of photoactived localization microscopy. The 2D guassian fitting was used to obtain the centriod of each fluorescent molecule. The Ripley K value was calculated to measure the association of molecules. The dimension of telomere can be resolved by the cluster analysis.
Access is limited to the campuses of the University of Missouri.