Opening the book…
An experimental unit is the smallest unit independently assigned a treatment. In a typical well-based experiment, it is the well. Several image fields from one well are subsamples of that well, and repeated readings of one sample are repeated measurements. They improve measurement precision but do not create additional independent cultures.
A technical replicate repeats a measurement or closely related preparation to examine measurement variation. Biological replication examines independent biological preparations or sources relevant to the intended conclusion. Separate donors can address donor variation; independent culture runs can address variation across preparations or days. State exactly what was replicated, because these labels alone do not explain the design.
Three treated wells on one plate can be separate treatment units, but all may share the same parent culture, reagent preparation, and incubation history. Repeating the study with independently prepared cultures provides information that one plate cannot supply. If the intended conclusion concerns donor variation, three passages from one donor do not replace three donors.
Randomization assigns treatments by chance to reduce systematic bias. Blocking groups units that share a source of variation, such as a plate, donor, or day. Place treatment comparisons and their controls within each block. If treatment is on one plate and control on another, a plate effect becomes another possible explanation.
Figure 13.1. Example balanced allocation on a 12-well plate. V is vehicle; L, M, and H are three treatment concentrations. Each condition occurs once in every row. This is an illustrative allocation, not a complete assay plate: blanks and other controls require planned additional wells or plates. Original diagram.
Evaporation, uneven seeding, delayed processing, and plate position can introduce systematic differences. Distribute groups rather than placing all controls on one side and all high-dose wells on the other. Use a model-appropriate plan to evaluate position effects. Record the full map, including blank wells and wells excluded before analysis.
Blinding means that the person measuring or scoring samples does not know the treatment assignment during that task. Coded images reduce the temptation to select especially convincing fields. Choose imaging positions, colony thresholds, and exclusion rules before viewing treatment outcomes. A documented contamination event or failed assay control may justify exclusion; an inconvenient value alone does not.
When multiple fields are measured per well, first summarize them by a stated method to obtain a well-level result. Preserve the individual measurements for checking variation. For three wells with four fields each, the dataset contains twelve field measurements and three well-level units, not twelve independently treated cultures.
This page titled 13.4: Replication and Plate Layout was last modified on Wed, 21 Oct 2026 01:34:26 GMT and is shared under a CC BY-NC-SA 4.0 license and was authored, remixed, and/or curated by Emalee MacKenzie.
The book could not be loaded just now.