Experiments where we think percent replicating is potentially wrongly penalized

[bethac07]

Percent replicating is calculated with only two replicates in the following places, which we think unfairly penalizes those cases since the null is less robust. We will not say in the paper that those conditions are worse by replicating, since we can't be certain; we will only discuss their matching.

• In the time delay experiment, days 1 and 4 only half of the rows were imaged (Days 0, 14, and 28 had full 4 replicates)
• In the drug selection experiment, the selected cells (unselected had 4 replicates)
• In the cell density experiment, the +/- 20% plates (median density had 4 replicates)

This is outside of plates we think may have had small technical artifacts - see this link for more information.