WORKFLOW_EXISTING.md

Loading an Existing Ontology into Breedbase

These steps outline the general workflow of updating an ontology (by modifying its Trait Workbook), creating the standard- and sgn-obo files, and loading the sgn-obo file into a breedbase instance.

1. Edit / Add Traits in the Trait Workbook

   The Trait Workbook is an excel file containing separate worksheets for each data type: Variables, Traits, Methods, Scales, Trait Classes, and Ontology Root Information. Example: traits.xlsx file in the sugar kelp example directory.

   If there is an exisiting Trait Dictionary for the crop ontology, the Trait Workbook can be created using the create_tw.pl script.

   For more information on creating a Trait Workbook for a new crop (when an existing Trait Dictionary does not exist), see the Creating a New Ontology workflow

2. Generate the Trait Dictionary and/or standard-obo files:

   perl build_traits.pl -t traits.csv -o traits.obo -u DJW -i T3 -v traits.xlsx

   The traits.csv file can be used to update the Crop Ontology website

   The traits.obo file will be used to create the obo file that is necessary to load the trait ontology into breedbase

   The -i T3 option filters the variables by the Institution column and will only include variables that contain T3 as the Institution.

3. Convert the standard-obo file to an sgn-obo file:

   perl convert_obo.pl --remove-synonyms -d wheat_trait -o sgn.obo -v traits.obo

   The sgn.obo file can be used to load the traits into breeDBase

4. Load the traits into breeDBase

   4A) Load the ontology

    cd /home/production/cxgn/Chado/chado/bin
    perl ./gmod_load_cvterms.pl 
        -H localhost -D {db_name} -d Pg -r postgres -p "{postgress password}"
        -s CO_360 -n sugar_kelp_trait -uv /path/to/sgn.obo
   

   4B) Connect the ontology terms

    perl ./gmod_make_cvtermpath.pl 
        -H localhost -D {db_name} -d Pg -u postgres -p "{postgress password}" 
        -c sugar_kelp_trait -v
   

   4C) Tag the ontology (first time only)

    Update the  cvprop table to mark new cv as a 'trait_ontology'
        cv_id = cv.id of 'sugar_kelp_trait'
        type_id = cvterm.cvterm_id of 'trait_ontology'
   

   4D) Make sure sgn_local.conf variables are correctly set

    trait_ontology_db_name          CO_360
    trait_cv_name                   sugar_kelp_trait
    onto_root_namespaces            CO_360 (Sugar Kelp Traits), ...
    trait_variable_onto_namespace   CO_360 (Sugar Kelp Traits), ...