Design of Experiments (DoE) made easy
One of the most powerful tools in an experimenter’s toolkit is known as Design of Experiments, or DoE for short. It’s also one of the most intimidating to most experimenters. In this post, I’ll discuss briefly what is a DoE and provide a video demo showing how to create one using the Riffyn AddIn for JMP. We strongly encourage you to try DoEs to improve your assay, screening, and production processes!
What’s a DoE?
DoE typically refers to a basket of statistical tools, developed over the past 100 years, for optimal experimental designs. “Optimal” means an experiment that delivers the greatest amount of information about the behavior of your system under study using the fewest number of experimental samples or trials.
DoEs are intended for testing the effect of multiple parameters at the same time. These parameters are varied in different combinations across multiple trials (or runs) of your experimental system. The impact of each separate parameter is sorted out during data analysis using multivariate regression.
Modern statistical software applications provide a multitude of old and new DoE methodologies to choose from. But in recent decades a certain class of them, called “Optimal Design” methods, have come to dominate all the others. They’re great because they are able handle any complex design specification with almost push-button ease. They also give you the best possible designs for a fixed number of runs (if you are limited in how many runs you can afford to do).
Creating DoEs in Riffyn
Our own favorite tool for DoE is the JMP software. Accordingly, Riffyn SDE provides an AddIn for JMP that allows you to easily create Optimal DoEs for your Riffyn experiments, and then push the designs directly back into Riffyn.
In JMP, the Optimal Design methods are called “Custom Designs”. Riffyn calls this functionally automatically when you start a DoE using the Riffyn AddIn for JMP.
Rather than waste any more words here, we encourage you to take a look at this 6 minute video showing it in action:
The result is a set of runs with optimal combinations of parameters. When you execute the experiment with those parameter combinations, the resulting data set can be analyzed and modeled in JMP using linear regression to understand if and how the parameters affect your process or biological system.
Check out Riffyn's Discover Blog to learn more about optimal experimental design and analysis, such as how to optimize your measurement systems, how to structure data for analysis, or to clarify what machine learning is and how to apply it to your systems.
Fun Fact: In addition to Riffyn's JMP add-in, Riffyn also has a Data Connector for Tableau!