Comparison of Algorithms for Analyzing Fluorescent Microscopy Images and Computation of Transcription Factor Profiles

Obtaining quantitative data about protein concentrations is an important component for many key concepts in systems biology, however, only a few options for generating such data exists. One of these is to use GFP reporter systems as an indicator of protein concentration, however, the measurements consist of a series of fluorescent microscopy images which need to be analyzed to derive time-dependent quantitative data.

This paper presents two techniques for determining data from fluorescent microscopy images. The first technique uses wavelets to sharpen the contrast between cells and the background which is followed up by a bi-directional search to identify the cell region. The second technique is based on K-means clustering and uses principal component analysis (PCA). A comparison of these two methods is made and a series of fluorescent microscopy images are analyzed by both techniques to investigate the dynamics of NF-kB in the TNF-a signaling pathway.