Automated detection of fluorescent probes in molecular imaging

Abstract

Complex biological features at the molecular, organelle and cellular levels, which were traditionally evaluated and quantified visually by a trained expert, are now subjected to computational analytics. The use of machine learning techniques allows one to extend the computational imaging approach by considering various markers based on DNA, mRNA, microRNA (miRNA) and proteins that could be used for classification of disease taxonomy, response to therapy and patient outcome. One method employed to investigate these markers is Fluorescent In Situ Hybridization (FISH). FISH employs probes designed to hybridise to specific sequences of DNA in order to display the locations of regions of interest. We have developed a method to identify individual interphase nuclei and record the positions of different coloured probes attached to chromatin regions within these nuclei. Our method could be used for obtaining information such as pairwise distances between probes and inferring properties of chromatin structure.