1. Global settings

2. Experiment tab

Place where users will be able to run complete experiments: create models and optionally project them.
Labels should be as generic as possible (not biased by ecological niche modelling).

Users should be able to choose one or more set of points to run different models from each one.
Each set of points can have an associated (user defined) label.
Points may come from different data sources (local text files, local database, other remote sources).
Need to decide about how to represent spatial references (such as wkt) and how to include them in the interface when they are not provided by the data sources.
When including a set of occurrence points users will probably need another interface to select an existing data source (or create a new one).
Should we provide a SQLLite database to serve as an internal data source?
Masks can be individually selected for each set of points.

Users should be able to choose one or more algorithms.
Since algorithm parameters can have predefined settings according to users preferences, each predefined setting could be a kind of algorithm profile with an associated label (better then showing a long line of parameters).
By clicking on the algorithm profile users will be able to see the parameter values and change them.

Removed the original "Trainning dataset" group box (not related to reprojection).