Biology Reference
In-Depth Information
two different cameras. Detector variability is thus spatially uncorrelated
and may strongly impact the ratio values (
Fig. 5.24B and C
). It is thus ab-
solutely necessary to suppress this field illumination inhomogeneity by ac-
quiring reference images using fluorescent solutions. When differences in
field illumination persist, corrections can be carried out by normalizing
pixel intensities using reference images. This procedure is called “shading
correction”.
141
Besides correcting for spatial field inhomogeneities, shading correction
also intrinsically calibrates the absolute ratio value. Shading images are ac-
quired with a uniform sample of a fluorescent chromophore chosen to emit
fluorescence at both wavelengths of interest. In the particular case of CFP-
like and YFP-like chromophores, a suitable fluorescent dye is Coumarin
343, which emits at both wavelengths. When each acquired image is divided
by its corresponding shading image, by definition the ratio of the Coumarin
343 dye is equal to 1.
74,142
This method thus defines ratio values with respect
to a reference chromophore and therefore defines the “balance” between
the two channels. This procedure allows easier comparison between
setups that may have different spectral sensitivities.
7.1.2 Image misregistration between acquisition channels
In the sequential ratio setup, the images acquired at two wavelengths may
not be perfectly registered. This may be due to wedge effects caused by
the nonparallel faces of the emission filter. This is easily corrected by the
translation of one of the images by a few pixels, and algorithms exist that
determine the optimal subpixel registration correction.
143
In addition, if
the objective suffers from chromatic aberration, sequential ratio allows a fo-
cus correction by a fixed value which can be accurately determined using
fluorescent beads. When cellular movement occurs at the timescale of the
sequential acquisition, no correction can be applied and the system is just
not fast enough for measuring the biological event of interest, and simulta-
neous detection with dual view or dual channels may be needed. In these
configurations, in addition to the linear translation of one image with respect
to the other, nonlinear deformation may occur as a result of different optical
paths followed by both wavelengths, and it will be necessary to apply more
elaborate algorithms that correct for image skewing. Misregistration effects
are nonetheless a common cause of data misinterpretation and the experi-
menter should look carefully at the images for border effects which usually
show a typical rainbow appearance (
Fig. 5.24D
)
Search WWH ::
Custom Search