The randomization of study subjects is commonly discussed for in vivo studies and is in general an accepted practice. However, for in vitro experiments, randomization is often seen by scientists as unnecessary effort with questionable impact.
In our Featured Publication, an assay artefact called Edge Effect was identified as a major factor leading to a decrease in data reproducibility and robustness when performing microplate cell culture-based studies.
The microplate is an important element of cell-based assay protocols and is designed so that multiple tests can run in parallel. During culturing, however, water and media commonly evaporate from the wells that are closest to the perimeter of the plate. The result is a variation in cell growth across the plate, while any media components, such as salt, can become concentrated to the point where they are harmful to the cells.
What can be done to reduce the edge effect?
Commonly used mitigation strategies include
       –  Excluding outer wells
       –  Use of humidified secondary containers
       –  Use of less sensitive cell lines
       –  Decreasing incubation time and duration of the experiment
       –  Limiting the number of inspections and unnecessary incubator openings
Unfortunately, these strategies often only partially mitigate positional artefacts. To reduce positional bias more effectively, randomization protocols may help. For example, a block randomization scheme has been developed, that coordinates placement of pre-defined blocks on the plate based on key experimental findings and assumptions about the distribution of assay bias and variability (see HERE).
Also for the multi-center study by Niepel et al, attempting to reproduce the results of an assay in which cultured cells were treated with anti-cancer drugs, randomized compound dispensing was recommended as a valuable way to reduce biases introduced by edge effects and irregular growth.
However, such randomization protocols increase study complexity significantly and usually requires automated processes via up-front programming of liquid handlers or pipet guides.
At this point, we would like to invite our readers to discuss with us how you manage to find the right balance between what is feasible (i.e. keeping pipetting errors and time requirements to a minimum) and what is necessary to increase reproducibility and robustness of in vitro microplate experiments. Please let us know what are your preferred methods and procedures – contact us or leave your comments HERE.