Hosted by the Jean Golding Institute and the Heilbronn Institute
Point processes are a rich and flexible class of models, that can be put to use to answer quantitative questions in science and engineering. Determining the best way to map a real-world application to the models is not always straight forward and the interpretation of any findings needs to be done in light of the subject matter interpretation of the point process model.
We have used point processes for quality assessment of digital X-ray detectors, based one dead pixel formation. We have developed distribution-free procedures, to quantify the significance of calculated differences between point pattern structures in fluorescent microscopic images of protein abundance.
The method is illustrated by experimental data shedding light on the interplay between subcellular structures called microtubules and chemicals involved in mitosis. Other applications in microscopy relate to the question of colocalisation, which is of interest for understanding the protein interaction. We further used a specific class of point processes, Hawkes processes, to fit COVID infection and death data.