Quasi-replicate factors create runs that are needed for you to calculate a response of interest, rather than to investigate a possible effect on the system you’re studying. As such, quasi-replicates do not contribute to the calculation or power of your DOE design. They allow you to define a set of levels by which each run in your calculated design will be replicated against.
For example, imagine you want a time course measurement as your response, but you are unable to sample from the same reaction at each time point. What you might require instead is a replicate of each of your reactions - one for each time point measurement. In this instance, you could define a factor called “Time Series”, with levels that describe each of the time points that you want to include in your time series (Figure 1).
Figure 1. Quasi-replicate factors can be used to make sure every run in the calculated DOE design is replicated once for each of the levels defined. A Time Series factor can be defined with three time points as levels. Each of these levels will generate a quasi-replicate (QR) of the runs in your design. Ultimately, a time series data set can be generated, from which you might want to analyse the gradient of a linear fit as your response.
Another example is a choice of treatments, which can include an inhibitor or activator compound for the system you are investigating, so that you can calculate a Z’ assay performance response (ref). Each run in your design will be replicated once for each of the two treatments, such that data from the inhibitor quasi-replicate is the baseline measurement, or negative control, and data from the activator quasi-replicate is the maximum signal, or positive control. These two measurements, along with technical replicates, which give a measure of variability, can be used to calculate the Z’ response for analysis (Figure 2).
Figure 2. Calculating the assay performance response Z’ from data generated by quasi-replicates. To calculate the Z’ assay performance response, you need runs where the assay is run with an inhibitor, to generate the minimum signal, and an activator, to obtain the maximum signal, which can be done by quasi-replicating an entire DOE design by treatment variable. Z’ is calculated by the equation where mt and mb are the highest and lowest measured signals when in the presence of activator or inhibitor, respectively, and st and sb are the standard deviation at the high and low measured signals. Z’ values >0.5 are considered excellent, between 0 and 0.5 are ok, and below 0 is reflective of an unsuitable assay.
The most important thing to remember when using quasi-replicates is that they are used to help generate the data to calculate a response. If you want to investigate the impact of variables in your actual system, then you should consider setting these as design factors, not quasi-replicates.
To learn how to define a quasi-replicate factor in Synthace, click here.
To learn how to define multiple quasi-replicate factors in Synthace, click here.