Derivation of an Algorithm for the Analysis of Images of DNA molecules prepared with Denaturation Mapping

Denaturation mapping is a powerful and fast method for optically analyzing DNA molecules.It can be used to characterize DNA molecules on a scale of a few hundred base pairs which is sufficient for applications. Denaturation mapping requires a sufficiently high optical resolution of the imaging system used, which typically translates to an expensive, bulky microscope. For point-of-care testing or usage in poor, remote regions of the world, it would be desirable to perform denaturation mapping without the need of a high-end microscope. In this thesis, an algorithm is presented and analyzed which reconstructs a high-resolution barcode signal from a series of low-resolution images. The implementation of this algorithm in imaging systems with optical resolutions that are too low to be useful otherwise could unlock the possibility of performing denaturation mapping. A necessary precondition for the algorithm to work is that the series of low resolution images have a known, unidirectional shift between them. With this knowledge, the signals of the different images can be merged into a single signal which can then be deconvolved with a square function of the same size as a single pixel. This would, in theory, lead to a resolution that is only dependent on the step size between the images, which is very suitable for imaging systems with large pixel sizes. Those imaging systems generally have the advantage of having a better signal to noise ratio than imaging systems with smaller pixel sizes. The aim was to provide a proof-of-principle of the idea behind the algorithm and to verify it with experiments. During this master thesis the algorithm was tested on artificial data and afterwards, its performance was analyzed with over 60 DNA molecules.Will you be able to see DNA with your phone in the future?Can you do me a favor and picture a microscope? Got it? Great. Can I guess that the microscope you imagined looked something like the one shown in figure 1. A super technical manifestation of over 300 years of research. Can you do me another favor? Can you imagine transporting this microscope from village to village in the midst of the African jungle? Not the most pleasant thought is it? That is why scientist are currently trying to develop a new method of microscopy that is only using a normal off-the-shelf smartphone and some intelligent algorithms.The most important characteristic that determines the quality of a microscope is its resolution. The resolution of a microscope is given as a distance and it indicates the closest two individual object scan get while still being distinguishable. The resolution of a smartphone camera can generally be computed by one characteristic number: the amount of (mega)pixels used by the sensor of the camera. For an image of a given size the resolution is then determined by the size of the image divided by the amount of pixels. This resolution is normally way too low for a smartphone to work as a microscope. However, if the camera records a video where the object you are trying to image is moving less then the pixel size between two consecutive frames of the video then the resolution can be increased. The image sequence that has been recorded can be used to reconstruct a single image of a high resolution which is determined by the product of the time between two consecutive frames and the velocity with which the image you are trying to image is moving and this product can be much smaller than the aforementioned pixel size.This technology has the potential to revolutionize microscopy by abandoning (almost) all of the bulky, expensive equipment used in a conventional microscope, such as the one shown in figure 1, therefore lowering the costs tremendously and increasing the usability for point-of-care testing enormousl