Flag high coefficient of variation

[jakebeal]

Steve Haase is looking for flags when a sample has a high coefficient of variation, which can indicate machine failures, such as air bubbles and malfunctioning fluidics.

Some key questions to resolve in order to define this well are:

• Is this CV in arithmetic or geometric statistics?
• Is this an absolute threshold or relative to other samples? (e.g., CV >50% higher than median sample)
• Is this on certain channels, on every individual channel, or jointly across channels?
• Should this be a parametric measure, when dealing with transient transfections that naturally have a wide range?