Biomedical Engineering Reference
In-Depth Information
10.2.5 Use of Factorial Design
During all phases of assay development, critical parameters may be assessed that
have the potential to influence one another, and if tested together, can be individually
identified [13]. Identification of these interactions can be exploited to make
decisions allowing one to pursue more robust assay conditions. Design of experi-
ment (DoE) is an approach that may be used to assess these factors and their
influence on assay performance, such as assay variability and signal-to-noise (S/N)
ratio. Such factors may include, but are not limited to, the following: cell growth
media, serum (bovine or human) levels and exposure time, cell number, state of cell
confluence, passage number, staining conditions, cell fixation, sample matrix
interferences, or effect of different anticoagulants. Use of DoE during the optimiza-
tion phase of assay development is advantageous since it can identify assay
conditions in a relatively small-scale experiment and isolate critical factors from
those having no effect, or those that may interact or influence one another in the
performance of the assay. These interactions may not be easily identified by testing
one factor at a time (OFAT). Appropriate factors and relevant levels for evaluation
are selected at the start of DoE, followed by the identification of the appropriate
responses to measure and how to set the numerical value [14-16]. Analysis of the
data will highlight the factors that have the greatest effects as well as the relevant
factor interactions that can be expanded upon with further experimentation. Once
conditions have been identified and the effects on particular responses such as signal
output have been assessed, further assay optimization can occur to ensure that the
cell line, assay output, and other reagents in the system have been appropriately
selected.
Since each drug candidate has a unique mode of action, the corresponding cell
source and assay reagents may also be unique. Therefore, appropriate characteriza-
tion of each component is required. Figure 10.4 shows the data output from a factorial
experiment designed to investigate multiple factors, including two concentrations of
cells (0.5
10 6 and 1
10 6 ), three concentrations of the fluorescent detector reagents
(1.0
g/mL), and three different fluorochromes (Alexa 448,
Alexa 647, and FITC). The assay responses measured were geometric mean fluor-
escent intensity (GMFI) and signal-to-noise ratio. Drug conjugated to Alexa 647 gave
the highest GMFI in comparison to Alexa 448 or FITC (Figure 10.4a), resulting in a
slight improvement of signal with a lower number of cells (Figure 10.4b). There was a
noted interaction between drug conjugated to FITC and the number of cells, with the
signal-to-noise ratio decreasing as the number of cells increased (Figure 10.4c).
Identification of this type of interaction allows the assay developer to select against
such combinations to develop a more robust assay. The condition selected for further
optimization used Alexa 647 as the conjugated fluorochrome and 0.5
m
g/mL, 2.5
m
g/mL, and 5
m
10 6 cells, and
even if this did not give the highest GMFI or S/N ratio, it was the lowest drug-
fluorochrome concentration that resulted in a reasonable signal. This lower concen-
tration of drug-fluorochrome improved the sensitivity of the assay since less
neutralizing antibody was required to inhibit a lower amount of drug than if a higher
concentration was chosen.
Search WWH ::




Custom Search