Currently one experiment can have multiple alternate phasing configurations. User can try phasing the same data in a different way and has all the information recorded in the configuration files with different ids.
This model can be extended to the preprocessing data with different configurations.