Deconvolution and Denoising for Confocal Microscopy - Modeling in Computational Biology and Biomedicine - page 114

label virtually all cell components with chemically synthesized fluorophores lead
to a large variety of fluorescence imaging strategies. Furthermore, the discovery of
inherently fluorescent gene products such as the Green Fluorescent Protein (GFP),
opened the way for biologists to genetically tag a specific protein in a living
organism. Fluorescence imaging, being essentially a dark-field method, exhibits an
exceptionally high image contrast compared to other wide field methods and there-
fore became the method of choice to localize molecules in cells. It is no wonder that
Osamu Shimomura, Martin Chalfie and Roger Y. Tsien [ 81 ] were jointly awarded
the 2008 Nobel Prize in Chemistry for the discovery of the GFP, demonstrating it's
value and contributing to our understanding of how they fluoresce.

4.1.2.1 Biological Context

Nowadays, a major interest for cell biologists is to identify and precisely localize
macromolecules, i.e., the building blocks of cellular organelles and supra-molecular
complexes. A first breakthrough to allow such molecular imaging was due to the
development of methods based on the use of antibodies and in-situ hybridization
techniques. These approaches combined with fluorescence microscopy made it
possible to image with high contrast individual cellular components using fluo-
rescent dyes. A second breakthrough resulted from the discovery of the so called
“fluorescent proteins”. These auto-fluorescent molecules can be physically attached
to any protein, for which the gene is known, and re-introduced into a living cell.
It has the ability to non-destructively image tagged-molecules in real-time with
optical resolution. It was mainly this tool that boosted microscopy in recent years.
In animal cells, the cellular structures can be labeled with fluorescence dyes, using
histochemical techniques, and efficiently imaged, if little or no background auto-
fluorescence is present. A confocal microscope can generate several optical sections,
at different depths, inside a chemically fixed or living specimen. With the help
of these optical sections, a 3-D data set of the specimen, representing the spatial
distribution of the fluorescent labeled cell components, can be constructed. Due
to this optical sectioning capacity, the confocal microscope has become a major
workhorse instrument in many biomedical research laboratories. Variants of the
confocal microscope have been developed to increase imaging speed (spinning disk
confocal) and tissue penetration depth (multi-photon) [ 37 ]. For an overview of these
confocal variants, see the handbook of Pawley [ 57 ].
In life-cell imaging, questions on the functioning of macromolecular machines
remain largely unanswered since many cell components have dimensions below the
200 nm x-y resolution limit (and 400 nm axial resolution limit) obtainable from a
standard light microscope. Recently introduced super resolution optical imaging
techniques in combination with adapted deconvolution strategies will be crucial to
address some of these questions.
However, imaging biological samples, especially living specimens, with a micro-
scope is difficult and tricky. Although a multitude of parameters will influence the
final quality and faithfulness of the image, we can simplify and state that imaging