Bioimage Informatics Algorithms: Detection, Localization, and Tracking Fluorescent Particles
D. Sage
Workshop "Collective Motion of Active Swimmers" (MOTIMO'13), Valbonne, French Republic, September 25-26, 2013.
The emergence of the time-lapse fluorescence microscopy have allowed the observation and the quantification of molecular dynamics within living cells. The quantification is significant after the analysis of large number of cells in high-throughput screening environment that requires advanced image-analysis tools. These tools mainly involve the detection, the localization and the tracking of fluorescent proteins that appear as small bright spot in images.
In this context, the localization microscopy (PALM/STORM) achieves super-resolution by rendering positions of accurate localizations of single-molecule emitters. For this task, many software have been developed. Here, we summarized a study to benchmark these localization software. To conduct the evaluation, we designed bio-inspired datasets generated with a faithful image formation model and we established fair metrics for the assessment.
In molecular biology, numerous relevant questions are related to the motility and the diffusion of specific proteins. These information are extracted from fluorescence images by tracking algorithms. The tracking task is complicated by the crossing trajectories of similar particles. We adopted a global optimization framework based on the dynamic programming algorithm. This algorithm code the motion model in a merit function. It is able to track single particle in noisy image even when the photons emission is very dim. For certain applications, additional image modality (e.g. phase contrast) is integrated in the image-analysis system, helping to identify the cell.
Finally, we present several implementations of the detection and tracking modules as ImageJ plugins for biological applications: study dynamics of telomeres, recording the circadian gene expression, measure the growth-rate of a colony of bacteria, construction of the cell lineage in FUCCI images, quantification of the asymmetric cell division.
@INPROCEEDINGS(http://bigwww.epfl.ch/publications/sage1302.html,
AUTHOR="Sage, D.",
TITLE="Bioimage Informatics Algorithms: {D}etection, Localization, and
Tracking Fluorescent Particles",
BOOKTITLE="Workshop ``Collective Motion of Active Swimmers''
({MOTIMO'13})",
YEAR="2013",
editor="",
volume="",
series="",
pages="",
address="Valbonne, French Republic",
month="September 25-26,",
organization="",
publisher="",
note="")