Using custom datasets in SNAP

I am interested in using SNAP to perform geolocation on images I have created by processing ALOS 1.0 data sets (or other data sets that contain orbit information but not scene lat/lon coordinates).

I have gone through the georectification tutorial using ALOS 1.1 level data sets with SNAP - what I want to know is: What is the bare minimum information SNAP needs to geolocate the image? I removed the lat/lon coordinates from an ALOS level 1.1 data set’s IMG and LED files, and SNAP was still able to geolocate the image correctly. If I write my processed image to the same CEOS format used in the ALOS 1.1 TRL file, can I use that along with the VOL, LED, and IMG files in 1.0 data sets to get the scene lat/lon coordinates in SNAP?

I had a look in the code.
It looks like that the reader tries to create the geolocation information from several sources. Depending on which it can find first and also depending on the type of the product and if the product is created by the IPF.
If you like you can have a look at the code too. This is where the generation of the geo-coding starts.

In most cases, the information is taken from the leader file.
Maybe @lveci can tell more about the minimum requirements.