ADF creation steps:

1. The user is asked for array application (GE | CGH | ChIP) and the type of microarray design (simple or complex)
   Depending on the application choice, all mandatory items fields for the use case will be created in the corresponding tabular window. In case of simple array design (one-to-one reporter composite sequence relation), the composite sequence data will be automatically generated from reporter data. For complex array design, the user is expected to provide the composite sequence data.

2. The user will fill in data in the following order:
   
   • first, Header;
   • then, FeatureReporter;
   • finally, CompositeSequence (if the user has selected the complex array design option);
     
   The user will have the possibility to go to the previous data part, at any moment (e.g.: from reporter part to header part), for modification.
   FeatureReporter and CompositeSequence (if applicable) data can be pasted in the corresponding table or imported from a spotter output file. In the former case, the application will check the file format and convert to « ADF format »way. That data could check during the import or only during the checking of all data at the end of form filling.

   In the help process, there are two interactive processes:

   • For database entry fields, it is asked to the user which database identifiers are in use by selecting the database from a list of approved databases or a user specific database. For a new user specific database, the name will take from user input and it will be automatically added in the Header user defined item.

   • For controlled vocabulary fields, it is possible to provide an auto-complete mechanism to reduce typing errors and implemented embedded controlled vocabulary (CV) checks. The user will just have to choose between possible values retrieved from ontology (MGED or others) or he will have the possibility to used a new term -by using new_term[NEW TERM]-. In this case, the new term should be proposed to the MGED ontology, as well (by an e-mail to MGED Ontology mailing list[mgeb] or simply by a popup window).

   After each part filling, there is a data check. For FeatureReporter and Composites data, the process happen in the following order:

   1. Name check (based on user feedback) - empty, not unique, unstandard characters, ...-> Stop and display erroneous entries . Ask for corrective action;
   2. Check header order;
   3. Validate: check whole data from ADF data checklist (from ADF checker functionality)

3. Once checking is completed, the user is prompted for an optional "Automatic annotation":
   • The user could ask to retrieve annotation from G.O.[go]:"Get GO annotation"; or from Ensembl EnsMart:"get EnsMart annotation";
   • It is asked to the user for database identifier to be used for query;
   • A query is send query to EnsMart form or annotations are retrieved from GO annotation files (see Automatic annotation modules).
4. Finally, the user can export the ADF to file or convert array design in MAGE-ML.

The version of the used database will be specified in the data, to avoid confusion between annotation versions.