PhenoRipper does not perform supervised classification. By default, when PhenoRipper compares microscopy images, it does not use prior information about the experimental conditions underlying the data. However, users have the option to allow PhenoRipper to make use of experimental annotation in two ways:
1-Displaying results: The scatter plots of profiles in PhenoBrowser can be grouped, colored, or labelled based on user annotation. This visualization helps users explore and interpret the results produced by PhenoRipper, even though the annotation itself is not used during the analysis.
2-Subsampling data: To identify stereotypical block types (the features used in profiling and comparing images), PhenoRipper analyzes a random subset of blocks selected from a random subset of all images. The user has the option to bias this default behavior by telling PhenoRipper which sets of images could be considered as replicates. This option can be useful when the user choses to subset only a very small number of images and wants to avoid redundant sampling.
(
hide)