Novel detection schemes and automated image analysis algorithms for planar chromatography and gel electrophoresis.

摘要

After a detailed introduction to Charge Coupled Devices (CCDs) and a discussion of personal computer capabilities, examples of their applications involving novel analytical techniques are presented.;A new laser-based indirect fluorometric detection method for thin-layer chromatography is demonstrated with digitally controlled two-dimensional scanning and computer-aided image acquisition and vision. Utilizing this mode, universal detection is possible without derivatization, with a detection limit of 6 pg; 100 times lower than with human vision.;A detection scheme for DNA in agarose slab gel electrophoresis based on ultraviolet absorption allows for convenient in situ monitoring of conventional gels via a CCD camera system, which greatly improves detectability compared to previous applications of this detection mode. With the use of the CCD system, low light level images can be acquired rapidly with minimal noise. Further on-line processing of the image allows one to correct for background and lighting variations and other non-random noise. The detection limit is about 5 ng of double-stranded DNA per band, which is at least 25 times lower than those obtained previously for UV absorption.;The UV imaging method, along with a native fluorescence scheme, is demonstrated for detection of proteins in agarose slab gels. Detectability is comparable to traditional visualization schemes based on coomassie blue staining.;A computerized method for correcting systematic signal distortions of images from slab gel electrophoresis is presented. The method is useful for interpretation of one-dimensional slab gel electrophoresis images where the information is manifested in rectangular shaped bands and the distortions can be described by low order polynomial functions. The purpose is to eliminate human judgement and interaction from the process, which is necessary for automated analysis of electrophoretic gels.;A computerized expert system is introduced for the reduction of data acquisition in random access two-dimensional scanners or imaging devices where multiple scans are necessary to identify all features. The various areas of the image are probed to different extents based on the accumulated signal to noise ratio (SNR) of each area, resulting in a constant SNR throughout the final image. The method is applied to sequencing gel interpretation.