Biology Reference
In-Depth Information
architectures, spectral bleed-through due to the large spectral overlap of
fluorescent proteins prevents the detection of accurate, specific fluorescence
intensities for every fluorescent protein. Spectral unmixing (or compensa-
tion) is therefore required to distinguish each specific signal from cells
expressing multiple fluorescent proteins; in this method, a sample with a sin-
gle fluorescent molecule is imaged onto several spectral channels to deter-
mine the extent of detector bleed-through.
31-34
Under the assumption
that the spectrum of an individual chromophore linearly contaminates
each of the other image channels (
Fig. 8.1C
), spectral unmixing can be
executed by
subtraction of
the
spectral bleed-through of
each
chromophore from each individual spectrum (
Fig. 8.1D
).
33
Here is a practical example. Fluorescent dyes
D
1
and
D
2
are observed
through optical filter sets f
1
and f
2
that are prepared for the observation of
D
1
and
D
2
, respectively, but there is bleed-through of spectra of
D
1
and
D
2
to f
2
and f
1
, respectively. To compensate for the bleed-through, cells
expressing
D
1
or
D
2
alone are prepared and observed through both f
1
and
f
2
. When the fluorescence intensity of
D
i
obtained through the filter f
j
is
represented as
F
ij
, bleed-through of
D
1
to f
2
is
F
12
¼
a
12
F
11
¼
a
12
D
1
;
½
8
:
1
where
a
12
is an inclination of the linear function representing bleed-through
of
D
1
to the
D
2
channel (through the f
1
filter;
Fig. 8.1D
). Similarly, bleed-
through of
D
2
into the f
1
channel can be expressed as follows:
F
21
¼
a
21
F
22
¼
a
21
D
2
:
½
8
:
2
In cells expressing both
D
1
and
D
2
, the intensity obtained through the f
1
channel (
F
1
) contains the emission of
D
1
and the bleed-through of
D
2
(which is provided by Eq.
8.2
); similarly,
F
2
is the sum of
D
2
and
a
12
D
1
.
Therefore,
F
1
and
F
2
are given by
F
1
¼
D
1
þ
a
21
D
2
;
½
8
:
3
F
2
¼
a
12
D
1
þ
D
2
:
½
8
:
4
By defining the matrix
A
, the equations can be rewritten as
¼
A
F
1
F
2
D
1
D
2
1
a
21
a
12
A
¼
:
½
8
:
5
1
Thus, multiplying both sides of each equation by the inverse of the ma-
trix (
A
1
) from the left provides the corrected fluorescence intensities of
D
1
and
D
2
.
Search WWH ::
Custom Search