This video demonstrates the DOE data analysis functionalities in Synthace.
Analysing the results from your DOE, or Response Analysis, you should think of as a scientific experiment in its own right. It is an iterative process of investigating and choosing which of your data to analyse, fitting models, assessing the quality of the models, rinse and repeat until you get to a useful model you have confidence in.
All models are wrong, but some models are useful. Always remember that the purpose of your response analysis is to generate a useful model that helps you make the decisions of what you should do next. It is not the process of trying to fit a model that you think is perfect, but never will be… but more on this later.
In the following sets of documents we will take you through all you need to know to perform your Response Analysis in Synthace aiming to demystify the complexities of data modelling.
Just the name can often send shivers down the spines of Biological Scientists, but really there is no need. Let’s break down the complexities and you will soon learn that really this is quite simple!
Sanity check time! Before you spend time in trying to fit models to data that maybe don’t have any signal to fit to let’s just take a look at some simple plots to get acquainted with the shape of your data and get an initial look at any obvious trends.
Of all the different variables you tested in your DOE which ones are important? Continue reading to learn how to identify what should be modelled and how to fit the models.
Clothes rarely come in a one size fits all, just because you can climb into a t-shirt doesn’t mean that it fits well. The same is true for models and data. Learn how to identify when a model is a good fit and what to look out for when it maybe isn’t.
Often you need to try on a few t-shirts before you find the right one. Again the same is true for models, you need to try a few to find the right one. Learn how to iterate on your model building to improve its suitability.
Once you have a model with a suitable fit to your data what can you do with it? Read on to learn what your models are useful for and what you can learn from them.
Useful models can help you better understand the biological space you are studying, how can you leverage that insight to decide what to do next? Learn what to do next when you have a useful model or what the next move is when you are not able to get to a useful model.
Once you have created several models from different underlying sets of data or transformed responses you may need a quick reference to see what you’ve done. This page explains how Synthace’s features for browsing models are implemented and how to use them.